There’s a first time for everything A study of the decision-making strategies of first-time voters

Dina Heider Hov

Master’s Thesis

Department of Political Science Faculty of Social Sciences UNIVERSITY OF Spring 2019

Number of words: 33 208

I

II There’s a first time for everything A study of the decision-making strategies of first-time voters

III

© Dina Heider Hov 2019 There’s a first time for everything. A study of decision-making strategies employed by first-time voters http://www.duo.uio.no/ Printed: Allkopi Word count: 33 208

IV Abstract

How to first-time voters decide which party to vote for? This thesis studies decision processes and strategies among first-time voters in . I argue that the traditional way of explaining voter behavior, using issue opinions and political values, might not give the whole picture when looking at first-time voters, or other voters that have not yet developed a stance on many of these issues. Thus, this thesis approaches this electoral group in a more process- oriented way, focusing not only on the decision, but more on the cognitive processes leading up to the decision. I ask “what type of heuristics or decision strategies do people employ when deciding on their vote?”, with a special emphasis on first-time voters.

The theoretical point of departure is the framework and the five voter decision-making strategies proposed by Lau and Redlawsk (2006, 2018). I conducted semi-structured interviews with first-time voters from three different high schools, which I later categorized within the preexisting strategies developed by Lau and Redlawsk. I identified four of the five decision-making strategies. I found that the best predictor for decision-making strategy is the voters subjectively perceived importance of task. The more important the perception of the decision task, the higher probability of employing a cognitively demanding strategy. The lower the interest, the higher probability of basing one’s decision on cognitive short-cuts like decision heuristics. I also found that voters in this sample, different from the strategies proposed by Lau and Redlawsk, went through a two-step decision process. I argue that due to the Norwegian multi-party system, voters in the first step decides which parties they perceive as potential candidates for their vote, more or less consciously eliminating several alternatives, and then in the second step they decide whom to vote for. The type of decision strategy employed in the second step, ultimately depend on the voters perceived importance of the decision task.

V

VI Acknowledgements

First of all, I am forever grateful to Johannes Bergh and Bernt Aardal for giving me the opportunity to work as their research assistant on the Norwegian National Election Studies. It has truly been, and still is, the most fulfilling job I have ever had. It has given me confidence and faith in my own abilities as a researcher, and it has motivated my future aspirations.

I want to thank Atle Haugsgjerd, your genuine passion for the field is truly inspiring, I appreciate all of our conversations, and the overall encouragement you have given me. Øyvind Bugge Solheim, thank you so much for invaluable guidance, discussions, comments and motivation throughout the whole process. Your faith in me has meant a lot, and I am grateful that I have met someone as selfless as you. Audun Bayer and Audun Fladmoe, thank you for helping me getting in contact with informants, and for sharing the same enthusiasm for cross-country skiing and Strava. Thank you to Hanna, Live, Sahra, Milla, Karin, Stine and last but not least Simon for making my time at ISF a truly wonderful experience. Your friendship has brought joy and inspiration and has sat the bar high for future colleagues.

I also want to thank the Institute for Social Research, which has introduced me to a truly fascinating world of great research and great people. Also, special thanks to the participants of Politikkseminaret, for taking time to read and comment on my thesis.

I also have to thank my best friends, Catharina and Julie for being my biggest supporters and motivators, not only in this process but in life in general. I love you. And Jacob, thank you for your support and love. I also have to thank Mimmi and bestefar, you have been so involved and caring throughout this whole process.

And again, thank you so much Johannes Bergh, you have not only been my boss, but my supervisor. Thank you for answering hundreds of “stupid questions”, and for always taking time to discuss and comment on my work, and for being very positive and motivational when doing so. Thank you.

Dina Heider Hov Torshov, Oslo, Norway July 2019

VII

VIII Contents

1 Introduction ...... 1

2 Theory ...... 5

2.1 Classical theories of voting ...... 5 2.1.1 The sociological theory ...... 6 2.1.2 The Michigan Model ...... 7 2.1.3 Norwegian election studies ...... 8

2.2 Lau and Redlawsk’ “New Theory of Voter Decision Making” ...... 9

2.3 Behavioral Decision Theory and bounded rationality ...... 11 2.3.1 Our cognitive structure and its limits ...... 12

2.4 How do we cope with our inescapable cognitive limitations? ...... 14 2.4.1 Judgement heuristics ...... 14 2.4.2 Simplifying mechanisms for making choices ...... 15 2.4.3 Decision-making strategies and cognitive styles in general ...... 17

2.5 Lau and Redlawsk’ five voter decision strategies ...... 18

2.6 What other factors can affect the decision strategy? ...... 23

2.7 Summary ...... 26

3 Method -Research design ...... 27

3.1 Data ...... 27 3.1.1 Population: First-time voters ...... 27 3.1.2 Sample frame ...... 29 3.1.3 Final sample ...... 30

3.2 Internal and external validity within qualitative research: A discussion ...... 31 3.2.1 Sample frame and sample: External validity ...... 32 3.2.2 Data: Internal validity ...... 33

3.3 On interviewing ...... 34 3.3.1 Semi-structured interviews ...... 34 3.3.2 The interview process and the interview guide ...... 36

3.4 On context – How can the high school context affect the decision strategies employed by Norwegian first-time voters? ...... 38

3.5 The analyzing process of the data ...... 39

IX 4 Analysis ...... 40

4.1 Strategy 1: Rational choice decision-making ...... 40

4.2 Strategy 2: Confirmatory decision-making ...... 44

4.3 Strategy 3: Fast and Frugal decision-making ...... 50

4.4 Strategy 4: Heuristic-Based decision making ...... 53

4.5 Strategy 5: Gut decision-making strategy ...... 60

5 Discussion and concluding remarks ...... 62

5.1 Do first-time voters in Norway use the same decisions strategies as general voters in the US? ...... 62

5.2 General findings when testing the framework by Lau and Redlawsk ...... 66

5.3 Two-step decision process and the concept of “party set” ...... 68

5.4 No “gut decision making” detected among first-time voters; a case for voting rights for 16-year-old’s? ...... 71

5.5 Future research ...... 72

References ...... 74

Appendix ...... 79

A.3.1 Overview of interview ...... 79

A.3.2 Interview guide ...... 80

X

Two days after the 2017 national election I asked my friend what she had voted. She answered the Red Party. When I asked why, she said “I didn’t know what party to vote for, but I saw a post on Snapchat that one of my politically active friends had posted, and it said: If you’re in doubt vote for the Red Party, they promote gender equality, and I thought that sounded good, so I just voted for them”

XI

I 1 Introduction

The conversation you just read was what inspired the theme of this thesis. That a simple post on Snapchat could be the decisive factor for a voter’s vote choice was for me, a political science student who has spent years learning about complex statistical models explaining voter behavior, thought-provoking. How does the episode I just described fit into the classical theories of voting?

This thesis studies the decision strategies employed by first-time voters. Based on theory from traditional election studies, modern psychology and behavioral decision theory, the goal is to identify different decision strategies employed by voters when deciding on their vote. How do they make the decision?

The outcome of an election has far-reaching implications. Hence, the desire to explain voter behavior has long been a subject of great interest. Whom will people vote for, and why? Many have tried to answer these questions but with different underlying theories of human behavior, mainly originated from the disciplines of economics, psychology and sociology. This has resulted in a wide range of voter behavior models over the past half-century. Theories of voting behavior span a wide range of alternative factors, from sociological effects of group membership, to psychological factors of motivated reasoning. There is also research emphasizing class (class voting), political issues (issue voting), and the electorates evaluation of the economy (economic voting). Many of these models have indeed proved capable of predicting and explaining voter behavior, and much of the information we have today about voters and voter behavior are much due to the earlier theories on voting, such as the party identification model (Campbell, 1960, p. 2). This theory, also known as the Michigan model, emphasize the significance of a person’s social characteristics because of how it influences the formation of one’s broad political orientation and group identity, which impact values and issue opinions, which again affect voting choice. By putting several independent variables into a regression the model can, with high accuracy, predict which party voters will choose (Lau & Redlawsk, 2006, p. 3).

Humans are cognitively limited information processors (Anderson, 1983, 2002), and it takes tremendous amount of effort to gather and understand the information necessary to employ

1 many of the standard predictors used in such models (Lau, Kleinberg, & Ditonto, 2018, p. 326). Elections constitute a highly complex decision task, and even if we as voters were willing to put in the hours to make the “perfect” vote choice, we may not have the cognitive capabilities required to do so. No matter how motivated a decision maker may be, the architecture of human cognition creates very real limits on how much information that can be processed, stored and utilized at a given moment (Anderson, 1983, 2002).We are not able to process all available information about each party, their politics and their value priorities, speculate about possible government coalitions, tactical evaluations, and in addition consider one’s own interests and values. Elections are overwhelming; however, most people do cast a ballot. In Norway election turnout have been over 75 percent since 1930 (Aardal, 2019)(aardal.info), which means that three out of four manage to choose a party one way or another. How do we reach that decision?

When facing decision tasks in general, we are often confronted with the dilemma that we want to make good decisions with as little effort as possible. In order to do so, decision theory research posit that people employ different types of cognitive decision strategies (Plous, 1993). These strategies is called heuristics, and are also known as mental short cuts that people more or less automatically or unconsciously employ to simplify the process of decision making. Think about all the decisions we make every day; what to eat, what to wear, what to buy, what TV-show to watch, which podcast to listen to etc. If we were to reflect carefully on every decision task we meet, we would probably not even make it out of bed. Thus, we employ heuristics which are both time-saving and help us make as good decisions as possible without having to evaluate every alternative and possible outcome. It can for example be that you buy sports gear from the same brand as used by a medal winning athlete, trusting that this athlete will only use the best, or that you eat at a restaurant because a friend who you consider having good taste recommended it, or you chose the same career path as your parents because it is the most familiar to you. How these heuristics are constructed and employed varies between individuals and the subjectively perceived “nature” of the decision task. If you are not very interested in sports gear you have probably not noticed what brand certain athletes use and will probably just buy whatever the shop assistant recommends. If we now turn to the election, and consider the complexity of the voting decision task, and the varied level of interest in politics in general, what type of heuristics or decision strategies do people employ when deciding on their vote?

2 This question will be the main concern of this thesis. By studying what decision strategies voters employ, we could get a better understanding not only of the outcome of an election but also the cognitive process voters go through before an election, which again could contribute to a more comprehensive insight on voter behavior. In order to get a deeper understanding of people’s voting behavior and the different decision strategies voters employ, I have conducted interviews with first-time voters from three different high schools in Norway. I chose to limit the study to first-time voters for several reasons. First of all, voting for the first time must involve some kind of active decision process, considering the fact that this is their very first election. Several studies have shown that voting in general is habitual, casting a ballot in one election increases one’s propensity to go to the polls in the future (Denny & Doyle, 2008; Gerber, Green, & Shachar, 2003). There are also studies showing that vote choice in itself is habitual (Shachar, 2003, p. 267). Looking at data from the Norwegian election studies we see that the average voting stability1 in Norway since 1965 has been around 72 % (Aardal & Bergh, 2015, p. 20). Thus, the probability of detecting a decision strategy is higher amongst first-time voters, compared to veteran voters. Also, many of the classical voter explanation models that are used today explain vote choice based on a voter’s social characteristics, issue opinions and political values. All these variables need not yet to be present among first time voters. This is not to say that teenagers are disengaged and uninterested in politics, a notion that has long since been refuted (Ødegård, 2003). We know that many adolescents are engaged, but that this is often not expressed through the traditional political channels such as elections and participation in organizations, where they are relatively underrepresented. Their engagement could be more case specific and sporadic, and their values and opinions do not necessarily have clear social ties. This indicates that the traditional way of explaining voter behavior might not be as useful when studying first-time voters. By approaching this group of the electorate in a more process-oriented way, by analyzing how people voting for the first time reach a decision, and what strategies they utilize when doing so, this thesis will hopefully contribute to enhance our overall understanding of voter behavior and decision making in general, knowing that the first election might for many be decisive in the future.

To study voter’s decision strategies is not a very common approach in the field of electoral studies. There does however exist some research on this topic. In the book How Voters Decide from 2006, Lau and Redlawsk propose what they humbly call a “paradigm-shifting framework for studying voter decision making” (Lau & Redlawsk, 2006). Their main

1 Voting stability means how many voters who votes for the same party in more than two or more consecutive elections. The opposite is volatility, which is how many that change their vote from one party in one election to another party in another election.

3 concern is how voters process information during election campaigns, and how differences of various factors such as political sophistication and cognitive capacity is reflected in how they reach decisions. By utilizing methods such as process-tracing and experiments, they were able to identify four different decision strategies which voters applied when deciding on their vote. In 2018, a fifth category was added by Lau and his colleagues in a follow-up article called “Measuring Voter Decision Strategies in Political Behavior and Public Opinion Research”. The original framework proposed by Lau and Redlawsk in 2006, and the extended one from 2018 will be the point of departure for this study. I will use their strategies as a theoretical backdrop when analyzing the interviews. However, Lau and Redlawsk’ framework was created within the social and political context of the US, which is, in many important respects2, different from the Norwegian context of this study. In addition, the average age of those who participated in Lau and Redlawsk’s research was 50 years, whereas the focus in this study is on first-time voters. Thus, I expect that there will be differences, if not in the strategies employed per se, but in how and which factors that influence the strategies.

The main contribution of this thesis is an in-depth analysis of the decision strategies first-time voters in Norway employ. By using the unique data collected specifically for this purpose, the goal is to identify and categorize the different strategies found in the sample and discuss to what extent the framework proposed by Lau and Ledlawsk fits the case of Norway and first- time voters.

Outline of thesis Thus far, the topic and research question have been presented. In chapter 2, an account of the theoretical framework of the thesis. The assumptions made by these theories will be utilized to derive some general expectations. In chapter 3, the method for data collection, the sample and a methodological discussion is presented. The analysis is presented in Chapter 4. In the last chapter, a discussion on the most important findings is accounted for, followed by concluding remarks and suggestions for future research.

2 The political system; multiparty system in Norway versus two-party system in the US, the frequency of elections is also very different in Norway compared to the US.

4 2 Theory

There are numerous studies, theories and models aimed at predicting and explaining voting and voter behavior. However, the focus of this study is mainly on the decision process itself, and how people voting for the first time are able to maneuver through the complex sea of information and complete the cognitively difficult task of selecting a party. Classical voting theories focus on how different social characteristics affects voting choice (Lau & Redlawsk, 2006, p. 21). What this study does is to incorporate elements from psychology and behavior decision theory about the structure of our cognition, and how bounded rationality forces us to employ heuristics (cognitive short cuts) and strategies in order to make decisions. By combining these new elements with the traditional factors, the goal is to obtain a more comprehensive understanding of the voting process. First, the theory chapter begin with a short review on some of the most influential voting theories, which for decades have functioned as the corner stones of most electoral studies. The key arguments of these theories are still relevant in this study, even if they are not the main focus. Secondly, I will give an account of some of the basic assumptions from behavioral decision theory. It is the understanding of humans as cognitively limited information processors that is the premise upon which the different voter decision strategies are built. In the third and last section of the chapter, a review of the “new process-oriented framework”, followed by an account of the five different voter decision strategies developed by Lau and Redlawsk.

2.1 Classical theories of voting Some of the first forms of election studies can be dated as far back as the 19th century (Aardal, 2017, p. 293) when geographers attempted to explain voting choice by looking at the official election results and maps. They were able to show how the support of political parties varied in relation to economic, cultural and social characteristics of the electorates living areas. Even if these first analyses could give some information about the relationship between geographical/social features and voting choice, they lacked depth and explanatory power. In addition, using aggregated data, which was the only data available at the time, could often lead to what we call the ecological fallacy, which is when inferences are drawn based on characteristics of a constituency or area of residence to individual attitudes and actions (Aardal, 2017, p. 293). To avoid this fallacy, electoral researchers started to collect data at an individual level through surveys. The information obtained through these surveys enabled the development of several voting theories aiming at explaining voting behavior. In the following, I will begin with outlining some of the most dominant theories of voting; the

5 sociological theory of voting (Columbia school), and the psychological theory of voting (i.e. the party identification theory/the Michigan school).

2.1.1 The sociological theory One of the most prominent theories of voting came in the 1960s; the sociological theory, also known as the Columbia school model. The supporters of this theory emphasized the group basis of voting, and argued that the social layer one belonged to would significantly determine one´s political preferences. Lazarsfeld, Berelson and Gaudet were some of the first to present a sociological voting theory approach in their well-known study The People’s Choice (1954). They looked at the relationship between certain social characteristics such as class, race, and religion and the support for political parties. They found a systematic structure in what type of citizens that voted for particular political parties. This discovery suggested that as long as the social origin of a voter was known, it was possible to determine their vote choice. One of the assumptions in the sociological theory is that politics is based on social (or political) cleavages. This concept was first presented by Lipset and Rokkan in their well-known work Party Systems and Voter Alignments (1967), in their attempt at understanding the development of the European party system. By looking at historical events such as the industrial revolution, they saw how a new political structure, which consisted of several conflicting interests, created the basis for different political and social cleavages. The classical division between center and periphery, religious and secular, and the economic right-left was long used to explain voting choice. These geographical, cultural and economic cleavages were for a long time the basis for our understanding of voter behavior within the first Norwegian electoral studies (Aardal & Bergh, 2015, p. 247).

However, as Western democracies became post-industrial societies, the traditional class structure, that had been used to explain party choice, dissolved. There was no longer a sharp division between the middle class and the working class, and the interests of the emerging “new middle-class” were difficult to position. The impact of class cues has generally decreased since the mid-twentieth century (Knutsen & Thrasher, 2007) which meant that the sociological group-based explanation model was no longer as sufficient. Also, the concept of crossing cleavages meant that a person who is concerned about the environment could at the same time support a solid social welfare system. It should be added that even if the relationship between voters’ social background and party choice has weakened, that does not mean that all socio-structural elements are irrelevant (Aardal & Bergh, 2015, p. 247). After

6 modernizing and adjusting the concepts of class and cleavages, later research still finds some geographical and “class-oriented“ patterns in the support of different parties (Berglund, 2007, p. 169). Sector and gender have been found to have increased importance on vote choice and is recognized as some of the “new” cleavages (Berglund, 2007, p. 169).

2.1.2 The Michigan Model In The American Voter (1960), one of the most influential books in political science, Campbell et al. presented what is now known as the Michigan model. In their socio- psychological approach they describe the voting process as a funnel of causality. At the width of the funnel are the socioeconomic conditions that generate the basic conflicts of interests within society: economic structure, social divisions (such as race, ethnicity, gender and sexual orientation) and historical patterns (Campbell, 1960, p. 24). These factors do not have a direct impact on the vote, but they influence which social group we identify with, and our value orientations, which again shape our political attitudes. Campbell et al. explain voting behavior with three attitudes: party attachment, issue opinions and candidate images. They argue that these attitudes have a very strong impact on the final vote (Dalton, 2014, p. 185). There are also other external factors that could matter; factors such as media reports, campaign activity, and economic and political conditions. These are also considered to have an effect on the voter´s issue opinion and candidate image (Dalton, 2014, p. 185). In sum, the model differentiates between long-term factors such as social class, group affiliation, ideology and party identification, and short-term factors such as characteristics of the candidates, political issues and the incumbent parties’ efforts. In the model party identification is the factor that is especially emphasized, which is why the model is also known as the party identification model. Party identification is an early-socialized, enduring, and affective psychological identification with a specific political party. Campbell et al. (1960) claimed that people tend to attribute their votes to long-established family traditions, or positions they had held across many elections. Perhaps the most famous example is the saying that used to be applied to voters in the American South: “they would vote for anyone, even a yellow dog, if they were a Democrat” (Dalton, 2016, p. 2). The theory has gained a great deal of recognition, and is the base for one of the most extraordinary data collection efforts in the social sciences; the American National Election Studies (ANES). The model is also foundation for multiple voting models used today (e.g. in Scandinavia), and has had a great impact overall on the way we study voters and voter behavior.

7 However influential, the model has also received criticism, and there are few who employ the model in its original form today. Many argue that the relationship between party identification on one side and voting choice on the other was so strong that its explanatory power becomes minimal. The Swedish election researcher, Søren Holmberg, has argued that the connection between party identification and voting is “too close to comfort” (Holmberg, 1994, p. 96).

2.1.3 Norwegian election studies From the beginning of the Norwegian National Election Survey (NNES), with its first survey conducted in 1957, the electorates socio-structural position has been a core element both theoretically and empirically (Aardal & Bergh, 2015, p. 247). The geographical, cultural and economic cleavages, first introduced by Lipset and Rokkan in 1967, have played a central part of how we understand voter behavior in Norway. However, the structural changes that found place in most western European countries did also have an impact on the social structure in Norway. In later years, the election studies have focused more on the electorate’s subjective perceptions, partly fundamental political orientations and partly personal stance on current political issues (Aardal & Bergh, 2015, p. 248). It is not enough to look at your fathers, or your own profession, to predict voting choice. Today, the model used to explain voting behavior contains several explanatory factors such as attitudes towards political issues and parties and politicians, fundamental ideological attitudes and values, and voter’s assessment of the economy. As I proposed initially, it would be reasonable to assume that these explanatory factors might not be as successful when explaining voting behavior amongst first- time voters. Political issues like taxation, pension, child-support, infrastructure, size of the welfare state etc., are not something that most 18-year-olds have had to take a stand on yet. Looking at this group of voters in a different, more process-oriented/decision-maker perspective could contribute to enhance our understanding of how first-time voters manage to decide on their vote, which is this thesis’ main goal. Even if this type of study is unique within a Norwegian context, there are others that has studied voter behavior from this perspective before. In the following section, a review of some of the most relevant studies.

8 2.2 Lau and Redlawsk’ “New Theory of Voter Decision Making”

As mentioned in the introduction, efforts have been made towards a more process-oriented framework for understanding voters’ decision-making. Most relevant for this thesis is the work by Lau and Redlawsk (2006) and the updated version by Lau et al. (2018). Lau and Redlawsk wanted to get “inside the heads” of the decision makers, the voters. Through experiments and process-tracing, they studied how voters make their decisions. They propose four broad theoretically defined types of decision strategies that voters employ to help decide which party or candidate to vote for (a more detailed review on these strategies will be accounted for in section 2.5). Which strategy a voter would employ depends, according to Lau and Redlawsk, on several factors. These factors are illustrated in Figure 2.1. In their proposed framework, they look at how the (1)traditional antecedents of the vote - characteristics such as ideology, partisanship, economic status and other personal characteristics, along with (2)political experience/sophistication and (3)campaign factors affect, not vote choice directly as is the traditional focus, but how a decision task is subjectively perceived, what they call (4)“nature” of decision task. How a voter perceives the decision task of voting (i.e. the level of importance and/or complexity associated with the task) will in turn influence the (5)information processing, (how much and what type of information that is acquired). Finally, the information process together with our (7)memory, which is where information is stored, and where the information we base our decision on is drawn from, will influence our evaluation, our vote choice and the quality of our decision (Lau & Redlawsk, 2006, p. 40).

9 Figure 2.1. Process-oriented framework for studying voter decision making

(2)Political Sophistication

(6)Memory

(7)Evaluation, Vote Choice, (1)Background (4)«Nature» Decision Quality Characteristics of Decision of Voter Task

(5)Information Processing

(3)Campaign Factors

Source: Lau and Redlawsk 2006:22.

The effect of the factors (2)political expertise and (3)campaign factors which psychologists would call task demands (Lau & Redlawsk, 2006), will be discussed in more detail later in the this chapter, but the notion in short is, first, that one’s level of expertise within a field, which in this case is politics, will affect what and how much information that is gathered and how it is processed. The second factor, campaign factors, is the situational or contextual nature of the campaign in itself (Lau & Redlawsk, 2006, p. 38). For instance how the political system is structured in a state or country will often determine how many parties or candidates that run for government, meaning how many alternatives and potential outcomes there are. The more alternatives and outcomes one has to consider, the more complex the decision task.

There are, as we have seen, many potential factors that influence how a voter will go about making their vote decision. However, I will now turn the focus on the (5)information processing factor, which is the factor that is the most unique about Lau and Redlawsk framework. Thus, in the following paragraphs I will start with a somewhat detailed review on this study’s basic assumptions; the view that we as decision-makers are cognitively limited information processors, followed by a discussion on what mechanisms we employ to overcome these limitations. These assumptions are drawn from behavioral decision theory,

10 the insights of which form the core premises of the process-oriented framework and the decision strategies developed by Lau and Redlawsk. Finally, an account of the five decision strategies found by Lau and Redlawsk (2006) and Lau et.al (2018).

2.3 Behavioral Decision Theory and bounded rationality

How do human beings reason when the conditions for rationality postulated by the model of neoclassical economics are not met? - Simon (1989, p. 377).

To get a complete overview of an election would require great amount of cognitive capacity. One would have to figure out the stance of all the different parties, but also how the parties will prioritize if elected, and what the issues at stake are in that particular election. Moreover, one would have to consider what issues one cares about, and then evaluate all the potential outcomes. This would entail a serious amount of time and cognitive effort, and the task of finding the most value-maximizing alternative would for most not be worth it, if even possible.

Studies of real-world decision making has shown that we more often than not make decisions based on incomplete information and without evaluating all potential outcomes. People tend to settle with alternatives that are considered “good enough”, rather than focusing on maximizing outcomes (Lau & Redlawsk, 2006, p. 22). This notion refers to the concept of bounded rationality, a concept coined by the Nobel prize winning social scientist Herbert Simon in 19573, which states that human judgements are limited by available information, time constraint and cognitive limitations. He proposed the concept as an alternative basis to the stricter rational choice theory, which up until then had been the dominant theory used in decision-making research. Bounded rationality is the fundamental theory assumed in behavioral decision theory which is what Lau and Redlawsk use as point of departure in their study. Next, a short introduction to human cognition and its limits will be presented. This will hopefully provide useful as a backdrop to the later, more extensive discussion on cognitive coping mechanisms and decision strategies.

3 The concept presumably first appeared in the book “Models of man, social and rational” written by Simon (1957).

11 2.3.1 Our cognitive structure and its limits As Anderson argues in his seminal work from 1983, the architecture of human cognition creates very real limits, which cannot be escaped no matter how motivated a decision maker might be (Anderson, 1983). Everyday our senses are being bombarded with stimuli, only a fraction of this, however, is processed and becomes part of our short-term memory (STM). Of the millions of bits of information that our senses are capable of processing in a brief period of time, our STM can maintain only 7 (+/-) chunks of information at any given time (Miller, 1956). Even if the unit of measure “chunks” is ambiguous, the main point is that the amount of information that most people can receive, process and remember is severely limited. For information to become part of our long-term memory, perceptions have to be activated strongly or repeatedly rehearsed. According to the Decomposition Thesis, only learning that takes place over 100 hours can be decomposed into learning events and eventually become part of our long-term memory or “knowledge bank” (Anderson, 2002, p. 86). This part of our memory is constructed as an associative network. Our memories are linked together in such a way so that when we perceive external stimuli, memories or nodes that are linked to it have a potential to become activated as well and thus become available in our STM (Lau & Redlawsk, 2006, p. 24). However, only small portions of what is activated can be processed, and only one item at a time, not simultaneously, which presents another significant cognitive limitation (Lau & Redlawsk, 2006, p. 24).

Drawing on what we now know about human cognition makes the complexity of decision - making, and especially vote decision making, much more evident. In order to make a somewhat satisfying decision, one would have to pay attention and perceive information that activates the short-term memory, then attend to it in such a way that it becomes part of the long-term memory, and finally one would have to be able to recall all of the relevant information from one’s memory when needed. This process is both time consuming and requires high cognitive skills. Add the fact that elections for many are not considered a life- changing event, it would be reasonable to presume that most voters do probably not put in the cognitive effort necessary to support the process assumed by many prior models of the vote choice (Lau & Redlawsk, 2006, p. 24). How are decisions then made?

12 Skepticism about voters’ cognitive abilities and knowledge has been expressed before, and the question whether voters are considered “sophisticated”4 or not has been debated for decades. Already in 1964, Converse in his well-known work The Nature of Belief Systems in Mass Publics reported on voters’ lack of political knowledge and ideological thinking. He had asked the respondents to explain the terms “liberal” and “conservative”, and found that many struggled to link those concepts to either political parties or other ideological concepts. He also found that, contrary to the political elite, the issue preferences amongst the mass public were inconsistent and showed low constraints, and the attitudes were highly unstable over time (Converse, 2006). This fact is perhaps more understandable now after learning what is required of cognitive effort for something to become part of our knowledge, especially if you are not very interested in the topic in the first place. Years later, Zaller (1992) argued in his famous work The Nature and the Origin of Mass Opinion that most voters do not hold single political preferences. Public opinion is in fact heavily influenced by exposure to elite discourse on political matters. He argued that voters do not derive policy preferences from a structured coherent belief system, rather they receive political “messages” from the elites, and form attitudes based on how these messages fits with prior beliefs and awareness in general (Zaller, 1992). This finding could actually be explained by behavioral decision theory; people employ heuristics to simplify the decision-making process. In Zaller’s study, the mass uses the elites and their opinions as a tool for making their own.

There is a wide range of approaches within the literature of decision theory on how people cope with our inescapable cognitive limitations, some more extensive than others. I found the research and the concepts developed by Kahneman and Tversky to be the most applicable for my study. There are of course more fine-meshed psychological theories which give very detailed and complex accounts of hundreds of cognitive mechanisms, but as this study is written within the field of social science, I chose to limit the usage of such psychological concepts, including only those I considered fruitful for the purpose of this analysis.

4 Sophisticated voters, according to Converse, are those who structure their opinions in a larger ideological framework. This level of political sophistication is correlated positively with the respondent's level of education, degree of political involvement, and amount of political information.

13 2.4 How do we cope with our inescapable cognitive limitations?

2.4.1 Judgement heuristics People have developed a large number of cognitive mechanisms in order to maneuver within a world of information overload, time constraint, and cognitive limits. One of these mechanisms, as introduced earlier, is heuristics. Heuristics are problem-solving strategies, often employed automatically or unconsciously, and represent mental shortcuts which simplifies the process of making judgements or inferences (Lau & Redlawsk, 2006, p. 24). In the early 1970s, Tversky and Kahneman identified three general cognitive heuristics that decision makers utilize when making judgements. The first is representativeness – which is when we assign specific instances to broader categories according to how well that particular instance fit the essential properties of one category (Lau & Redlawsk, 2006, p. 26; Tversky & Kahneman, 1974). An instance or a person is judged within a preexisting scheme, which has been developed through one’s prior life experiences. The second heuristic is availability, which is useful when assessing frequency or probability, where instances of larger classes are usually better and faster recalled than instances of less frequent classes (Tversky & Kahneman, 1974, p. 1127). If you are considering whether you should buy a bus ticket or not (you are only taking it two stops), you try remembering how many instances of ticket control you have witnessed; if there are many, you buy a ticket, if there a few, you don’t. However, this heuristic is often affected by other factors than frequency and probability, factors such as familiarity and salience, which consequently can lead to judgement biases (Tversky & Kahneman, 1974, p. 1127). Salience, for example, affects the retrievability of instances, which means that the impact of a personal experience (e.g. being caught sneaking on the bus last week) on the subjective probability of such accidents is probably greater than the impact of reading or hearing about that same experience (Tversky & Kahneman, 1974, p. 1127). The third heuristic is adjusting and anchoring, which is when people start with an implicitly suggested reference point, and then make incremental adjustments based on additional information (Tversky & Kahneman, 1974, p. 1128). Biases in this case can occur when we make judgments based on the “wrong” anchor.

These heuristics simplifies the process of making inferences and judgments, not only in our day to day life, but also in situations that initially would require great cognitive efforts, such as elections. When we evaluate parties, party leaders and the issues at stake, most of us do this based on some preexisting values, or notion of political belonging, even if we don’t have a specific party affiliation. The anchoring and adjustment heuristics could be created, more or

14 less unconsciously, by several factors such as the sociopolitical environment one is brought up in, or the influence of friends and family. This initial “anchor” creates the basis on which the other heuristics, representativeness and availability, are created. Growing up in a working- class family or amongst the bourgeoisie would probably affect one’s outlook on political issues such as redistribution and taxation, where the former would argue that everyone should have equal possibilities, while the latter would argue the unfairness of being forced to give some of their earnings away. One’s initial values, or anchor, will then “color” one’s continuous evaluation of persons and instances, in other words what heuristics we form and employ. This is how Lau and Redlawsk in their framework (figure2.1) explain the relationship between the traditional socio-structural factors and the psychological factors; information processing and memory. The social and political characteristics of our upbringing influence what mental short-cuts we establish and employ, what information we process and what we remember.

2.4.2 Simplifying mechanisms for making choices Research on decision-making distinguish between judgement and choice, whereas the former requires an explicit evaluation of each alternative, the latter in contrast, requires only that one alternative be selected and the rest rejected (Billings & Scherer, 1988, p. 2). Lau and Redlawsk identify three fundamental ways decision- makers simplify the task of making choices. The first is decomposition, which is when you break down a task into its composing parts, each of which are presumably easier to evaluate than the entire decision (Lau & Redlawsk, 2006, p. 27). The second strategy is called editing, or more precisely pruning. You simplify the decision by eliminating or ignoring otherwise relevant aspects of the decision. (Lau & Redlawsk, 2006, p. 27). In an election setting, this could for example entail concentrating only on the major parties which would effectively eliminate several alternatives from one’s choice set. On the other hand, if you are passionate about only one issue, single- issue voter, you would only have to consider the parties that include that particular topic in their party program. The third is decision heuristics, which are similar to the judgement heuristics described above, but are more directly concerned with the task of simplifying choice, not judgements. A heuristic in this sense is a strategy that ignores part of the information, with the goal of making decisions more quickly, frugally, and/or accurately than more complex methods (Gigerenzer & Gaissmaier, 2011, p. 454).

Focusing on vote choice, Lau and Redlawsk describe five heuristics/cognitive short-cuts that people commonly employ when deciding on their vote. Affect Referral is the first; if you

15 already are familiar with the different parties, vote for the party you like the most. The second is called Endorsements, which is when you decide your vote based on cues from close acquaintances or from e.g. a political elite or a close friend whose values you believe correspond with your own. You leave the hard work of figuring out who to vote for to someone else. The third is Familiarity; you simply vote for the party that you are familiar with, assuming that their policy is somewhat tolerable. This strategy is similar to Tversky and Kahneman’s judgment heuristic availability. Chose the alternative that first comes to mind when presented with a decision task. The forth is Habit. You vote for the same party you voted for last time. The fifth is Viability; You only consider those parties that actually have a fair chance of winning (Lau & Redlawsk, 2006, pp. 27-30).

All of these simplification mechanisms; judgement and decision heuristics and decomposition and pruning, all help to solve the problems connected to our bounded rationality. However practical these cognitive shorts-cut are, it can sometimes affect the quality of our decisions and subsequently lead to unfortunate outcomes. A professor at the University of Oslo, Raino Malnes, told the university newspaper that he considered refraining from voting one year because he noticed that he was too caught up in a local issue regarding a potential shutdown of his daughters’ school. Politics, he argued, is something bigger than the school of his daughter. He did not want to risk voting for a party who indeed opposed the shutdown, but in other areas supported politics he did not stand for (Universitas, 2015). Even if the normative aspects of single-issue voting could be discussed, the message is that not all heuristics lead to decisions we in the aftermath of the decision process is satisfied with. We are, according to Lau and Redlawsk, potentially faced with a real dilemma coping with these cognitive limits:

On the one hand, because we are not omniscient calculators, we simply need to use some cognitive shortcuts, some means of simplifying decisions so that a choice can be made. But on the other hand, whatever shortcuts or simplifications we adopt come with a potential cost: inaccurate judgements and something short of value-maximizing decisions. -Lau and Redlawsk (2006).

Behavioral decision theory research suggest that this dilemma occurs because people generally have two competing goals when making a decision: (1) the desire to make a good decision and (2) the desire to reach a decision with minimal cognitive effort (Lau & Redlawsk, 2006, p. 29). However, the data that I have collected is not meant and con not be used to say anything normative regarding whether some heuristics or strategies leads to better

16 decisions than others. My focus will mainly be on what type of cognitive decision strategies people voting for the first time employ. Based on the theories just presented, two main conclusions can be drawn: we as decision makers have cognitive limits, but we utilize cognitive shortcuts to mitigate the effects of those limits. In the next section, a definition of decision-making styles (cognitive styles), and a review of the five decision strategies discovered by Lau and Redlawsk is presented.

2.4.3 Decision-making strategies and cognitive styles in general Decision strategies, or decision-making styles, can be defined as “a set of mental and physical orientations that an individual use to reach a decision” (Lau & Redlawsk, 2006, p. 30) or individuals’ characteristic mode of perceiving and responding to decision-making tasks (Harren, 1979, p. 119). They constitute a subset of broader cognitive styles, generally defined as the way people deploy their intellectual abilities or the manner in which they approach cognitive tasks (Galotti et al., 2006, p. 630). Cognitive style, a concept carefully reviewed by Kozhevnikov in her 2007 article, refer to consistencies in an individual’s manner of cognitive functions, particularly with respect to acquiring and processing information. There are fairly stable individual differences in the way people perceive, think, solve problems, learn, and relate to others (Kozhevnikov, 2007, p. 464). Such cognitive styles have, according to research reviewed by Kozhevnikov, proved better predictors of success than many other more obvious factors such as intelligence (Kozhevnikov, 2007, p. 464).

The study of decision-making and the concept of cognitive styles is not new. For over seventy years researchers, especially within the field of psychology, economics, and marketing, have studied and developed various versions of cognitive styles. The first experimental studies revealing the existence of individual differences in simple cognitive tasks involving perception and categorization was conducted already in the 1940s and early 1950s (Hanfmann, 1941; G. Klein & Schlesinger, 1951; G. S. Klein & Krech, 1951; Witkin, 1950). These studies though, have commonly been limited to self-report measures of general decision-making, unconnected from specific decisions made (Galotti et al., 2006, p. 630). People do, however, adopt different strategies depending on several aspects of the decision task, also known as the adaptive decision maker hypothesis (Payne, 1993). For example, what strategy one employs could depend on the level of conflict that arise when confronted with a decision (Lau & Redlawsk, 2006, p. 30). If you find yourself in a situation where you preferer one alternative over all other alternatives, the decision process is easy and without conflict. If

17 you however prefer one alternative on one dimension, but another alternative on another dimension, the level of mental conflict will rise. Voting, for most voters, essentially involve some type of conflict. The decision strategies identified by Lau and Redlawsk (2006) and Lau et al. (2018) have been developed particularly for the sake of studying decision-making processes amongst voters, which is why their strategies will be the starting point and the foundation of this study’s analysis. In the next section, a description of the five voter decision strategies identified by Lau and Redlawsk is presented, followed by a review of other factors that are argued to have an impact on information processing.

2.5 Lau and Redlawsk’ five voter decision strategies

Building on the theories presented above, about our limited cognition and how we overcome them utilizing tools such as judgment and decision heuristics, and on findings from conducted surveys and experiments studying voters’ decision processes, Lau and Redlawsk propose four broad voter decision strategies, or models as they call it. A fifth style has later been suggested by Lau et al. (2018) in their newest article. Initially, they distinguish these models, or strategies which I will refer to them as of now, into two very broad categories, which the more specific strategies fall under. According to Lau and Redlawsk, decision-making in situations containing conflict, such as in an electoral setting, can be made in two different ways; one is called a compensatory strategy, and the other a non-compensatory strategy. Compensatory strategies are “cognitively complex information integration rules where decision makers are assumed to assign value to every salient attribute associated with each alternative” (Lau & Redlawsk, 2006, p. 31). A positive value on one dimension can compensate for a negative value on another dimension, as long as they are commensurable. Thus, conflict is confronted by integrating the positive and the negative information or values associated with a choice. Noncompensatory strategies, on the other hand, rely on incomplete information search to avoid conflict (Lau & Redlawsk, 2006, p. 31). In this case, negative values or outcomes do not trade-off against positive values or outcomes on another dimension, because alternatives are usually eliminated once negative information about them is obtained (Lau & Redlawsk, 2006, p. 31). Research has found that people often utilize noncompensatory strategies and try to avoid value trade-offs because it is more time- consuming and often more cognitively demanding (Lau & Redlawsk, 2006, p. 31).

18 Another dimension on which these five strategies are categorized by is information processing. One important finding from much behavioral decision theory research is that individuals’ pattern of information acquisition reflects distinguishable decision strategies (Lau & Redlawsk, 2006, p. 32). Lau and Redlawsk’ models therefore differ on two main dimensions; (1) Depth of search, i.e. how much of the available information is gathered and considered before a decision is made, and (2) Comparability of search, which is how evenly that search is distributed across alternatives (Lau et al., 2018, p. 327). Based on these broad dimensions, Lau and Redlawsk propose four (later five) decision strategies, or models as they call it.

The first decision strategy (Model 1) is “Classic Rational Choice”, which is a compensatory strategy based on assumptions from rational choice theory. Voters employing this decision strategy actively gather as much information as possible, about every candidate or party on the ballot. The information search is in other words deep and highly comparable. By evaluating the candidates in terms of expected consequences for one’s own self-interest of each of them winning the election, one has a high probability of finding a value maximizing outcome (Lau & Redlawsk, 2006, p. 6). There is however, a differentiation within this model between what is called constrained and unconstrained rational choice. The constrained rational choice model is just as rational in its orientation as the pure unconstrained one but has incorporated elements about our cognitive limitations. More information is always considered to be better than less information, but “constrained” rationality realizes that the cost of gathering all that information may sometimes exceed the marginal benefit from having it (Lau & Redlawsk, 2006, p. 9). To the extent that there are people utilizing a decision strategy as the Classic Rational Choice model, Lau and Redlawsk expect that the constrained one is the most commonly employed.

The second decision strategy (Model 2) was initially called “Early Socialization and Cognitive Consistency” (Lau & Redlawsk, 2006, p. 9) but was later renamed “Confirmatory” decision strategy by Lau et al. 2018. This model is based on principles from the Michigan/ party identification model described earlier in this chapter. Voters makes decisions based on early affective based socialization toward or against prominent symbols such as political parties, and motivation to maintain cognitive consistency with early learned affect (Lau et al., 2018, p. 328). Party identification becomes the “lens” through which political information is selectively perceived, and the information seeking is often biased in favor of the “in-party” candidate or party, which translates to shallow to deep- but clearly unequal- information

19 search (Lau et al., 2018, p. 328). Whereas Model 1 decision making is based on explicit calculations of self-interest and value -maximizing outcomes, Model 2 decisions are “strongly influenced by early-learned social identifications, which like all of such identifications, tend to be accepted with little or no consideration of alternatives” (Lau & Redlawsk, 2006, p. 10).

The third strategy (Model 3) is “Fast and Frugal decision-making”. Voters are motivated primarily by efficiency, actively seeking only the most diagnostic information, which will allow them to make the correct choice quickly. The information-gathering process is limited to the one or two most important/diagnostic criteria, but those few criteria should be evenhandedly applied to every alternative in the choice set- that is shallow but comparable information search (Gigerenzer & Goldstein, 1996; Gigerenzer & Todd, 1999; Goldstein & Gigerenzer, 2009). Voters employing this type of strategy are within the political science field labeled single-issue voters. There is a distinction between “easy”-issue voters, and “hard”- issue voters. Where the former are characterized by issues that are more symbolic rather than technical, and dealing with policy ends rather than policy means (Lau & Redlawsk, 2006, p. 13). The latter, hard issues are more technical, dealing with the means to achieve a universally valued end (Lau & Redlawsk, 2006, p. 13). Easy issue voting does not require much “sophistication”, whereas voting on hard issue are more similar to the model 1 strategy.

The fourth, “Heuristic-Based decision-making” (Model 4), is a strategy that views voters as cognitively limited information processors who are generally motivated to make good, though not the best, decisions, as easily as possible. The information search is generally quite limited; various cognitive shortcuts and heuristics are heavily utilized, particularly those that avoid having to make cognitively difficult value trade-offs. Non-comparable search across alternatives. Lau and Redlawsk explain that “this approach argues that most decisions (including political decisions) are better understood as semiautomatic responses to frequently encountered situations than as a carefully weighted probabilistic calculations of the consequences associated with the different alternatives” (Lau & Redlawsk, 2006).

As mentioned above, Lau et.al (2018) propose a fifth decision-making style. The “Gut decision-making”-strategy (Model 5) involves gut-feeling and intuition. Decisions are made strictly in affect and usually unconscious (Lau et al., 2018, p. 328). There is no deliberate external search for information, nor any effort trying to compare alternatives. Thus, the information search can be categorized as shallow and non-comparable. Traditionally, intuition have been viewed as frequently biased, leading to systematic and

20 predictable errors, thus considered risky to base decisions upon (Tversky & Kahneman, 1974, p. 1124). However, newer empirical research show that there are situations, involving non- decomposable tasks5, where decisions based on “gut-feeling” exceeds the effectiveness of analytical decision-making. According to Dane, Rockmann, and Pratt (2012), the effectiveness of intuitive decision- making relative to decisions based on thorough analysis depends on one’s level of domain expertise (Dane et al., 2012, p. 193). In addition, gut- decision making often provides decisions that the decision-makers feel good about.

Model 1, 2, 3 and 4 was theoretically defined by Lau and Redlawsk in their study from 2006. Years later, based on the findings from this first study, Lau and his colleagues developed a 13-item battery, which they tested on six different samples, some of them experiments (mock elections) and some national representative surveys. After running a confirmatory factor analysis, they did find support for the four hypothesized factors (strategies), but in addition they identified a fifth, the gut-decision making. However, they are open for the possibility that there might be even more than five decision-strategies, which motivates the continuation of this type of study, especially within a new context. In table 2.1, an overview of the characteristics of the five voter decision strategies.

5 Research has shown that tasks that can be decomposed, such as math or logic, are better solved by analytically executing a sequence of procedures (Sadler-Smith and Sparrow 2008:317). When a task is non-decomposable however (e.g. judging the quality of artwork, the taste of food etc.) intuition can prove effective because there is no definitive solution to the problem, nor any clear rules of how to approach the task (Dane et al. 2012:188).

21

Table 2.1. Five strategies of individual decision making *

Compensatory Noncompensatory Model 1 Model 2 Model 3 Model 4 Model 5 Classic Rational Confirmatory Fast and Frugal Heuristic-Based Gut decision-making Choice Assumptions Actively seeks Party identification is sought Seeks out only a few Seeks out only Information search is about information out as much out early and become the attributes of judgement enough information strictly affective, search information as “lens” through which which they care about, to allow them to usually unconscious, possible, about political information is or the most diagnostic reach a decision. and involves no every available selectively perceived. information that will Cognitive shortcuts deliberate external alternative (until Information search is often allow them to quickly and various decision searching for the cost exceeds more passive than active, make the right heuristics are information. the benefit). depending on the decision, and ignore heavily, and almost circumstances of the everything else automatically, decision task. utilized Depth of search Deep and Shallow to deep, but Relatively shallow but Shallow and Shallow and and comparable unequal search comparable search noncomparable noncomparable search comparability of search search serach Method of Explicit, Memory-based evaluations Explicit memory- Satisficing or related Alternatives are decision making conscious, of what is known (long- based consideration of methods which considered based on cognitively term) and has recently been the one or two positive attempt to make how they make you difficult memory- learned (short-term) about and negative choice relatively feel. Going with your based the different alternatives. consequences easy by restricting gut. consideration of associated with each information search. positive and alternative. negative consequesces of each alternative. Motivations for Self-interest Cognitive Consistency Efficiency Making the best Motivated reasoning Choice possible decision with the least amount of effort; Avoiding value tradeoffs. Electoral Inputs Mainly Primarily party Candidates´ ”stands” Cognitive shortcuts No electoral input to Decision retrospective identification, but also issue on the few attributes a (stereotypes, (e.g., job stands, economic voter considers (but schemas, etc.) and performance) and evaluations, perceptions of certainly not limited to other political prospective (e.g., the candidates, and policy stands) heuristics. issue stands) evaluations of the judgement about incumbent´s job the candidates. performance.

*The table incorporates information from Lau and Redlawsk (2006:8 and 36) figures, and Lau et al. (2018:328-329).

22 2.6 What other factors can affect the decision strategy?

So far, the focus has been on our cognitive limitations and how we, when confronted with decision tasks, manage to overcome these limitations by employing simplifying decision strategies. Thus, the attention has been on the (5) information processing factor (from Figure 2.1). I will now take a step back and discuss the two antecedent factors that Lau and Redlawsk in their framework argue affect information processing directly. How information is processed and what decision strategy that is employed is influenced by the subjective (4)“nature” of the decision task, but Lau and Redlawsk argues that it is also directly affected by (2)political sophistication and (3)campaign factors. They say that the level of complexity associated with a decision task, how cognitively demanding it is, will vary according to level of expertise within the field (political sophistication), and the situational or contextual nature of the campaign itself (campaign factors )(Lau & Redlawsk, 2006, p. 38). In the following section, I will take a closer look on how these factors can affect decision strategy, and how the characteristics of the political context in Norway might influence the outcome of these different effects.

Political expertise According to previous studies on decision making, the amount of knowledge or experience that an individual has in a particular domain is the only individual difference which is found to affect information processing in decision making (Lau & Redlawsk, 2006, p. 37). One might think that those who are considered political experts would spend more time gathering information, carefully considering and evaluating all relevant aspects before making a choice, compared to non-experts. However, this is often not the case (Lau & Redlawsk, 2006, p. 37). In a study of chess players Chase and Simon (1973) found that there was no difference between experts players and novices in their thought processes; the number of moves considered, search heuristics and depth of search (Chase & Simon, 1973, p. 55). The experts search through the same, or even fewer, number of possibilities as weak players, however they were very good at coming up with the “right” moves, whereas the weaker players spend considerable time analyzing the consequences of bad moves (Chase & Simon, 1973, p. 55). This can be explained by the fact that experts, through multiple experiences within their domain have developed cognitive schemas that allow them to make reasonable inferences without paying attention to all the details; they learn to pay particular attention to specific heuristic information which makes them able to discriminate between relevant and irrelevant cues (Shanteau, 1992, p. 75). The more experienced the decision maker, the easier the

23 decision and vica versa (Lau & Redlawsk, 2006, p. 40). Thus, if we can transfer this logic to voting, political experts do not spend more time on information gathering and thorough evaluation than rest of the electorate, they are just better at collecting relevant information, and only enough to make a satisfying choice.

However, how this factor will play out in my study is not certain, because it will be a question of how political experts are defined. Lau and Redlawsk included in their experiment questions that measured the number of campaign-related and community-oriented political behaviors, a stringent political knowledge test and measures of following politics in the media, frequency of talking about politics and self-proclaimed interest in politics (Lau & Redlawsk, 2006, p. 120). They combined these indicators into an overall measure of political expertise/sophistication. Because my first priority was to get the respondents to give honest and forthright accounts of their vote choice decision process, I chose to exclude questions asking explicitly about political knowledge. I was afraid that if the respondents felt incompetent, not being able to answer the questions, this might affect their answers on the other questions, perhaps leading them to say that they spent more time on researching than they actually did to compensate for their lack of knowledge. Nonetheless, I do have some questions which could be used as an indirect measure of political expertise. A more detailed account of this is presented in the next chapter. This will not be equivalent to the measure used by Lau and Redlawsk, but it can give us some indication on the effects of at least political awareness, if not expertise, on information processing and decision strategies.

Campaign factors/ task demand The second factor that is argued to have an impact on information processing and decision strategies, is “campaign factors”, or “task demands” as psychologists call it (Lau & Redlawsk, 2006, p. 38). One of the aspects of campaign factors which is thought to have an influence is the complexity of the task. Complexity is here understood as how many alternatives that are under consideration times the number of different attributes across which they vary (Lau & Redlawsk, 2006, p. 38). In Norway we have a multiparty system. More than twenty different parties ran for parliament in the 2017 general election, and nine of them obtained representation in the parliament. Because of this system, one party is rarely able to obtain a majority and form a government alone, meaning that most parties need coalition partners. The different constellation parties forming a coalition, who supports whom, has proved to be crucial for the outcome of an election, and more than once the achievements of some of the smallest parties have been decisive for the final result. For instance, the getting

24 over the electoral threshold in the 2017 election was crucial for the forming of a right-wing government. Thus, a vote for one party could in effect be a vote for several other parties as well. “Tactical voting” is a relevant concept within the Norwegian context. Moreover, the potential coalitions are not always explicitly stated before the Election Day and are sometimes decided after the election, making it even more difficult to distinguish between the alternatives. It seems reasonable to say that the level of complexity posed in Norwegian elections are relatively high compared to the United States presidential elections where the alternatives often are limited to only two candidates. According to Tversky (1972), decision makers who are confronted with more than two alternatives rely more heavily on simplifying decision heuristics, and employ more noncompensatory strategies, whereas those who are given only two options tend to employ compensatory strategies. Based on this we would expect that Norwegian voters who are forced to decide within a multiparty system, in general employ noncompensatory strategies to a larger extent than those who vote in presidential elections. However, there will of course be a distinction between the actual alternatives and the subjectively perceived alternatives. Even if there exist many alternatives, this does not mean that a voter will consider them as alternatives. I would guess that many have a pre- existing idea of which parties they consider as potential candidates for their vote. Which these are could be determined by for example the social context of one’s upbringing.

Another aspect that is thought to affect the difficulty of choice making is the similarity of the alternatives (Lau & Redlawsk, 2006, p. 39). If the alternatives are very dissimilar, it is easy to distinguish between them, and this means that just a simple decision strategy is enough to make a satisfying choice. If the alternatives are very similar, however, the task of finding the best option becomes much more difficult. Research on decision-making behavior has shown that in these cases depth of search increases, in addition to the employment of a compensatory decision strategy (Biggs et al.1985:970). As mentioned, there are multiple parties that run for government in Norway, and it can be a challenging task to distinguish between the policies of the different parties. According to the latest Norwegian election study, when asked “how big of a difference do you think there is between the parties” 53 percent said “not very big” or “very little”, while only 8 percent said “very big” (and 39 percent said “quite big”)6. When over half of the respondents/electorate consider the parties to be that similar, this would suggest that there will be many who employ a compensatory decision strategy.

6 I used the data from the latest NNES to run a frequency on the question “Hvor stor forskjell synes du at det er mellom partiene på Stortinget? (Q42). Dataset available from www.nsd.no

25 2.7 Summary

I began this chapter by reviewing some of the most prominent theories on voting, namely the sociological model and the Michigan/party identification model. These theories have for decades been the base of many of the biggest election studies, and a lot of the information we have about voters and voter behavior is due to these early studies. However influential, today there are few who uses these classical theories in their initial form, although many elements are still relevant. Further, I gave an account of what the Norwegian Election Studies emphasizes when studying voters, and I argued that using voter’s attitudes and values on different political issues, and their assessment of the government and the economy might not provide as useful when looking at first-time voters, or those voters who do not have clear stands or views on these political topics. Hence, I argued that by approaching this group of voters in a more process-oriented way, we could obtain a more comprehensive understanding of how vote decisions are made. I then presented the framework proposed by Lau and Redlawsk. In their framework they combine elements from the traditional voting theories with assumptions from behavioral decision theory and modern psychology. They argue that the traditional antecedents of the vote, such as the demographic background characteristic of a voter affect not vote choice directly, as traditionally assumed, but which decision strategy a voter employ when faced with a decision task. In order to understand the concept of when and why decision strategies are used and how they are created, I gave an account of the theoretical fundament Lau and Redlawsk use as their point of departure. Their fundament is based on assumptions about bounded rationality, humans as limited information processors and what coping mechanisms we as decision-makers employ to overcome our inescapable cognitive limits. After this, I reviewed the five decision strategies proposed by Lau and Redlawsk followed by a discussion on how the factors, political sophistication and campaign factors, are hypothesizes to affect decision strategy within a Norwegian context. In the next chapter, I will go through the research process, and the method I have used trying to answer the thesis question; what decision strategies do first-time voter in Norway employ when deciding on their vote.

26 3 Method -Research design

3.1 Data

3.1.1 Population: First-time voters This study’s population is first-time voters attending high school in Norway. First-time voters are defined as people who have the right to vote for the first time, which in Norway is Norwegian citizens aged 18 and 19 years at the time of the election. I chose to interview first- time voters for several reasons. First of all, interviewing people voting for the first time increases the probability of detecting an active decision-making process, considering that this in fact is their first time making such a decision and that voting later in life has been proved to become more of a habit. We know that the act of voting in itself is habit-forming (Gerber et al., 2003; Green & Shachar, 2000), meaning that just participating in one election increases the probability of voting in future elections. Research has shown that this applies to vote choice as well (Shachar, 2003). In most democracies, two out of three vote for the same party in successive elections (Shachar, 2003, p. 251). The average voting stability level in Norway has for the last fifty years been higher than 70 percent (Aardal & Bergh, 2015, p. 20)7. This is often explained by party identification; the notion that people establish life-long ties to parties who correspond to their ideological beliefs, which can be used to predict vote choice. However, some research argues that vote choice for many simply is a matter of habit (Shachar, 2003, p. 267), and should not necessarily be regarded as an act of ideological expression. Regardless of which of the theories are most plausible, by studying first-time voters the chances of detecting an active decision strategy are enhanced, limiting observations of habits and passive decision-making. The first vote choice decision seems for many to be a decisive one strongly influencing future vote decisions. Thus, the study of first-time voters is important because the first vote choice might have great implications one the vote choices we make later in life.

Second, as described in chapter 2, vote choice has traditionally been explained by social characteristics such as social ties, issue opinions, ideological beliefs and political values. Even if there are many adolescents who are politically interested and engaged, the main tendency is that they are less so than the older generations. Looking at table 3.1, we see that the political

7 “Voting stability level” refers to how many people who votes for the same party from one election to the next, the opposite is what is called volatility level, which is how many who changes party from one election to another.

27 interest among first-time voters has for decades been way below the average of the rest of the population. And the turnout amongst this group has typically been 20 percentage points below the total turnout of the population in general (Bergh 2013). Most first-time voters have perhaps heard the names of some of the biggest political parties, but never really known what they stand for, or they might have heard about some of the biggest political issues but have not yet had to take a stand. This is not to say that young people are ignorant or careless, it is only natural that one has not spent time reflecting on issues that up until now has not been relevant in one’s life. Thus, explaining the vote choice of 18 and 19-year-olds with variables such as; their evaluation of the state of the economy, of political leaders, their attitudes towards privatization, unemployment and disability benefits might not be highly fruitful. The traditional way of explaining vote choice therefore might not be as fitting when studying first- time voters.

Table 3.1 Percentage of first-time voters that are “very” or “quite” interested in politics compared to all voters, 1993-2017. Percent. 1993 1997 2001 2005 2009 2013 2017 All 59 61 65 67 67 65 62 18-21 47 45 55 41 43 57 41 N 2177 2051 2041 2003 1776 1726 1965 Source: Norwegian National Election Study

Finally, I chose to limit the study to first-time voters who attend high school because it was the best way to assure some variation in the sample. There is no public list of all first-time voters in Norway that I could access, which meant that it was never possible to draw a fully randomized sample limiting the possibilities of generalization. However, knowing that the majority of my population (first-time voters) attend high school, made the selection process more controllable (i.e. controlling the school profile, draw a random sample within a class and so forth). Even if the majority of first-time voters do attend high school8, the contextual differences between those who do attend high school and those who do not could result in differences in their levels of political knowledge, participation and also decision processes.

8 92,3 percent is the last updated number from 23th of Feb 2018 (ssb.no).

28 The high schools in Norway are very focused on preparing and informing the students about the upcoming election; arranging school elections, inviting the youth party leaders to debate and putting up party stands on the school ground. In addition, teachers have their own assignments related to the election (the respondents talked about having presentations, and role-plays where they represented a party). There are in general many mandatory activities that high school students have to participate in, which means that most first-time voters attending high school do have at least a minimum of political knowledge before casting a ballot. The fact that many of the first-time voters already are familiar with the process of voting, assuming that most do participate in the school election, would suggest that they are better equipped than those who have not been forced to go through the process before.

3.1.2 Sample frame

The main goal of this study was to contribute to the relatively new theoretical field of studying and understanding voter’s decision processes and strategies. By interviewing first- time voters within a Norwegian context, the focus was to use the data for further development of theory rather than testing and making broad generalizations. The sampling frame employed is therefore aligned with this research goal. For researchers using interviews as their main source of data, it will always be a question of how many interviews that are considered satisfactory for meeting certain scientific criteria. Within the natural limits of this thesis, how much material I would be able to process and how many interviews I would be able to conduct alone, I decided on 10-20 interviews.

In order to target the population, I began by contacting (e-mailed and called) several high schools in Oslo asking for permission to interview their graduating students (first-time voters). This was in the beginning of April 2018. Unfortunately, this was not the optimal time to recruit first-time voters, because of the fast approaching final exams and the “russetid”. I only got one interview from this initial approach9. I then changed my selection frame and used my network, colleagues at work,10 to get in contact with two teachers who at the time taught graduating students in two different high schools in Bærum. These teachers were friends of my colleagues which was crucial and proved to be the only way to get access to

9 This was a student from my old high school who, when the teacher asked out loud in class, volunteered to be interviewed after school. 10 I work as a research assistant at Institute of Social Research. My colleagues knew two teachers who worked at Stabekk and Eikeli and put me in contact with them.

29 first-time voters in such close proximity to their final exams. After informing the teachers about my study, both expressed interest and facilitated for me to come and interview their students during one of their less “important” classes11. On both occasions I was given a private room in which I could interview the students one by one, without distraction from classmates. The interview process started with me introducing myself in front of the whole class, telling them a little bit about myself and my research, followed by the teacher asking if anyone would like to participate in my study. Fortunately, some did, and I interviewed as many as I could during those two sessions. The respondents in my sample are thus partly self- selected, meaning that they chose themselves to be interviewed, but the context in which they volunteered (it was the teacher that invited me to come and not the students themselves) I believe provided some variation to the final sample.

3.1.3 Final sample I ended up interviewing 13 first-time voters from three different high schools; one located in Oslo named Oslo Handelsgymnasium and the other two located in Bærum named Stabekk videregående skole and Eikeli videregående skole. The respondents’ age between 18 to 19 years old. As can been seen in table A.3.1 in the appendix, I conducted five interviews (3 men and 2 women) at Stabekk, seven interviews (2 men and 5 women) at Eikeli and one interview (1 man) at Oslo Handelsgymnasium. The sample is thus fairly gender balanced (7 women and 6 men). All participants were in their last year of high school when interviewed. The time period of my data collection was from May 8th to June 11th 2018.

Both Eikeli and Stabekk are high schools located in Bærum, Norway’s fifth biggest municipality and neighboring municipality of Oslo. It is known to be an area with relatively high socioeconomic status with a fairly homogeneous group of inhabitants12. In the 2017 national election over 64 percent of the electorates in Bærum voted for the parties towards the right of the traditional political spectrum (The Conservative Party, The Liberal Party and The ) (Valgresultat.no-Bærum). The same results was found at the school election, where the right-wing parties gained over 65 percent of the votes at Eikeli and 61 percent at Stabekk (Skolevalg 2017-Stabekk) (Skolevalg 2017-Eikeli). In 2015, which was the year my

11 In the last year of high school, the students have final exams but only in two subjects. The teachers know beforehand which subjects the class have to take an exam in, meaning that if they know that their class is not going to have to take an exam in their subject, the classes right before summer are of minimal importance. 12 Over 50 percent of the workforce in Bærum work in either sales, hotel and restaurant management, property or in finance. The median income in Bærum is twice as large as the median income in rest of Norway. (https://www.ssb.no/kommunefakta/baerum).

30 respondents started high school, the grade level of entrance at Eikeli was 3,96 and 4,27 at Stabekk. Oslo Handelsgymnasium is located in the centre of Oslo, but consists mainly of students from Vestre Aker, and Nordstrand which are the wealthiest districts in Oslo. The right-wing parties gained on average over 63 percentage of the votes in the last national election in these districts. And looking at the 2017 results from the school election, 80 percent of the students voted for either The Conservative Party, the Liberal Party or The Progress Party, suggests that “the apple does not fall far from the tree” (Skolevalg 2017-OHG). The grade level of entrance in 2015 was 46,5 points, which is considerably higher compared to Eikeli.

Table 3.2 School profile. Percent. Eikeli Stabekk Oslo Handelsgym Grade level of 3,96 4,27 4,65 entrance 2015/2016 Turnout 91,4 89 62,2 Support for right- 65 61 80 wing parties in the School Election 2017

3.2 Internal and external validity within qualitative research: A discussion

There is a methodological disagreement regarding the use of the concepts internal and external validity within qualitative research, which is often rooted in the more fundamental epistemological discussion between positivists and constructivists.13 These concepts, internal and external validity, are usually associated with quantitative research, which again is often associated with a positivistic research approach. Internal validity refers to whether an instrument actually measures what it intends to measure or whether it gives the correct or truthful answer (Kirk & Miller, 1986). External validity is concerned with generalizability; to which degree can we use the results from a study and draw inferences about a broader population. Some argue that these concepts are only useful and relevant when assessing the

13 Positivism is the notion that it exists an external truth outside our own perceptions which we with the right data and method can detect. Constructivism on the other hand do not think that there is an external truth that exist outside our subjective perception, but that the truth is a constructed entity and the “truth” is context dependent.

31 quality of quantitative data and does not serve any purpose when applied to the qualitative field, because of the interpretive nature of this approach (Thagaard, 2009). Guba and Lincoln (1994) has for example proposed alternative evaluation criteria which qualitative researchers can use when assessing their data; replacing internal validity with credibility and external validity with transferability (Guba & Lincoln, 1994). However, without diving further into a complex metaphysical discussion of whether or not there exists an external world of which we objectively can observe, the evaluation criteria I will use to assess the empirical findings in this thesis will nonetheless be internal and external validity. How these concepts are defined will of course vary to some degree between the qualitative and the quantitative field, depending on the research goal and the data material employed. However, I believe that both methods should aim to meet certain scientific standards, and by using “common” methodological concepts/tools within quantitative and qualitative research, this might contribute to a reconciliation between these two traditions, from which both can benefit.

3.2.1 Sample frame and sample: External validity When discussing external validity in this study, I consider to what degree I managed to obtain some variation in the final sample. Even if the main focus is not on broad generalization, a discussion regarding the final sample and its representativeness is still relevant. What determines whether the data collected is representative or not depends in large on the sampling method. The fact that I was able to get access to two entire classes and interview them during the school day, and not after school (which was the case in the first interview), strengthens the external validity of the sample in many ways. Firstly, because the cost of participating was relatively low, the number of respondents increased substantially from one interview in the first selection process to twelve interviews in the second. Further, the accessibility also increased the probability of getting a wider variety of students participating. It is not unreasonable to assume that the students who were willing to be interviewed during class are more representative of the study’s population than the person who agreed to be interviewed after class. The fact that the teacher had invited me to class and encouraged the students to participate in the study while they were having a lecture increased the chances of getting not only the “teacher’s favorite”, but also those who were simply curious about the situation, those seizing an opportunity to get out of class, and those wanting to score some easy points from the teacher before the final grades were set. It should be said that the students in this selection process did also self-select, no one was forced to participate, but I consider the contextual differences in this case, compared to the first, to reduce the potential

32 effect of self-selection bias. Lastly, because of the relatively modest number of interviews, I consider it a strength that the socio-structural profile of the three highs schools represented in the sample are so similar. The fact that the participants had more or less the same starting point, in terms of being socialized within a certain political context makes it possible to isolate a potential “school profile” effect. If I had interviewed students from various high schools, schools with different geographical and political orientation, it would perhaps be even more difficult to explain the variation in the data.

3.2.2 Data: Internal validity When discussing internal validity, I consider to what extent I was able to measure the concepts I intended to measure. There are different aspects that are important to discuss in this regard. Firstly, do the respondents give truthful accounts of how they decided on their vote? One thing is of course that they could purposely be lying or modifying their answerers, portraying themselves in a more favorable way etc., which is always a possibility when employing self-reporting surveys and interviews. This could be explained by a social desirability bias:

“Much of what we think we know about human behavior comes from self-report measures. Unfortunately, the basic human tendency to present oneself in the best possible light can significantly distort the information gained from self-reports. Respondents are often unwilling or unable to report accurately on sensitive topics for ego-defensive or impression management reasons” -Fisher (1993).

Politics is for many a sensitive topic, which has the potential for creating social desirability- affected answers. Political engagement and interest are often considered ideal goals, citizens should be concerned about how their country is ruled. However, many find politics confusing and/or not that interesting. When asking first-time voters questions (face-to-face) about their voting process, their political interest and their political involvement, and about important political issues, the probability of detecting some social desirability bias is relatively high. I tried to minimize the effect by explicitly telling the respondents beforehand that the interview was not in any way a political knowledge test, and I encouraged them to be as honest as possible. However, there is still a chance that the respondents gave answers that made their decision process and decision choice seem more rational and well-thought out than it actually was.

33

Second, the goal was to obtain information about their decision process, and although people can usually produce an explanation of their behavior, that explanation may not be accurate because people do not have direct introspective access to many of their mental processes. This is the main argument in the seminal article “Telling more than we know” by Nisbett and Wilson (1977). In the article they demonstrated the limits of people’s ability to introspect about the nature of cognitive processing responsible for one’s preferences, choices, and behavioral inclinations, in both the social world and the non-social, physical world (Berger, Dennehy, Bargh, & Morsella, 2016, p. 168). However, later research has argued that even if there are some aspects about our cognitive processes that are difficult to report accurately or to recollect, such as processing speed, people are very much conscious and capable of recalling other aspects such as objects and inclinations towards them (Berger et al., 2016). Lastly, the fact that the interviews were conducted 8 months after the election pose a threat to the internal validity of the answers. A lot has probably happened in the life of the interviewees since September 11th 2017, and there is no guarantee that some parts of the decision process has not been lost on the way. I tried to structure the interview in such a way that their memories could be invoked, and that they were allowed to ask questions which could help stimulate their memory. The time-gap has different effects on the outcome; one being that their decision at the time was actually more informed than the respondent is capable of remembering, or that they can rationalize their decision after having made their decision e.g. saying that they were more politically interested or that they cared more about the outcome of the election than they actually did at the time. It is of course difficult to avoid such effects and it is important to take note that they might exist.

3.3 On interviewing

3.3.1 Semi-structured interviews The aim of this study was to detect different decision strategies employed by first time voters, a research subject I considered best explored qualitatively. In contrast to Lau and Redlwask who used experiments and process-tracing as their method, I choose semi-structured interviews. What characterizes a semi-structured interview is that the questions are premade, but open-ended. This form is advantageous for the purpose of this study, firstly it allows the respondents to express their own perceptions freely, potentially providing more honest and accurate responses, compared to a survey where the answers has to fit preexisting categories. This is a strength when the goal is not to quantify or measure already existing concepts, but

34 possibly detect or explore new ones. The field of voters’ decision-making strategies is as mentioned fairly new with many aspects left to investigate. Giving the respondents the opportunity to answer the questions in their own words increases the chances of detecting these unknown aspects. Second, the fact that many of the questions are premade and structured in an interview guide, provides a basis for comparability, a trait that is important when wanting to detect some broader, more general tendencies in the material (Bryman, 2012, s. 472).

A third aspect of semi-structured interview is that you can vary the order and the wording of the questions, adjusting it to the natural flow of the conversation. I wanted to create an informal atmosphere where the subjects could feel comfortable and safe, and not like they were being tested. I considered this important because politics and motives behind political actions or non-actions could for many be a sensitive topic. I reasoned that if I had followed a very strict scheme, this might have disrupted the flow and the natural dynamic of the conversation, which in turn could cause some unwanted effects; making the students feel like they had to impress me or say things just because they thought it was the correct thing to say. In fact, I explicitly told them before the interview that this is not a test, I am only interested in how first-time voters make their decision and there are no wrong answers.

Another reason for choosing the semi-structured interview, rather than the structured version, is the characteristics of the respondents; first time voters are young adults and most of them have never been in this type of situation before. Being asked about the motives behind one´s political actions and orientation can for some be considered frightening. Also, the respondents are high school students while I am a student at the university, the balance of power is thus not equally distributed. Such imbalance in an interview setting can as mentioned have unfortunate effects, and it was important for me to equalize this power relationship. By using the semi-structured interview, I was able to guide the respondents when needed, e.g. elaborate on a difficult question or explain a concept they did not understand. This dynamic would not have been possible if I had chosen the structured interview.

Methodological considerations There are both positive and negative effects with such an informal format. On the one hand, creating a casual interview setting I think enhances the chances of truthful answers, which again increased the internal validity of the data. Making the interview into more of a conversation I think helped the students to open up and be honest on some potential sensitive

35 issues, which of course was crucial for the quality of the data. On the other hand, the bigger the difference between the interviews, the bigger the possibility of measuring errors and comparability problems. The questions in the interview were as mentioned open-ended, meaning that the respondents were free to provide as long or short answers as they wanted. The length of the answers varied between the respondents. Those who gave very short answers were asked follow-up questions. These cues were premade, and were created so as to stimulate the respondent’s memory, making them think about the question in a new way, and possibly help them remember more details. Others, who would speak more freely and give elaborate answers, could receive one question and then answer in such an extensive manner that it included answers on many of the other questions as well. In those cases, it was a challenge to make sure that all of the questions in the interview guide were covered, without having to make the respondents repeat themselves. However, if it was a very important question which the responded had only slightly touched upon, I would ask them to elaborate.

Thus, the actual unfolding of the interviews varied from respondent to respondent. Some could argue that these differences might affect the comparability of the interviews, and in turn the external validity. However, such trade-offs are always an issue when choosing a research method. In this case, because this study is more about the exploration of a fairly new topic (voter decision strategies), at least within the Norwegian context, the primary goal was to make sure that the answers the respondents gave were as truthful as possible (high internal validity) and less on broad generalization (external validity). However, the purpose of the interview guide is to provide some basis for comparison, making sure that everyone answered more or less the same questions. This makes it possible to see if there are some broader systematic tendencies in the data.

3.3.2 The interview process and the interview guide I began the interview by introducing myself and the project. I then assured them of my full confidentiality and let them know that they could withdraw from the interview at any point. The introduction and the formalities were written beforehand and read up directly from the interview guide, this was to secure a similar starting point for all of the respondents. I then gave them a declaration of consent form to sign and asked permission to record the conversation. Everyone signed the form and allowed me to record.

36 The interview guide consists of 19 (superior) questions, each of them with between two to seven additional, premade follow-up questions also known as “probes”. The probes were used as a tool to stimulate the respondent’s memory; giving examples of potential answers, or to obtain additional information that was not given after asking the initial question. As mentioned before, some people are less talkative than others and do not give supplementary answers unless they are explicitly asked to do so. In addition, the interviews were conducted 8 months after the election, so to provide “cues” for those who needed a little extra stimulus seemed to be reasonable. Because the interviews were conducted face-to-face, I could easily tell based on their facial expression if there was something about the questions that was unclear, e.g. a difficult concept or something they did not quite remember. In those cases, I could explain and help the respondent. In addition to the premade probes, I asked follow-up questions that was not explicitly written in the interview guide as well. If the interview reached a state of a natural conversation, I would use this advantage to dig even deeper, asking them to elaborate on certain topics. I must admit that the more interviews I did, the better I became at improvising fruitful follow-up questions. In the following section will I discuss some of the questions that were asked during the interview and their theoretical anchor.

Political interest and expertise The first two questions I asked was about their degree of interest in politics. “How interested in politics would you in general say you are?” and “Did you care about which party or parties that would win the election?”. These questions had two purposes: First, because I assumed that many of the respondents would not be that interested, I considered it clever to start the interview with questions specifically about this, so that I could show them that their level of interest/disinterest would not influence their participation value. After hearing the topic of the interview, many did perhaps enter the interview worrying that they might not be the right fit. Letting them know immediately that a high level of interest in politics is not a criterion for participation. I think encouraged them to be more open and honest during the rest of the interview as well. This is method I learned from the Norwegian National Election studies, who always start their survey with a question about political interest. Second, interest in a decision task is one of the factors that are hypothesized to affect what decision strategy one will employ in a decision-making process, and it is also one of the indicators that Lau and Redlawsk used to measure political expertise. I also asked questions about whether they had discussed the election with family and friends, and political participation and if they

37 considered it important to familiarize with the election campaign. Even if I did not explicitly measure political knowledge, I included the rest of the indicators which I believe will be sufficient when looking at the effects of political expertise on decision strategy.

Resources spent on the election- Depth and comparability of search I also included questions about resources spent on the election. Questions like “Ahead of the election, did you try to familiarize yourself with the different political parties and their stance on different political topics?” and if yes “How did you go about doing this?”. These questions say something about which parties they did or did not consider, which is a measure of the comparability of search indicator. In addition, questions on how they did their research can tell us something about the depth of search.

Other questions I asked if they thought it was easy or difficult to decide which party to vote for, followed by a probe asking, “why was it easy” or “why was it difficult”. As we know, the level of perceived difficulty associated with a decision task is theorized to affect the decision strategy. Additional questions I asked were about; their vote choice, when they decided to vote, if there were any other parties they considered, if there were any specific issues that influenced their vote choice and if they themselves could point to the factor that ended up being decisive for their vote. The interview guide in its completeness can be found in appendix A.3.2.

3.4 On context – How can the high school context affect the decision strategies employed by Norwegian first-time voters?

Almost all high schools in Norway have mock elections, a small-scale election arranged at the schools two or three weeks before the actual election. The schools invite representatives from the biggest youth parties to debate in front of the students. The idea is that through this pretend election and debate the students learn about the parties and are familiarized with the process of voting. The school debate is mandatory for all students. The public high schools strongly focus on political elections and giving first-time voters as much information as possible. This is done by arranging role-plays where each student represents a different party making them debate against each other, visiting party stands in the city and class discussions in a broad range of subjects. Most of these activities are mandatory and the students are often graded on them as well. This means that even the most uninterested student will have a basic

38 knowledge of the different parties (at least the biggest parties and their policies). This is probably one of the reasons why first-time voters (those aged between 18-21) participate on a higher level than those who are in the next age cohort (22-29-year-olds). This makes it more complicated to measure the depth and comparability of search amongst this group of voters. When I asked if they had tried to familiarize themselves with the parties and their politics, many answered that they had to because of school assignments. In order to correct some of this noise, I asked the respondents if they think they had done this if it had not been for those school assignments. This question will not eliminate all the potential errors, and it will still make it difficult to detect those who would otherwise employ a “gut-feeling” decision.

3.5 The analyzing process of the data

Before the analysis, a short description of how the data has been processed and analyzed will be given. First, I transcribed the recorded interviews. Second, I read through the transcripts and structured the respondent’s answers in an excel sheet, so as to get an overview of the material. Third, I tried to place the respondents into the pre-existing categories made by Rau and Redlawsk. By doing this I could see which categories that fitted, and which that needed modifications. The interviews were conducted in Norwegian, so the quotes have been translated.

39 4 Analysis

According to the process-oriented framework proposed by Lau and Redlawsk (Fig.2.1), different factors influence what decision strategy voters employ. Background characteristics (socio-demographic factors), political sophistication and campaign factors/task demand all influence how a voter subjectively perceive a decision task (“nature” of decision task), which again influence the information processing. How a decision task is perceived, the level of importance and complexity associated with the task, depends on the preceding factors and affect the (more or less conscious) choice of decision strategy.

Through the interviews, the goal was to obtain information on as many of these factors as possible and see how they translate into different decision strategies. In the following chapter, I have categorized the respondents, according to their answers on decisive questions, into the five categories identified by Lau and Redlawsk (2006) and Lau et al. (2018). The strategies are used as the starting point for the analysis, and there will be a comprehensive discussion within each strategy. As a means to structure a potential complex analysis, I have identified four factors that I found to be useful “predictors” of decision-making styles; (1) perceived importance of decision task, (2) information processing, (3) campaign factors/task demands and (4) political expertise/sophistication. It is the variance within these subthemes/factors that I found to be the most decisive and fruitful for studying decision strategies. There will be a discussion within each category on the compatibleness between the original categories and the theoretical expectations and the actual strategies employed by the respondents. Pseudonyms are used to maintain the anonymity of the respondents.

4.1 Strategy 1: Rational choice decision-making

“I just did the research and found the party I agreed with the most” -Nora (Eikeli)

The rational choice decision-making model is a compensatory strategy, based on the assumptions from rational choice theory. It presupposes deep and comparable information search ahead of the decision. There were two respondents, Emma and Nora, that employed a strategy that could be categorized as a moderate version of the classic rational choice strategy. I do not think that Emma or Nora in reality considered all of the alternatives (parties) nor did they carefully evaluate every possible outcome trying to find the most value-maximizing alternative, which is strictly speaking the definition of the classic rational choice model.

40 However, I still consider their decision strategy a version of the rational choice, at least the “constrained” version, which reasons I will now discuss.

Perceived importance of decision task. First of all, from the interview it was obvious that both of them had some general interest in politics, they did not have sleepless nights thinking about the decision task, but they both acknowledged the importance of the election. When I asked Emma about her political interest, she said “What I think about the most is that I pay taxes, and I don’t want my tax-money to be spent on something stupid, so I care in the sense that it matters some, but it is not what I think is the most interesting”. When I asked if they cared about which party or parties that would win the election Emma said “Yes, but again, in the same way, I care about how my money is spent, my tax money, and if there are any big changes in society”. Nora´s response to this question was “Yes, because it is them, how can I say… that will represent us…and I care about which issues the parties stand for”. Both also expressed that they considered it important for them personally to familiarize with the election campaign before casting the ballot. Nora stressed that she did not want to just vote randomly, she needed to be able to stand for her decision. In addition, both explicitly said that they thought it was important that their decision was made without the influence from others, which stood out compared to many of the other respondents who were very open about the fact that their vote was heavily influenced by friends and family. When I asked if they had discussed the election with their families, and if that in anyway had effected their choice Emma said “I did for a long time consider voting for the Conservative Party, because my mother is a very strong supporter of that party, and I think that would have influenced me if I didn’t decide to read up on the issues myself. I think that when you familiarize (with the political campaign) you find your own little meaning in it”. Nora said “It could, I mean how you grow up and what you have and don’t have, do affect you, but it’s not like I voted the same as my family or friends”. Both wanted to make an independent decision, they were reluctant not to rely on other people’s opinions. For Emma and Nora, it was not only the decision task in itself that was considered important, it was also important that their decision was independently made, motivating a deeper information search.

Information search: It is not clear exactly how many parties they really did consider, which makes it difficult to say anything specific about the indicator “comparability of search”. However, based on their statements, we know that they approached the decision task without a preexisting party preference, they had an open mind and wanted to make their own, independent decision. Of course, this does not mean that their research effort was equality

41 distributed between all the parties, which it most likely was not. However, they often referred to different parties and what they liked and not liked about them, suggesting that they had at least familiarized with more than just the biggest parties. When I asked how they managed to reach their decision it became evident that they both had done extensive research. Nora said that she had a subject in school called Politics and Human Rights which she was really into, and where she had learned a lot about the election and the parties, in addition she had read up on the different parties on their respective webpages and taken several election quizzes. She had also had many discussions about the election with her friends. Emma had mostly read on the parties’ webpages and watched the news. She had also discussed the election with her family and friends and taken several election quizzes. When I asked if there were any specific issue that they cared about they both mentioned several political topics. Nora for example said that she specifically liked the Liberal party’s politics on “immigration, because they have a more moderate position than the others, and integration and health and that they support animal welfare.”. Emma said that she liked the whole party (the Liberal), but specifically mentioned their stance on environmental issues and education. According to cognitive theory, the more time you spend on research, the more likely it is that this research is memorized and becomes a part of one’s long term memory. The fact that both Emma and Nora, eight months after the election, was able to mention what political issues they specifically liked about their chosen party (the Liberal Party) suggest that their information search was considerably deep, which is one of the main indicators of the rational choice strategy.

Campaign factors/task demands. Campaign factors, as explained before, are the situational or contextual nature of the campaign itself. One important factor is the complexity of task, the number of alternatives times the number of different attributes across which they vary, this is something that is said to influence decision strategy (Einhorn, 1970). When I asked if they thought it was easy or difficult to decide on their vote both said it was difficult but had different reasons for why that was. Nora said “It was quite hard… There are so many parties that have parts that I liked, but there is not one party that I agreed with hundred percent, and it is difficult to weigh what’s most important”. From this, we learn that Nora did consider the decision task complex, because she had considered many of the parties as potential candidates for her vote. She had read about their politics, found some parts that she liked, and others she did not like, in the end she had to weigh all of these aspects against her own initial preferences and combine them into a decision. Emma also thought the decision was difficult because “I don’t feel that they (parties) are honest. They just say the things we want to hear, without being able to fulfil half of the things they say they are going to do…”. Emma’s

42 understanding of the decision task is perhaps even more complex, considering the fact that she does not even trust them. Before making a decision, she first had to familiarize with the party’s politics, and then she had to assess their credibility.

According to the theory by Tversky, if there are more than two alternatives the probability of employing a noncompensatory strategy is increased (Tversky, 1972). I suggested that in Norway, because how the party- and election system is structured, the decision task is objectively complex, resulting in a higher use of noncompensatory decision strategies. Of course, this requires that there is a correspondence between the perceived alternatives and the actual alternatives. Both Nora and Emma did consider several of the parties, however they still employed a compensatory strategy. I believe this could be explained by the notion that the effects of the complexity of task will initially depend on other factors, such as the perceived importance of task. Emma and Nora both considered the vote choice important, they did not want to use their parents or friends as heuristics, in order to reach a decision. Thus, in this case, the effects of complexity are overridden by the perceived importance of decision task. The goal was decision accuracy rather than decision efficiency, which means that it does not matter how many alternatives or potential outcomes there is, because the decision-maker consider the decision task important, resulting in a compensatory decision strategy.

Political expertise/sophistication. As discussed in chapter 3, political experts were assumed to spend less time on information processing, only paying attention to specific heuristic information. Lau and Redlawsk measured political expertise on campaign-related and community-oriented political behaviors, a stringent political knowledge test and measures of following politics in the media, frequency of talking about politics and self-proclaimed interest in politics. I do not have the same measure of political sophistication, although I think I am able to distinguish between the level of expertise within the sample, based on their interviews. Emma and Nora are, according to this study’s informal measure, considered if not experts, then more politically aware than most of the participants. However, they did employ a cognitively demanding compensatory decision strategy, when theory suggested otherwise. The problem is that the theoretical assumptions that Lau and Redlawsk use when making their hypothesis is not essentially valid in this case. They refer to studies where those who are considered experts, are defined by their experience within a decision domain, which affects the frequency and the recency with which possible decision strategies have been used (Lau & Redlawsk, 2006, p. 38). This definition of experts might be fruitful when looking at veteran

43 voters, but it perhaps not as useful when explaining decision strategies amongst first-time voters. One can be considered a political expert or politically sophisticated when being 18 or 19 years old, but because the vote decision is a new decision task, the effects of political expertise could prove to be rather the opposite (unless you have a strong political preference), because you want to make a good, and thoughtful decision without any short-cuts. When you are making a decision for the first time, a decision you consider important, you will consider all the aspects of this decision task, resulting in a compensatory decision strategy.

4.2 Strategy 2: Confirmatory decision-making

“I spent some time on the process, but it wasn’t that hard, of course the decision landed on what it did” – Elias (OHG)

The early socialization and cognitive consistency strategy, later renamed confirmatory decision-making, has its origin from the Michigan/party identification model. The notion is that a voter’s early socialization towards or against prominent symbols, such as party affiliation or social group, will affect how a voter process information related to the vote choice decision task. The information search is biased, so as to maintain cognitive consistency and avoid cognitive dissonance. This strategy can be placed somewhere between a compensatory and noncompensatory decision-strategy, meaning that the information search can be deep, depending on how important the decision task is, but not comparable. After analyzing the interviews, I identified three respondents who employed a version of this decision strategy; Sara, Ella and Elias. All three had a political orientation before the election campaign started, and even if they experienced moments of doubt, their final vote choice was as expected.

Perceived importance of the decision task. The respondents who employed the confirmatory decision strategy are above average interested in politics. This is not unexpected, it is rather obvious that people with clear political preference would care more about politics than those without. Sara said that she considers herself more politically interested than most her age. She reads the news, watches political debates and follows political parties on Facebook. Ella said that she thinks that it is interesting to discuss, but that she cares especially when there are very big issues, such as “the cases with Sylvi Listhaug”.

44 Elias also said that he considered himself above average interested in politics, and he thought it was stupid not to be, he also added “I really care about how our country should be ruled, and how our society will look when I am older…”. When asked if they though it was important which party or parties that would win the election all three agreed. For Elias it was important because he really wanted his party win, and it became an emotional thing for him. Sara said it was important because she was left-oriented, and she really wanted them to get the prime minister post. To the question if they personally felt that it was important to familiarize with the election campaign before voting, Ella said that she thinks there is no point in voting if you do not have a proper basis, that’s just stupid. Elias had the same sentiment saying: “Yes, it was. If you don’t know anything, and just walk in with an attitude that you are just going to vote the same as your parents, because then we can pay less taxes, that is so stupid… I mean, if you actually stand for the policies that that party promotes, then go ahead and vote. But don’t just vote because of a small issue you have heard you father talk about, or something similar. That’s just silly.”.

There is no doubt that all of them considered the decision task important, and they all explicitly expressed their feeling towards those who did not take it as serious. One interesting notion is that their political orientation must initially come from somewhere, and that somewhere can often be traced back to those close to you, like family or friends. But because they have a genuine interest in the decision topic, and find the decision task important, they want to make an educated decision, they want to distinguish themselves from those who just “vote the same as their parents” by doing research that supports their early learned predisposition. This gives them a feeling that their decision is to some degree independently made, although party identification is in reality just one big heuristic. So, what differentiates these decision makers from those who employ a pure heuristic-based strategy is, their initial interest in politics and their political preference, making them invest more time on research, research that is either consciously or unconsciously focused on strengthen their initial beliefs. If you do not have this interest in politics, or have any emotional attachment towards the decision task, you will not bother with extensive research to support the decision.

Information search. One of the main characteristics of those employing confirmation-based decision strategy is biased information search, you only gather information from sources you know that will strengthen your preexisting beliefs, in order to avoid cognitive dissonance which is known to be unpleasant. Ella was determined on her vote to the Liberal Party, and admitted that her information search was limited, and that the research she had done was due

45 to mandatory school assignments. However, when she was given the task of familiarizing with the different parties, she used this as an excuse to gather more information about her initial preference, which strengthen her feelings towards the Liberal Party. “I just compared the Liberal Party’s politics to the Progress Party and all of them, and the Conservative party, and then it was obvious that I wanted to vote for the Liberals”. Her approach was using the Liberal party’s policies as baseline which the other parties had to match up to, and when they did not, the choice was easy. When I asked how she reached her vote decision she said “I started to see that they fought for animal rights, animal police or something like that. And they are very focused on the environment, and I think that the Greens can take it too far sometimes, the Liberals have a fine balance on what they focus on, they don’t over- dramatize”. It is rather obvious that Ella’s determination on the Liberal Party is guiding her information search. She saying “I started to see that they …”, tells us that this was perhaps information she did not have initially, but after learning about it she felt even more convinced that this was the party for her. I think when you approach the decision task with a clear preconception, the task of finding at least one thing in that party program you like is relatively easy.

Elias was on the one side very open about the fact that his parents did obviously influence him, but at the same time he often stressed that he had also done his own research. He said that he matched the Labor Party’s politics with his own value orientation. When you grow up in a home where both your parents have a strong affiliation towards the Labor Party, the chances that this party’s values have in some way manifested into your own is relatively high. The idea that Elias, in order to make an educated decision, matched his own values, values which most likely has its roots from the Labor party ideology, with the Labor party values, is in itself the pure definition of biased information search. Another interesting fact is that we know that it takes a lot of effort to change someone’s ideological conviction, which means that when a person is faced with information that potentially could shifts his initial conviction, this information is often regarded as not important or fake so as to maintain cognitive consistency. When I asked Elias if there were any specific issues that he thought influenced his vote he said “There was a lot of fuzz ahead of the election, with Jonas Gahr Støre. But I don’t think that the cases that are brought up right before the election should be considered, unless they have a real impact…”. Indeed, there was a lot of “fuzz” around Jonas Gahr Støre, but how that information was perceived would in turn depend on one’s initial feelings toward Gahr Støre and the Labor Party. In the case with Elias, this information was considered

46 irrelevant, while it perhaps was considered highly relevant amongst those who did not have this early socialized affection towards the Labor Party.

Sara’s information search was in some ways a bit different than the others, by the fact that she did not have one particular party that she clearly preferred, rather her preference was oriented towards the political left in general. I still consider her decision strategy as a form of confirmation decision making for several reasons. First, she said that she discussed the election with her friends, all of whom were active members of the Socialist Left Party. They gave her very good arguments for why she should vote for their party and she thought the arguments were sound and convincing, but she said “They did not influence me, they just gave me good arguments”. She also said that she followed many political parties on Facebook, and that she had read many of their posts ahead of the election, but “it is typical that the things you get on your Facebook is things you are already interested in, so then it becomes like a confirmation bias”. She is definitely searching for information within a safe environment, not exposing herself to sources that potentially could cause insecurity. However, she stressed that she was really considering three to four parties, and that she systematically went on to eliminate the alternatives one by one after learning their political stance. This suggest that there was a two-step decision strategy, where the first one was her early learned predisposition towards the political left that automatically limited her alternatives, but once her room of maneuver was established, she employed a rational choice strategy, learning as much as she could about each of them. In the end she said that she voted for the Social Left Party because their arguments were the most appealing.

Campaign factors/task demand. A decision task is considered easy in instances where the decision-maker prefer one alternative above all other alternatives. This should be the case when people have a strong party affiliation. However, even if the respondents expressed a clear inclination towards one party, or political side, they all expressed that there had been a process, in the pursuit towards making their own independent decision, and not relying on others. Ella thought the decision task was somewhat easy because she had a preexisting party preference, however she became uncertain when her friends started to ask critical questions regarding her choice, she explained “Well, I knew who I was going to vote for, but then you get comments from friends being like “What, are you really voting for them? Why?” which made me insecure… but then when I was alone, I thought to myself “No, I am voting for who I want to vote for””. Elias could also report that there had been a process “I was back and forth, when you are a first-time voter you don’t know much, you only know what your

47 parents have done, and so you enter the decision task with a preconception. But I did a lot of research on my own, I took several election quizzes, and in the end, they all gave the same result, and there is a reason for that. So, the process took some time, but it was obvious that the decision would land on what it did.”. When I asked when he decided on his vote, he explained that both his parents are supporters of the Labor party, and throughout his life, that has been the party they have voted for, so “it had been lying in the back of my head for some time”. So, from this it sounds like the so called “process” he went through was not because he considered voting for another party, it was rather a process of gathering information that strengthened his preexisting, early socialized value orientation. This is a perfect illustration of a person employing a confirmatory decision strategy. Sara did also report that she considered the decision task somewhat difficult, “It was not difficult to eliminate some, but it was difficult to land on one”. She said that she thought the whole thing about “tactical voting” was confusing, and that it was difficult to determine what politics each party really stood for “because all parties say that they want better care of the elderly and schools, which is great, but in the end its about how it’s done”.

I think there are several reasons for why the respondents did not consider the decision task easy, which is in part contrary to what one would initially assume about those who employ the confirmatory decision strategy. First, because of their young age and political inexperience, their political orientation has not yet been fully developed. Their political attitudes and values are just starting to take form, and it has not reached a state where they have internalized their preferred party’s values into a comprehensive ideological schema. When I asked if they felt like they had a clear party identity, they all said that they would not go as far as to say that, they might vote differently next time. I was admittedly surprised to hear this, but it illustrates that fact that they are not completely convinced by their temporarily political conviction. This again could be due to the fact that they are inexperienced voters, and might feel that they do not have a good enough overview of the election process and the parties to truly settle and be confident with their choice.

Secondly, which is partly related to the first, when your decision departs from the norm, and you experience that friends and family questions this decision, and you lack the arguments necessary to answer them satisfyingly this might result in second throughs and confusion. All of the respondents who employed the confirmatory decision strategy voted for parties traditionally regarded as laying on the lefts side of the political spectrum (The Social Left, The Labor Party and The Liberal Party, although this party is defined as a right-wing party,

48 she emphasized the policies usually associated with the left). This is not common parties to vote in those areas where they are from, which is, as showed earlier traditionally very right- wing oriented. All of them were aware of this fact and made it a point that their vote choice deviated from the norm. Because they had not yet developed a strong party identity, only a preference, the pressure from peers and family might have complicated a decision process that otherwise should be relatively easy. This will most likely be the biggest difference between first-time voters saying that they have a party preference and the experienced voters with a party identification. The decision task is considered more difficult amongst the former, and as the party preference develops into a party identity, the decision task will be easier.

Political expertise. The respondents that employed the confirmatory decision strategy could, according to the measure used in this study be considered political experts, or above average politically aware. They, similar to those who employed the rational choice strategy, expressed a genuine interest and respect for the decision task, they all had engaged in political discussions with friends and family and kept themselves updated by watching political debates, reading the news and by following political parties on Facebook. This should then, according to the theory presented by Lau and Redlawsk, mean that these respondents spent less time on information processing, because of their previous experiences with decision- making within their expert domain. I argued that this understanding of political expertise was not fitting when looking at first-time voters, because of the fact that this is their first time voting. However, I would think that the respondents who employed this decision strategy spent less time than those who employed the rational choice strategy, even though both groups did their fair share of research. I do not have the exact number of hours the respondents spent searching for -and processing information, but it would not be unreasonable to assume that voters who had a preexisting idea of who they would vote for, the process of getting those extra supportive arguments is easier and requires less time than if you enter the decision task “carte blanche”. I think that it is difficult, in those cases where people have preexisting party preferences, to see the isolated effects of political expertise on the employment of decision strategy, not only in this study but in general. Those who have a strong party preference is often also interested in politics and are more politically engaged compared to those who do not have a party identification. I would argue that the effect of their political affiliation on decision strategy are stronger than the effect of their political expertise, however there might be an interaction effect.

49 4.3 Strategy 3: Fast and Frugal decision-making

“There is so much information, and of course, I understand that politics has to cover all aspects of a society, but it is just too much. So, you have to steel oneself on one issue that you especially care about” -Sofie (Eikeli)

Voters who employ the fast and frugal decision strategy is primarily motivated by efficiency. They seek out only the most diagnostic information that allows them to quickly make a decision. What this diagnostic criteria is could be either “hard” or “easy” issues. It is categorized as a compensatory strategy, meaning that the decision maker is willing to do some evaluation of the different alternatives, but not to the same degree as those employing the rational choice strategy. The information search is often shallow, but comparable. I identified one respondent, Sofie, that employed the fast and frugal decision strategy. Sofie’s decision strategy does not correspond perfectly with the strategy described by Lau and Redlwask, however there are some core elements that are similar, which I discuss in the following. I have excluded a discussion on the effects of political sophistication because of its limited relevance, which I accounted for in the discussion of strategy 1.

Perceived importance of decision task. Sofie said that she is average interested in politics. It depends on the issue, she is more interested in certain topics. When I asked if she cared about which party or parties that would win the election she answered “Well, yes to a certain degree. It does perhaps not really affect me that much yet, compared to those who work and stuff. So, a little, but not considerably.”. When I asked her if she considered it important to familiarize with the political campaign before voting she answered “Ehhm, yes-ish, it’s like, I really don’t feel that it matters that much when you are young but you hear people saying like “you have a voice, and it is important that you vote for your future”, so I felt like I had to. And then I wanted to focus on something and actually choose something that I want, and not just do what everybody else is doing”. Her perception of politics and the election is understandable, and I would assume quite widespread amongst first-time voters. She is conflicted by the fact that she knows that her decision is “important” but at the same time she feels like it does not really affect her yet. She then agrees on a compromise with herself, she wants to make a decision that she cognitively can be satisfied with, without having to spend hours on end researching every party and every political issue. She chose the environment as the diagnostic criteria but did not want to go deeper into the matter, and just decided on the

50 Liberals because they had a fair chance of winning. I will discuss this in more detail in the next section.

Information search. Because she was not too interested in politics, the amount of time and cognitive effort she was willing to spend on her decision was limited. However, she wanted to make an informed choice, not just vote the same as anyone else. She also said that she would not have been able to familiarize with everything either way, so she figured that her best chance at making a satisfying decision would be to focus her attention on just one topic; environmental issues. This made it possible for her to limit her information search, concentrating only on two parties she felt was relevant. This is a typical cognitive mechanism employed when wanting to simplify a decision task. Because she lacked the motivation necessary for doing the full rational choice information search, she employed the decision heuristic; editing or pruning, which is when you simplify the decision task by ignoring otherwise relevant aspects. By eliminating all other parties, and all other issues the decision task becomes much more feasible. When I asked if she tried to familiarize with the political campaign, she said that she had visited some of the party stands at school and listened to the political school debate and read on some of the parties’ webpages. However, she thought the political debates was boring because the politicians argued too much, and they were more focused on attacking each other than promoting their own politics. She also thought that the school debate was unprofessional, trying to appeal to the youth by talking in a certain way. It is clear that Sofie was not very impressed nor interested in all the “shenanigans” connected to election. The majority of the information she got was from school arranged events, and there was no external information search besides that. It is of course difficult to judge the depth and comparability of her search, when so much of her information was obtained through obligatory activates. Even if she did not engage in external information search, she might have obtained enough information to make a somewhat satisfying decision. This pose a more general problem when assessing the information search amongst the first-time voters attending high school; the fact that many did not seek out any additional information by themselves, does not necessarily mean that their decision choice was uninformed or that their information search was biased or shallow. It might as well be that they paid attention at school, leaving them with perhaps more information than they would otherwise have.

Campaign factors/task demands. Sofie said she thought the decision was easy because she chose to focus on the environment, and that her single issue focus left her with only two parties, the Liberal Party and the . However, she said that it was quite difficult

51 choosing amongst those two again. She ended up voting for the Liberal Party. When I asked her how she decided on that she said “I knew that the Greens would never be elected…”. This suggest that she employed the heuristic viability; only voting for a party you think have a chance of winning. This heuristic was employed, I think, because she (as expressed) was not that interested in the decision task to begin with, she had no motivation for really getting into the whole environmental debate.

Because she felt that her vote choice did not matter as much, she chose to only focus on one issue, simplify the decision task and the information process. She used the environment as a diagnostic criterion, concentrating only on the two parties she knew cared about the environment. Even if Sofie said that she thought the decision task was easy, it seems like she initially thought it was quite hard, only that she did not care enough for the decision task to really absorb the complexity. She expressed more than once that there was too much information to consider and that she would never be able to familiarize either way, so her best chance was to concentrate on one topic. When I asked if she managed to get an overview of the election campaign, she answered “There is just so much information, and there are of course many aspects of a society that politics has to cover, but it is just so so much. So, one has to focus on a topic that interests you, I think”. This suggests that she initially thought the decision task was difficult, and that the decision process in reality began before she was consciously aware of it. Her first impression of the election was that it was complex, so in order to avoid this complexity (because she lacked motivation) she chose to limit the decision task and focus one issue. Thus, the decision task was not considered easy in the beginning of the decision process, it is rather considered difficult, but it became easy after employing simplifying heuristics. I would think, even if I do not have more empiric evidence in this study, that many first-time voters have the same experience as Sofie. They feel overwhelmed by the whole process, and they have no intentions of even trying to get a complete overview, first because the outcome of the election will most likely not affect them, and second because they do not feel that they have the competence required to deal with such a complex matter.

The main difference between the strategy employed by Sofia and how the strategy is defined by Lau and Redlawsk is the motivation behind the employment of strategy. Lau and Redlawsk describes a mother with three children, her main focus is on logistics and organizing her daily life with these kids leaving very little time for thorough information search, the only thing she cares about is the educational system, so she finds out which party that has the best politics on this issue and gives her vote to them. I believe that Sofie’s motivation for employing such a

52 strategy is not primarily due to time contains, of course she does not want to spend too much time on it, but the main reason is that she does not view the decision task that important (at least yet) and that she is not capable to familiarize with the political campaign, but still she wanted to make an informed choice, leading her to focus on one issue. In addition, Lau and Redlawsk argued the difference between hard issues, and easy issues, and that the decision strategy employed by those single-issue voters emphasizing hard issues could reminisce the strategy used by rational choice-decision makers. Even if environmental issues could be characterizes as a “hard issue”, the depth of her evaluation (quite shallow) and consideration of the two alternatives, and how she finally decided on her vote (viability heuristic), show that the strategy she employed is more “fast and frugal”, than rational choice.

4.4 Strategy 4: Heuristic-Based decision making

«I was a little unsure who to vote for, because the environment is the reason the earth exists, so we should all actually vote for the Green Party. But I voted for the Conservatives. I think Erna is a good prime minister” -Olivia (Eikeli)

The heuristic-based decision strategy is as earlier mentioned a classic example of a noncompensatory strategy; decisions are based on incomplete information search to avoid conflict, and heuristics are heavily utilized so to simplify the decision process. It was this strategy that was the most common amongst the respondents in the sample. I identified seven respondents employing different variants of this decision strategy, all of which had relied their decision on more or less simple heuristics. I have excluded a discussion on “political sophistication” because of its irrelevance within this category (I do not consider any of them political experts), and because I have already discussed why the effect of political sophistications is not found in this study (see the discussion under strategy 1). I have prioritized a longer discussion on the different types of heuristics employed, describing each voter’s decision process in more detail.

Perceived importance of decision task. In general, the respondents employing this strategy expressed medium to low interest in politics, although many said that their interest had increased since the election. When I asked Philip about his political interest he answered “lately, I’ve been quite interested… but I wasn’t last year… I’ve become more interested

53 afterwards”. Olivia said “…It´s not super important, but it probably will over the years, I think, when I´ve become more engaged”. Maja did also express that it was not until after the election that she started to realize the consequences of her decision. She said “I voted for the Conservatives, but it was heavily influenced by everyone around me and I did not realize the consequences until afterwards…”, she later elaborated further saying “looking back, I regret my decision… if I had to vote again, I guess I would have chosen a party more to the left… …I just didn’t care that much at the time, it all happened so fast and we had a debate and the election was just two days later, so I was just influenced by my parents and friends, and did not really think too much about it”. Why did so many express that their political interest had increased since the election? This could either be explained by the fact that they actually have become more interested after they voted for the first time, perhaps curious about the outcome of their first political action, or it could be an expression of hindsight. The latter could be explained by the social desirability-effect discussed in chapter 3; at the same time as they have to admit that their engagement and interest in the election was not that high during the actual decision process, they still wanted me to know that now, eight months later, they do realize the importance and the consequences of their actions. Whichever is correct, the overall impression is that they in general was not very interested in the decision task, at least not at the time of the election.

Oskar, for example, was were open about the fact that he initially was not planning to vote at all, saying “I wouldn’t have voted if it hadn’t been for my parents. They said “You have to vote, you are 18 and you are entitled to vote”, and I was just like “ha-ha, I don’t care”, but then I ended up voting, I did.”. When I asked Philip if he cared about which party or parties that would win the election, he said that he had hoped that the party he voted for would win, but then added that he really would not have cared much if another party won instead. There were no real emotions involved he said. Lukas did not really care either, saying “A little, not much… I figured it would end the way it did…”. Oliva did also express that it did not really matter for her who won the election, and when I asked why she said “I don’t know, I just don’t think I care that much yet”. Thus, common for all the respondents was that they all seemed uninterested in the decision task’s topic, which meant that their motivation for doing a thorough information search was low. This is one of the main characteristics of those employing the heuristic-based decision strategy.

Information search. When the perceived importance of a decision-task is low, the time and effort one is willing to put into that decision is minimal. The decision will in such cases be

54 based on incomplete information and heuristics, which proved to be the common case amongst the respondents who employed this strategy. The difference between this strategy and the “fast and frugal” strategy is the type of heuristics employed. In this case, the decision- maker has no intention of going through a cognitively demanding information process, weighing values and alternatives against each other, rather their main focus is on making a quick decision which they can be somewhat satisfied with. The “fast and frugal” decision strategy is more compensatory in style and involves more deliberate search for information which is used actively in the decision-making process. Those who employed the “heuristic- based” strategy did not have the same motivation for doing such a search. However, the respondents did tell me about their mandatory assignments and the events arranged by their schools in relation to the election, explaining that they did have some information, but when I asked if they though that they would have obtained this information on their own, and if their interest in politics went beyond the mandatory activities at school most of them said no. Oskar adding “I’m not that interested”. Emil expressed the same saying “I’m not that interested that I wanted to spend my own spare time on it”.

Campaign factors/task demands. There was a variation in the level of perceived difficulty associated with the decision task. Some considered the decision easy while others thought it was very difficult. According to theory, those who consider the decision task easy will in a larger degree employ a compensatory strategy, while those who consider it difficult will employ a noncompensatory strategy. Why do people, categorized as employing a heuristic- based decision strategy, a noncompensatory strategy, express that the decision task was easy? I think maybe this variation could be explained by the fact that the respondents have different “stages” of the decision process in mind when answering. After reviewing their answerers, I believe that most of them did at some point in the decision process consider the election process somewhat difficult, but because they initially did not care about the decision task, they avoided the complexity utilizing heuristics which then made the decision task appear “easy”. Although in reality, the election and the process of familiarizing with the different political parties and issues was in fact so overwhelming that this was the reason they employed the heuristics in the first place. There were several that explained that they had talked to their parents in relation to the election, because they were much more experienced than them, and knew politics way better. From this I would make the assumption that they initially did not feel competent enough to make a decision based on their own information, thus had to rely on their parent’s expertise. This one of the main heuristics employed by the respondents.

55

Decision heuristics. The heuristic that was most commonly employed was the one called “endorsement”. This is one of the decision heuristics described by Lau and Redlawsk where the decision-maker bases their decision on recommendations from close acquaintances, trusted elites or social groups whom they identify themselves with. This decision heuristic could be employed in circumstanced where the decision-maker do not feel competent enough to make an informed choice and would rather leave the decision task to someone that is competent or who has experience within that decision-field. It might not come as a surprise that there were many who used their parents as such a decision heuristic. There were also many who said that their friends had influenced them. I also detected some instances of the decision heuristics viability and affect referral as well. The former is about only voting for those who you consider having a fair chance of winning, and the latter is simply voting for the party or candidate you like the most.

Philip and Lukas were one of those who considered the decision task easy, and when I asked why, Philip said “a lot of influence from others” and then followed up by saying “I cannot remember the reasons I had back then, but it was mostly because of my mom and dad, who knows politics way better than me”. Lukas explained “I figured it would end the way it did, and I voted for the Liberal Party because of my grandmother, she is a very strong supporter of that party and the rest of my family is too, although everyone on my father's side is voting for the Conservative Party... …so I knew beforehand, I will either vote for the Conservative Party or the Liberal Party, and they are basically the same, small differences”.

From Philips statement we learn that he views his parents as more experienced, which then makes it obvious that he should listen to their advice. However, and supporting my argument above, I believe that Philip did find the decision task complex initially, but because he did not have the motivation necessary for engaging in a cognitively demanding information process, he left it to his parents whom he figured had the expertise required to make a satisfying decision, which then in the second phase of the decision process made the decision task appear easy. When I asked Philip if he considered it important to familiarize with the political campaign before casting the ballot, he said “Not really no… because I had already decided whom I was voting for”. This suggests that Philip’s choice was fully influenced by his parents, there was no further effort trying to gather supplementary information that could support his choice. His parent’s advice was enough. This is a typical example of how a first-

56 time voter, not particularly interested in the decision task, nor in which party that will win, relying on the endorsement heuristic, his parent’s advice, to make a decision.

From Lukas’s statement we know that he thought the decision task was easy because for him, the were only two options. He would either vote for the Liberal Party, which was the common party amongst all of his family member on his mother’s side, or the Conservative Party, which was the common party on his father’s side. Again, I believe that the decision task was considered easy because he had already employed a decision heuristic, even before the “decision process” began. He has employed the decision heuristic editing/pruning and familiarity, which is when you eliminate or ignore otherwise relevant aspects of the decision task, resulting in him only considering the two parties that he was familiar with. According to decision theory, when there are only two options, this should result in an employment of a compensatory decision strategy (Biggs, Bedard, Gaber, & Linsmeier, 1985), however, Lau and Redlawsk argue that if the alternatives are considered so similar that it really does not matter what you choose, a noncompensatory decision strategy is understandable (Lau & Redlawsk, 2006, p. 39). We know that Lukas did consider the two alternatives as basically the same, which then could explain why he employed the heuristics-based decision-strategy. Lukas used is grandmother as his final decision heuristic. He explained that he did not really talk about politics at home with his parents, it was mostly through conversation with his grandmother that he obtained information about the election, and she was very eager to explain why she voted for the Liberals, and the issues she was passionate about. Because Lukas did not really express an interest in the decision topic, all he needed was a little push in one direction. This “push” could have come from his father, who he knew votes for the Conservatives, but because he did not really have a proper conversation with him, giving Lukas reasons to vote for that party, he relied on the cues given to him from his grandmother.

Oskar and Emil thought the decision task was difficult. Oskar managed to decide with the help from his parents. Emil used the familiarity heuristic. Oskar said “It was difficult… because some parties have very similar politics… and it all made me uncertain, but then I talked to my mom and dad and they helped me with that stuff. Because I know who they voted for, so I kinda went in the same direction”. Emil explained “It was difficult because I felt that all of them had very good arguments for why you should vote for them, and it was difficult to know what you should prioritize”. He ended up voting for the Labor party, saying “I was actually really tired of the whole election process, and I didn’t want to think more about it, so I concluded that I would rather vote for the Labor Party than the Conservatives.”

57 It became evident that he had some preexisting sentiments towards the Labor Party, he explained that when he was younger, he really liked Jens Stoltenberg. He did not explain why, but it could explain why he chose the Labor party; when he was faced with the decision that for him was between the Conservatives and Labor, the Labor Party was the party he was most familiar with and remembered that he once liked. This suggests that Emil’s final decision was based on a combination of the affect referral heuristic and the familiarity, voting for the party you know the best, and the party you like the best.

Oskar, compared to Philip who were honest about the fact that his vote was solely a result of his parents’ influence, did multiple times during the interview try to convince me that he, in addition to the influence from his parents, had other reasons for why he voted for the Conservatives, all of which ended up being not very convincing. When I asked Oskar if he tried to familiarize with the parties’ politics, he responded by saying:

“ Yes, that thing with Syria, or not Syria, but the refugee crises, that thing about we taking in 10.000 and all that. And then it was that thing with the absence-limit. Or wait, what am I saying, the absence-limit is not that important. It’s only the thing with the refugees. I didn’t care about the absence-limit to be honest”.

When I asked where he obtained information on this topic he said “the papers and such, and then I heard about the Labor Party, was is not the Labor Party that wanted to take in many more refugees, increase the limit or something like that? I actually don’t remember”. He also mentioned an issue with his municipality, Bærum, being a rich municipality, they are taxed higher than other less wealthy municipalities, and concluded the argument by saying “I don’t remember completely how that stuff is connected again, but I have heard about that stuff before, and that is quite important”. It seems like he wants me to see him in a certain light, he wants me to know that his decision was based on actual political issues that he cares about, only he does not combine these statements into a convincing argument leading up to his decision choice. He cannot fully explain why he care about these issues, but he has heard about them, and he knows that they are important, even if he does not understand in what way they are so. It could of course be that the time-lag between the election and the interview has affected his memory, or it could be a classic example of listening with one ear to one’s parents’ political rants around the dinner table. He knows that there is something about Syria, and there is something about the taxation level in his municipality, but he just do not care enough, yet, to really familiarize with the debate. Why he spends so much time trying to

58 convince me of his political reasons and attitudes could of course be explained by social desirability. He wants me to perceive him as a guy who do not just vote the same as his parents, but a guy who has viewpoints and opinions on difficult political topics, although I believe that his decision is first and foremost a product of his parents influence, which he later in the interview confirms by saying “a lot was explained to me by my parents, because they know this kinda stuff, so they did of course influence me”.

Maja also expressed that she thought the decision task was difficult, but decided to just “follow the herd”. She explained:

“It was difficult because there are so many different alternatives, and the problem is that I agree with many parties but then, how can I say, for example the Conservatives aren’t they in favor of shooting wolfs? I don’t remember. Anyway, it is these kinds of small issues that are very important, and kinda ruins it for the party. So, it was difficult in that sense, but after a while I was just like “you know what, I’m just gonna listen to what everybody else is saying and go with the flow” which was stupid, but whatever.”.

She also said that she is very good friends with the boys in her class “and they are very oriented towards the Conservatives and the Progress Party, and when I hear them talk about it, you get convinced…”. I would argue that this is a classic example of a noncompensatory decision strategy where Maja, learning something about the wolfs being shot, avoids this whole conflict by taking the easy way out, employing the decision heuristic endorsement, and decides to just follow her friends’ recommendations. To employ heuristics in situations where one does not have the motivation or competence necessary to make an independent decision, does not always entail bad decisions. However, Maja was one of those who in retrospect regretted her decision, she did not really prefer the Conservatives she says, but the decision was made quickly and without much consideration. This could be an example where the employment of heuristics did not result in the most satisfying outcome.

Oliva considered the decision task easy, because she thought Erna Solberg was a good prime minister. Although she did admit that the whole environmental aspect made the process a little more complicated. Olivia was very honest and open about the fact that she was not that interested in politics, resulting in her employing several decision heuristics. First, she used editing/pruning which is when you eliminate or ignore otherwise relevant aspects of the

59 decision, boiling the alternatives down to only two parties; the Greens and the Conservatives. The Conservatives because she thought they had done a good job while sitting in government, and the Greens because she was aware of the fact that a healthy environment is crucial for our existence. She said “I was a little unsure of who to vote for, I think that it is the environment that allows for our existence in a way, so we all should actually vote for the Greens, but then I voted for the Conservatives”. She did not explicitly say why she thought that the Conservatives had done a good job, however this could of course be explained by the social and political context of her upbringing. The Conservative party has had the prime minister since she was 14/15 years, which means that through most of her formative years, the country has been ruled by the right-wing parties, which most likely has been considered a positive thing in her social environment. Growing up, supposedly hearing only good things about the governing party crates a basis, and becomes her anchoring and adjustment heuristic, affecting how she evaluates and process information about the political parties. She expresses concern for the environment, but because she lacks a general political interest, she might not feel that she has the political competence needed to make more of this complex issue. In addition, the Greens have a particular bad reputation amongst the supporters of the right, often saying that their politic is “naïve” and unrealistic. Hearing this type of statements from older and more “competent” and experiences people might confuse someone who has not yet learned how to organize and evaluate different political opinions. Finding yourself divided on this complex and sensitive issue, without any means or interest in sorting it out, may then result in avoidance of the whole topic, landing on a safe and familiar choice: the Conservative party. She used the decision heuristic affect referral, voted for the party she liked the most. Olivia said that “Erna is very good and nice, and I feel that the others are more harsh in the way they speak and how they are as persons”, this she expressed several times during the interview, and illustrates her feelings towards Erna, without any more explicit references to the party’s politics. She chose the Conservatives because she just likes Erna the most.

4.5 Strategy 5: Gut decision-making strategy The gut decision-making strategy was the fifth strategy identified by Lau and his colleagues in their follow-up article from 2018. Voters employing this strategy make decisions in affect, based on their gut-feeling, and involves no deliberate external search for information. Thus, the information search is categorized as shallow and noncomparable. Decisions are based on how they make the decision-maker “feel”. As discussed before, The Norwegian high school context of my respondents made it difficult to measure the exact level of information search.

60 Because the high schools have such a big focus on the election, there was really no one that based their decision purely on their gut-feeling. They all had a minimum level of information. Thus, this category will probably become more relevant when voting later in life, when there is no one that “forces” you to familiarize with the political campaign and the parties.

61 5 Discussion and concluding remarks

The main goal of this thesis is to contribute to a more comprehensive understanding of voter behavior by studying and identifying decision processes and strategies among first-time voters in Norway. I argue that the traditional way of explaining voter behavior, using issue opinions and political values, might not give the whole picture when looking at first-time voters, or other voters that have not yet developed a stance on many of these issues. Thus, I approach this electoral group in a more process-oriented way, focusing not only on the decision, but more on the cognitive processes leading up to the decision. I ask “what type of heuristics or decision strategies do people employ when deciding on their vote?”, with a special emphasis on first-time voters.

The framework and the five decision strategies proposed by Lau and Redlawsk is used as this study’s point of departure. By combining elements from modern psychology and behavioral decision theory with the traditional antecedents of the vote, Lau and Redlawsk were able to identify different decision strategies employed by US voters. I expected, based on decision- making theory, that there were certain aspects about the political system in Norway, and the high school contexts of my respondents that would influence the decision strategies employed here. I conducted semi-structured interviews with first-time voters from three different high schools, which I later categorized within the preexisting strategies developed by Lau and Redlawsk.

I identify four of the five decision strategies, and I would in general argue that the strategies proposed by Lau and Redlawsk was to a large extent compatible with the strategies used by respondents in my sample. In this chapter I begin with a short review of the general observations from the analysis of the strategies, followed by a more detailed discussion on some of the more fundamental discoveries, and the implications if these. I will conclude by proposing suggestions for future research

5.1 Do first-time voters in Norway use the same decisions strategies as general voters in the US?

62 The immediate answer to this question is in many ways yes. I identify four of the five decision strategies proposed by Redlawsk and Lau. I did not detect the employment of strategy five because of the high school context of the respondents. The potential implications of this will be discussed later. In the following, a review and a discussion of the compatibleness of the strategies by Lau and Redlawsk and the strategies detected in this study.

Strategy 1: Rational choice decision-making. I identified two respondents as having employed the rational choice decision strategy. I do not believe that the respondents really considered all the parties, which mean that not all of the original criteria described by Lau and Redlawsk were met. I argue, however, that the most important criteria were met. What characterized these voters was that they perceived the decision task important and that they did not want their vote choice to be influenced by others. They were determined to make an independent decision. This motivated them to familiarize with the different parties and they processed that information, weighing the alternatives and the different issues up against each other. Their choice was ultimately based on the acquired information, voting for the party they felt corresponded best with their own values and beliefs. Also, in contrast to those who employed strategy 2 which also went through a more cognitively demanding decision process than the rest of the respondents, these respondents did not express any preexisting party affiliation. Thus, even if they did not actually consider all parties, they did not belong to one political party or side, which made the information search and processing more evenhandedly applied to multiple parties.

Also, I did not find the expected effect of political sophistication on decision strategy. Lau and Redlawsk argued that political experts would spend less time on information processing; because they have made so many decisions within their field of expertise, decision-making is easier for them compared to novices. This interpretation of political expertise was not valid when looking at first-time voters. Even if I did not have the same measure of political expertise as Lau and Redlawsk, I argue that I am able to distinguish between the level of political sophistication within the sample. I find that those who were more politically aware were also those who spent the most time on information processing, which is the opposite of what was hypothesized by Lau and Redlawsk. Hence, I would argue, based on the findings from this study, that first-time voters who express an interest in- and a perceived importance of the decision task are also those who will spend more time on thorough information processing and search. The effect of political sophistication might prove relevant after some

63 years of voting, when one has learned the stance of many of the parties and know how they will position themselves on new issues. However, in my sample of first-time voters, I found that the more political aware and the more interest in the decision task the more likely is it that the voter will employ a compensatory decision strategy.

Strategy 2: Confirmatory decision strategy. I also identify this strategy among the respondents on the sample. Those I categorize as having employed the confirmatory decision strategy all expressed that they had a preexisting party orientation, or a predisposition towards a specific political side, which resulted in a clearly biased information search. However, because they were first-time voters, all expressed that they had experienced uncertainties during the decision process. From the interviews I understand that their political orientation had not yet been fully been internalized, and none would go as far as to say that they had a “party identification”. The decision process among these first-time voters were probably more cognitively demanding than those who employ this strategy later in life, when their political experience and competence on their party has increased. I argue that the confirmation decision strategy employed by first-time voter is closer to a compensatory strategy, and that it later in life will become more noncompensatory, and requires less cognitive effort.

One main difference within this category was that I categorize Sara as employing a confirmation decision strategy, although her political orientation was towards one political side rather than one party, which was the initial criteria described by Lau and Redlawsk. Sara’s affiliation towards the left, she considered four parties, resulted in a decision process that required more information search and cognitive effort than those who identified with only one party. Nevertheless, the decision process Sara described is very similar to the process described by Lau and Redlawsk, only that it includes a little more comparability. Her information search was limited, including only the parties she considered relevant; she said that she only followed the parties she was actually interested in on Facebook, and she had discussed the election only with her like-minded friends. I believe there are many like Sara who identify with one political side, thus automatically eliminating many alternatives even before the official decision process, but then employs another decision strategy in order to make the final decision. This two-step decision strategy will be discussed in greater detail in the section 5.3, because I believe that this is one of the major differences between the decision strategies employed in the US and in Norway, and potentially in other countries with a similar multi-party system.

64 Strategy 3: Fast and Frugal decision strategy. There was only one person who used a fast and frugal decision-making strategy. She was average interested in politics, but thought the whole process was somewhat overwhelming and confusing. Nevertheless, she wanted to make an informed decision, so she chose to focus on one diagnostic issue, the environment. I argued that the difference between the strategy identified here and the one described by Lau and Redlawsk, is the motivation behind the decision strategy. Lau and Redlawsk described the typical employer of the fast and frugal strategy as someone who do not have time to familiarize with the election campaign, so they just chose the candidate that they believe have the best policy on one diagnostic issue they care about. They describe the information search as shallow, although somewhat comparable. In this case I believe that the main motivation was not time constraint per se, rather a feeling of incompetence and being overwhelmed by the whole process. She did not feel that she would have been able to familiarize with the whole political campaign and all the parties, but she still wanted to make a decision she could stand by. Her information process consisted of her familiarizing with the parties she had perceived as caring about the environment, which was the topic she chose as her diagnostic criteria. Because of the many parties that run for parliament in Norway, I would argue that even though she chose to focus on one issue, the information process she went through was more complex than the one described by Lau and Redlawsk. In addition, I do not believe like the name of the strategy, fast and frugal, is an appropriate description of the decision strategy employed. I would rather call it a “pragmatic decision strategy”. Pragmatic in the way that she was aware of her shortcomings when it came to political competence and knowledge, and she did not have the motivation necessary for doing the whole “rational choice” information search, thus decides to focus one issue she cares about, making the decision task more manageable.

Strategy 4. This strategy, the heuristics-based, was the most commonly employed by the respondents in this sample. What characterizes these voters was that they all lacked an interest in the decision task, thus the motivation for spending time on information processing was low. These voters just wanted to make a quick decision. However, I found variance in the expressed reasons the respondents gave for employing the heuristics-based strategy. Some said that they thought it was very difficult to deicide because there were so many parties and so much information to process and consider, and they did not feel that they were competent enough to make the decision on their own, nor did they have the motivation necessary to do so, thus ended up relying on other, more experienced people’s opinions such as family or peers that they considered to be more politically sophisticated. Others thought it was easy

65 because they knew that they were just going to vote the same as their parents, there were no need to dwell more on that. What was interesting was that almost all of those who I categorized as having employed the heuristic-based decision-making strategy expressed that they had not been that interested in the decision-task at the time of the election, but that their interest and the perceived importance of the decision task and its consequences had since increased. There were also those who explicitly said that they regretted their decision, and that they wished they had taken it more seriously.

As discussed, there were many who explained that they indeed had been participating in many election-related activities arranged by the high schools, thus the majority of the information they had was due to these high school evets. However, there were those who felt like all of these activities had happened within a very short period of time, and very close up to the actual election, thus they had not been able to reflect and process all of the information they had received before they had to make a choice. Maybe, if those who ended up using the heuristics-based decision strategy, because they felt overwhelmed by all of the information and the parties, had gotten more time to process and reflect, and then they might have been able to make a decision they in retrospect would be more satisfied with. This could be an argument for including more political information earlier in the school year. I do not believe that this will eventually mean that every first-time voter will employ a rational choice decision strategy, for example I would assume that those who said they though the decision task was easy because they either way was going to vote the same as their parents, would still use their parents as a decision heuristic. But I do believe that those who explained that their chosen decision strategy was a result of incompetence, these voters would probably benefit from more information earlier on, and perhaps a slower introduction to the election process and the parties. I elaborate further on the implications of this in section 5.4.

5.2 General findings when testing the framework by Lau and Redlawsk

Perceived importance of decision task better predictor than information search

One general finding was that I discovered that the level of interest in the decision topic was a better predictor for decision strategy than information search, at least in my case. Lau and Redlawsk used information search, the depth and comparability of search, as their main

66 indicator for which type of decision strategy a voter employed14. Because of the high school context of the respondents in this study, it proved difficult to measure the precise effect of this factor, because all of the students had been through a lot of election related activities before voting. This meant that even those who otherwise would not have bothered familiarizing with the election campaign and the parties, had a minimum insight on the election process and the different parties. In this case, because they had comparable, and potentially deep information about many political parties and issues, this should then, according to the criteria stated by Lau and Redlawsk, mean that they all employed a compensatory decision strategy, which we know that most of them did not. I would argue that there is a difference between the information you have and the information you actually process and consider. Even if you have information about many parties, this do not mean that all these parties will be perceived as potential candidates for one’s vote. Thus, I found that the questions that measured the perceived importance of decision topic15, to be a better indicator for the decision strategy used than what type, and how much information search they had done. Those who expressed an interest in the decision task employed compensatory strategies in a larger extent than those who did not care about the decision task.

Another interesting finding was the effects of complexity of task. According to Tversky (1972), the more complex a decision task is, the more likely it is that a decision-maker will employ a noncompensatory decisions strategy. If there are more than two alternatives, the chances are that a decision maker will need to simplify the decision task by using cognitive short-cut like heuristics, resulting in a noncompensatory strategy. I argu that the political party system in Norway makes the decision task of voting in the national elections objectively complex. Thus, I expected that many would employ a noncompensatory strategy, and my expectations was confirmed. Almost all of the respondents employed at some point of the decision process a version of a noncompensatory strategy, simplifying the decision task by automatically eliminating many parties, considering to two or three parties, or focusing on only one issue. Even if some relied more on heuristics than others, there were few who expressed that they had considered and researched all of the parties and all of the potential outcomes. When I explicitly asked them whether they thought it was difficult or easy to

14 Because they used experiments when studying decision strategies, they could track the amount of time the participants spent on information processing, they could also see what type of information they acquired (they could see if they only read about the republican candidate or if they read about both candidates). 15 The wording of the questions measuring perceived importance of the decisions task: How interested in politics would you say that you in general is? Did you care about which party or parties that won the election? Was it important for you, personally, to familiarize with the political campaign before voting? The wording of the questions measuring perceived complexity of the decisions task: Did you consider it easy or difficult to decide who to vote for? with a follow-up questions fitted to their answers.

67 decide on their vote, I received different answers. There were many who said it was easy, who I still identified as having employed a noncompensatory strategy. However, it became evident that many thought the decision task was easy because they had employed a noncompensatory strategy. I believe that this is common for almost all the respondents, and this leads me to a discussion on what I believe is the most unique with the decision strategies detected in this study, namely the two-step decision-making procsess.

5.3 Two-step decision process and the concept of “party set”

I argued that Sara, who I categorized as employing strategy 2, went through a two-step decision process. Sara expressed that she considered herself as typically left-oriented, which made it easy for her to eliminate many of the alternatives. However, she thought it was difficult to decide among the remaining alternatives. Her predisposition towards the left-wing parties made her automatically exclude many of the parties and resulted in a biased information search. Thus, the first step in her decision process, was her employing the confirmatory decision strategy. However, because she did not favor one particular party, she did an equal and thorough information search among the alternatives she considered relevant. Thus, in the second step of the decision process she employed a “constrained” version of the rational choice decision strategy. Based on the interviews I believe many of the respondents went through such a two-step decision process, although not as consciously as in the case with Sara. Sara was explicit about the parties she considered relevant. These parties were relevant because of her left-wing orientation.

Even if most of the respondents did not explicitly express a clear orientation towards a political party or side, it is clear that most did have a predefined conception of which parties that was potential candidates for their vote. When asked if there were any other parties, besides the party they ended up voting for, that they had considered, all of the respondents mentioned at least one, and some as many as three other parties. First of all, the parties that most respondents voted for were the Conservative Party and the Liberal Party. 10 out of 13 voted for either one of these two parties. This did not come as a surprise. As I discussed in chapter 3 (see section 3.1.3), the respondents came from high schools located in areas known to be politically center-right oriented, and we know from the analysis that the majority of the respondents based their vote decision on heuristic cues from family or friends, which then explains why so many voted for the center-right parties. In table 5.1, we see what other parties

68 the respondents also considered voting for. For example, three out of four respondents who voted for the Liberal Party did also consider voting for the Conservative party, and two out of six of those who voted for the Conservatives did also consider the Greens. Even if most of these voters do not have an explicit political orientation, as was the case with Sara, which mean that they do not verbally explain why they consider the parties that they did, most still have a clear idea of a set of parties they perceive as relevant for consideration.

Table 5.1 Which others parties the voters considered voting for.

The The The Liberal The The The The Christian The The Progress Conservative Party Labor Social Green Democratic Center Red Party Party Party Left Party Party Party Party Party

Voted for the 2 - 1 2 - 2 - - - Conservative Party (N=6) Voted for the - 3 1 1 - 1 - - - Liberal Party (N=4) Voted for the - - - 1 - 1 - - - Social Left Party (N=1) Voted for Labor - 1 1 ------(N=2)

This concept is known as a voter’s “party set”, first introduced by Oscarsson, Gilljam, and Granberg (1997) in their article “The Concept of Party Set -A Viable Approach or Just Another Way to Slice the Same Cheese?”. This concept was introduced as an alternative to the American party identification model, a model that was originally designed for two-party systems (Oscarsson et al., 1997, p. 17). Oscarsson and his colleagues argue that in a multi- party context, like the one we have in Norway, voters can have more than one party that they consider to be likely or probable alternatives. They define “party set” as the following:

«A voter’s party set consists of only those parties that the voter would really consider voting for at a given time. It is from the parties in the party set the person chooses when voting.» -Oscarsson et al. (1997)

69 Based on the interviews in this study, the concept of party set seems to be present among all the respondents. None of the respondents in the sample expressed that they had a strong affiliation towards one party, there were rather several different parties that they considered as potential candidates. Because of how the political system in Norway is structured, I suggest that the decision process among Norwegian voters, and perhaps voters in other multi-party systems, will almost always happen in a two steps. The first step of the process is where the choice set, or party set is defined. This could happen more or less consciously, depending on the political awareness and engagement of the voter. Drawing from earlier presented theory, I believe that which party that are included in a voter’s party set is initially influenced by the judgement heuristic adjusting and anchoring. It is the anchor (formed by the socio-political context of one’s upbringing) that influence the party set, in other word which parties you perceive as potential candidates for your vote. This anchor creates an implicitly suggested reference point, which all further evaluation of alternatives is colored by.

To sum up, from the interviews it became clear that all of the respondents had a party set, a set of parties they considered as potential candidates. It is the more or less conscious development of this party set that constitute step one in the decision-process. The second step in the process is however different from respondent to respondent, depending on the perceived importance of the decision task. I believe this is one of the biggest differences between the decision processes described by Lau and Redlawsk and the decision processes identified in this study. The decision strategies proposed by Lau and Redlawsk was developed within the American two-party system, whereas the strategies employed in this study is clearly affected by the Norwegian multi-party structure, resulting in a two-step decision strategy. Thus, to truly understand how the vote decision is made within a multi-party system, we need to obtain information, not only the last decision strategy employed, but also on the first. This could of course prove difficult, especially in instances where this first step of eliminating alternatives have happened rather unconsciously, but it is still an important aspect to remember when wanting to obtain the full picture of the decision process. If I were to conduct another study like this in the future, I would have asked more about the first step of the decision strategy, and made the respondents elaborate and reflect more on why they considered the parties that they did.

70 5.4 No “gut decision making” detected among first-time voters; a case for voting rights for 16-year-old’s?

I did not identify any instances of the gut decision-making strategy. This, I argued, was because all of the respondents had, through mandatory events and assignments arranged by the schools, obtained a minimum level of information about the parties and the election process in general. Even if I detected a variance within the level of actual processed information, all respondents expressed that they had obtained some information ahead of the election day. When I asked if they had familiarized with the parties and the political issues, almost all of them mentioned the mock elections, the debates with the political parties’ youth leaders, a roleplay, and general discussions in class, as sources of information. This suggests that even those who initially were not that interested in the decision task, still had a basic understanding of how the political system worked and what parties that were running for parliament and what these parties stood for. I would argue that the fact that I did not identify anyone who had just followed their gut when deciding in their vote is a positive finding looking at it from a democratically ideal perspective. Without entering a normative debate, I would think that it is safe to say that if democracy is the ideal political system then vote decisions that reflect careful evaluation and consideration is better than decisions that are based on nothing more than your gut. Thus, the fact that these voters were attending high school when voting for the first time, I believe contributed to minimizing instances of gut- decision making, which is preferable when wanting our democracy to reflect actual political attitudes and values.

This finding also has implications for the debate about voting rights for 16-year-old’s. Based on the interviews, I have argued that the focus the high schools has on informing and preparing the first-time voters for their first election does influence the decision strategies employed. Even if there were many who expressed that they felt that they did not have the competence necessary nor the motivation for doing the rational choice decision process, the school’s election-related activates did at least limit the instances of pure gut decision making. If this were to reflect a more common phenomena, this would suggest that all first-time voters should attend high school when voting for the first time, so at to make sure that all first-time voters obtain some basic information about the election process and on the parties and the most important issues. Today, this is not the case. Today, only those who are born in an uneven year of birth are participating high school when voting for the first time, those who are born in an even year of birth are 19 when they finish, thus has not been through the same

71 introductory process as those who are 18 when voting. We know that the turn-out among 18 year-olds is higher than those who are 19 and up to 22 years old (Bergh, 2013). This could be explained by the fact that those who are 18 years old when voting for the first time, still live at home with their parents which could mean that they are influenced and encouraged to participate by their parents. Or it could be an effect of all the effort the high school put in on education the first-time voters on the election process and the parties. Based on the finding from this study, I would argue that it in general would be better if all first-time voters had been “forced” to participate in some election-related activities before casting a ballot. By introducing voting for 16-year-olds, there would be a guarantee that at least those first-time voters who attend high school have been introduced to the basic concepts of political participation and has obtained at least a minimal information about the political parties and their policies.

Introducing the concept of voting and familiarization of politics at an earlier stage, could prove have at least three positive implications. First, it could contribute to an increased election turn out both among first-time voters in general knowing that those who attend high school participate to a higher degree than those who has finished high school when getting voting rights. Second, it might also prove to have even longer lasting consequences when we know, as I stated in the introduction, that voting is habit forming and participation in one election increases the probability of participation in future elections (Gerber et al., 2003). And third, it could limited the instance of gut-decision making, which I believe is a democratic ideal.

5.5 Future research

I believe in the possibility of creating a voting strategy measure which could be quantitatively tested and used to validate the categorizations identified here. This has been successfully done by Lau et al. 2018 within the context of US. I would propose a similar lay-out, although the measure would have to incorporate the characteristics of the strategies found here, with an emphasis on the two-step decision strategy. One would have to include questions aimed at detecting the first step of the decision process, the development of party set, as well as the second step, which is when the final decision is made. Lau et al. (2018) created a self-report scale that measured the five different political decision-making styles, and based on the collected data ran a confirmatory factor analysis that supported their theoretical expectations.

72 I believe that it would be possible to conduct similar studies in Norway, for example by including some batteries of questions specifically developed for measuring decision strategy within the NNES. I believe that by incorporating element aimed at detecting decision strategies, this would contribute to enhance our overall understanding of voter behavior, not only amongst first-time voters but among voters in general.

73 References

Anderson, J. R. (1983). The architecture of cognition (Vol. 5). Cambridge, Mass: Harvard University Press. Anderson, J. R. (2002). Spanning seven orders of magnitude: a challenge for cognitive modeling. Cognitive Science, 26(1), 85-112. doi:10.1207/s15516709cog2601_3 Berger, C. C., Dennehy, T. C., Bargh, J. A., & Morsella, E. (2016). Nisbett and Wilson (1977) revisited: the little that we can know and can tell.(Richard E. Nisbett)(Thomas D. Wilson). Social Cognition, 34(3), 167-195. doi:1999/xlink" type="doi"> Bergh, J. (2013). Valgdeltakelse i ulike aldersgrupper Historisk utvikling og oppdaterte tall fra stortingsvalget 2013. Retrieved from https://fido.nrk.no/bc1842136b3d79c3765daecbd3ba4db8dbd011735037f4e89e89bd1 da7004779/notat%20valgdeltakelse.pdf Berglund, F. (2007). Nye sosiale skiller: Sektor teller, ideologi avgjør. In B. Aardal (Ed.), Norske Velgere: en studie av stortingsvalget 2005. Oslo. Biggs, S. F., Bedard, J. C., Gaber, B. G., & Linsmeier, T. J. (1985). The Effects of Task Size and Similarity on the Decision Behavior of Bank Loan Officers. Management Science, 31(8), 970-987. doi:10.1287/mnsc.31.8.970 Billings, R. S., & Scherer, L. L. (1988). The effects of response mode and importance on decision-making strategies: Judgment versus choice. Organizational Behavior and Human Decision Processes, 41(1), 1-19. Retrieved from http://www.sciencedirect.com/science/article/pii/074959788890043X. doi:https://doi.org/10.1016/0749-5978(88)90043-X Campbell, A. (1960). The American voter. New York: Wiley. Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4(1), 55- 81. doi:10.1016/0010-0285(73)90004-2 Converse, P. (2006). The nature of belief systems in mass publics (1964). Critical Review, 18(1-3), 1-74. doi:10.1080/08913810608443650 Dalton, R. J. (2014). Citizen politics : public opinion and political parties in advanced industrial democracies (6th ed. ed.). Washington, D.C: CQ Press. Dalton, R. J. (2016). Party Identification and Its Implications. In: Oxford University Press. Dane, E., Rockmann, K. W., & Pratt, M. G. (2012). When Should I Trust My Gut? Linking Domain Expertise to Intuitive Decision-Making Effectiveness. Organizational Behavior and Human Decision Processes, 119(2), 187-194. doi:10.1016/j.obhdp.2012.07.009

74 Denny, K., & Doyle, O. (2008). Political Interest, Cognitive Ability and Personality: Determinants of Voter Turnout in Britain. Brit. J. Polit. Sci., 38(2), 291-310. doi:10.1017/S000712340800015X Einhorn, H. J. (1970). The use of nonlinear, noncompensatory models in decision making. Psychological Bulletin, 73(3), 221-230. doi:10.1037/h0028695 Fisher, R. (1993). Social desirability bias and the validity of indirect questioning. Journal of Consumer Research, 20(2), 303. doi:10.1086/209351 Galotti, K. M., Ciner, E., Altenbaumer, H. E., Geerts, H. J., Rupp, A., & Woulfe, J. (2006). Decision-making styles in a real-life decision: Choosing a college major. Personality and Individual Differences, 41(4), 629-639. Retrieved from http://www.sciencedirect.com/science/article/pii/S0191886906001206. doi:https://doi.org/10.1016/j.paid.2006.03.003 Gerber, A. S., Green, D. P., & Shachar, R. (2003). Voting May Be Habit‐Forming: Evidence from a Randomized Field Experiment. American Journal of Political Science, 47(3), 540-550. doi:10.1111/1540-5907.00038 Gigerenzer, G., & Gaissmaier, W. (2011). Heuristic Decision Making. Annu. Rev. Psychol., 62, 451-482. doi:10.1146/annurev-psych-120709-145346 Gigerenzer, G., & Goldstein. (1996). Reasoning the fast and frugal way: Models of bounded rationality. Psychological Review, 103(4), 650-669. Retrieved from https://www.scopus.com/inward/record.uri?eid=2-s2.0- 0030266119&doi=10.1037%2f0033- 295X.103.4.650&partnerID=40&md5=cec8f94283f594bdbc2a8cf516abc432. doi:10.1037/0033-295X.103.4.650 Gigerenzer, G., & Todd, P. M. (1999). Fast and frugal heuristics: The adaptive toolbox. In Simple heuristics that make us smart. (pp. 3-34). New York, NY, US: Oxford University Press. Goldstein, D. G., & Gigerenzer, G. (2009). Fast and frugal forecasting. International Journal of Forecasting, 25(4), 760-772. Retrieved from http://www.sciencedirect.com/science/article/pii/S0169207009000818. doi:https://doi.org/10.1016/j.ijforecast.2009.05.010 Green, D., & Shachar, R. (2000). Habit formation and political behaviour: Evidence of consuetude in voter turnout. Br. J. Polit. Sci., 30, 561-573. Guba, E. G., & Lincoln, Y. S. (1994). Competing paradigms in qualitative research. In Handbook of qualitative research. (pp. 105-117). Thousand Oaks, CA, US: Sage Publications, Inc.

75 Hanfmann, E. (1941). A study of personal patterns in an intellectual performance. Character & Personality; A Quarterly for Psychodiagnostic & Allied Studies, 9, 315-325. doi:10.1111/j.1467-6494.1941.tb02060.x Harren, V. A. (1979). A Model of Career Decision Making for College Students. Journal of Vocational Behavior, 14(2), 119-133. doi:10.1016/0001-8791(79)90065-4 Holmberg, S. (1994). Party Identification Compared Across the Atlantic. In I. M. K. M. Jennings, T. (Ed.), Elections at Home and Abroad. Essays in Honor of Warren Miller. Ann Arbor: The University of Michigan Press. Kirk, J., & Miller, M. L. (1986). Reliability and Validity in Qualitative Research. Retrieved from https://methods.sagepub.com/book/reliability-and-validity-in-qualitative- research doi:10.4135/9781412985659 Klein, G., & Schlesinger, H. J. (1951). PERCEPTUAL ATTITUDES TOWARD INSTABILITY: I. PREDICTION OF APPARENT MOVEMENT EXPERIENCES FROM RORSCHACH RESPONSES*. Journal of Personality, 19(3), 289-302. Retrieved from https://doi.org/10.1111/j.1467-6494.1951.tb01103.x. doi:10.1111/j.1467-6494.1951.tb01103.x Klein, G. S., & Krech, D. (1951). The problem of Personality and Its Theory*. Journal of Personality, 20(1), 2-23. Retrieved from https://doi.org/10.1111/j.1467- 6494.1951.tb01510.x. doi:10.1111/j.1467-6494.1951.tb01510.x Knutsen, O., & Thrasher, M. (2007). Class voting in Western Europe: a comparative longitudinal study. In O. Knutsen & M. Thrasher (Eds.), (Vol. 30, pp. 642-643). Kozhevnikov, M. (2007). Cognitive Styles in the Context of Modern Psychology: Toward an Integrated Framework of Cognitive Style. Psychological Bulletin, 133(3), 464-481. doi:10.1037/0033-2909.133.3.464 Lau, R. R., Kleinberg, M. S., & Ditonto, T. M. (2018). MEASURING VOTER DECISION STRATEGIES IN POLITICAL BEHAVIOR AND PUBLIC OPINION RESEARCH. Public Opin. Q., 82(S1), 325-350. doi:10.1093/poq/nfy004 Lau, R. R., & Redlawsk, D. P. (2006). How voters decide : information processing during election campaigns. Cambridge: Cambridge University Press. Miller, G. A. (1956). The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychological Review, 63(2), 81-97. doi:10.1037/h0043158 Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review, 84(3), 231-259.

76 Oscarsson, H., Gilljam, M., & Granberg, D. (1997). The Concept of Party Set - A Viable Approach or Just Another Way to Slice the Same Cheese? . Working paper presented at the annual meeting of the Swedish Political Science Association in Uppsala, October 5-7, 1997 Payne, J. W. (1993). The adaptive decision maker. In J. R. Bettman & E. J. Johnson (Eds.). Plous, S. (1993). The psychology of judgment and decision making. New York, NY, England: Mcgraw-Hill Book Company. Rokkan, S., & Lipset, S. M. (1967). Party systems and voter alignments : cross-national perspectives (Vol. 7). New York: Free Press. Shachar, R. (2003). Party loyalty as habit formation. Journal of Applied Econometrics, 18(3), 251-269. doi:10.1002/jae.698 Shanteau, J. (1992). How much information does an expert use? Is it relevant? Acta Psychologica, 81(1), 75-86. doi:https://doi.org/10.1016/0001-6918(92)90012-3 Simon, H. A. (1957). Models of man : social and rational : mathematical essays on rational human behavior in a social setting. New York: Wiley. Simon, H. A. (1989). The Scientist as Problem Solver. In I. Carnegie-Mellon Univ Pittsburgh Pa Artificial & P. Psychology (Eds.). Thagaard, T. (2009). Systematikk og innlevelse: en innføring i kvalitativ metode (3rd ed.). Bergen: Fagbokforlaget. Tversky, A. (1972). Elimination by aspects: A theory of choice. Psychological Review, 79(4), 281-299. doi:10.1037/h0032955 Tversky, A., & Kahneman, D. (1974). Judgment under Uncertainty: Heuristics and Biases. Science, 185(4157), 1124. Retrieved from http://science.sciencemag.org/content/185/4157/1124.abstract. doi:10.1126/science.185.4157.1124 Universitas. (2015, 11.09.2015). Usikker på hva du skal stemme? -Vurder sofaen. Universitas. Retrieved from https://universitas.no/sak/60807/vurder-sofaen/ Witkin, H. A. (1950). INDIVIDUAL DIFFERENCES IN EASE OF PERCEPTION OF EMBEDDED FIGURES*. Journal of Personality, 19(1), 1-15. Retrieved from https://doi.org/10.1111/j.1467-6494.1950.tb01084.x. doi:10.1111/j.1467- 6494.1950.tb01084.x Zaller, J. (1992). The nature and origins of mass opinionThe Nature & Origins of Mass Opinion.

77 Ødegård, G. (2003). Samfunnsengasjement blant dagens unge ; nye tall om ungdom. Tidsskrift for ungdomsforskning, 3(2), 89-98. Aardal, B. (2017). 60 år med valgforskning – hva har vi lært? Aardal, B. (2019). www.aardal.info. Aardal, B., & Bergh, J. (2015). Valg og velgere : en studie av stortingsvalget 2013. Oslo: Cappelen Damm akademisk. http://samfunnsveven.no/skolevalg/resultat/skole/10030 (10.10.18. Kl.10:02) http://samfunnsveven.no/skolevalg/resultat/skole/10025 (10.10.18. Kl.09:10) https://valgresultat.no/akershus/b%C3%A6rum?type=st&year=2017 (10.10.18. Kl.11:04) https://universitas.no/sak/60807/vurder-sofaen/ (17.01.19. kl.19:02)

78 Appendix

A.3.1 Overview of interview

Interviewee Status Source Format Length Recording Transcript School 1: Stabekk Lukas 11.06.18 Through a Semi- 07:20 Audio Yes colleague structured recorded Sara 11.06.18 Through a Semi- 06:49 Audio Yes colleague structured recorded Jakob 11.06.18 Through a Semi- 07:43 Audio Yes colleague structured recorded Emma 11.06.18 Through a Semi- 09:07 Audio Yes colleague structured recorded Oskar 11.06.18 Through a Semi- 09:40 Audio Yes colleague structured recorded School 2: Eikeli Emil 07.06.18 Through a Semi- 06:27 Audio Yes colleague structured recorded Ella 07.06.18 Through a Semi- 07:37 Audio Yes colleague structured recorded Philip 07.06.18 Through a Semi- 07:23 Audio Yes colleague structured recorded Maja 07.06.18 Through a Semi- 09:46 Audio Yes colleague structured recorded Olivia 07.06.18 Through a Semi- 07:18 Audio Yes colleague structured recorded Sofie 07.06.18 Through a Semi- 10:27 Audio Yes colleague structured recorded Nora 07.06.18 Through a Semi- 07:37 Audio Yes colleague structured recorded School 3: Response Handlesgym from email Elias 08.05.18 Semi- 09:52 Audio Yes structured recorded

79

A.3.2 Interview guide

Dina Heider Hov

Navn på skolen: Navn på intervjuobjekt: E-post: Dato: Kjønn: Alder:

Introduksjon Først vil jeg takke deg for at du tar deg tid til å snakke med meg. Som jeg tidligere har nevnt på bla. e-post, er jeg masterstudent ved Universitetet i Oslo, og i masteroppgaven min ønsker å undersøke hvordan førstegangsvelgere tenker når de skal bestemme seg for parti.

Før vi begynner vil jeg bare gjøre det klart at alt som blir sagt her vil være konfidensielt, og svarene dine vil bli behandlet på en slik måte at de ikke kan knyttes til deg som person. Alt som diskuteres her skal kun brukes til vitenskapelige formål. Du står også fritt til å trekke deg når som helst i løpet av intervjuet. Hvis du fortsatt ønsker å delta, kan du signere dette arket (gir IO samtykkeerklæring), der står alt jeg har sagt til nå og litt til. Jeg vil også dobbeltsjekke om det er greit at jeg tar opp samtalen? Da er formalitetene unnagjort, og da setter jeg på båndopptakeren (HVIS IO HAR GITT KLARSIGNAL).

Politisk interesse

Ø Hvor interessert i politikk vil du si at du generelt er? o Hvorfor er du interessert? o Hvorfor ikke? Ø Brydde du deg om hvilket parti eller partier som vant valget? o Hvorfor var det viktig? o Hvorfor brydde du deg ikke?

80

Stemmegivning

Ø Synes du det var enkelt eller vanskelig å bestemme deg for hvilket parti du skulle stemme på?

Hvis IO uttrykker at det var enkelt: Ø Hvorfor var det lett?

Hvis IO uttrykker at det var vanskelig: Ø Hvorfor var det vanskelig?

Ø Hvilket parti ente du med å stemme på ved Stortingsvalget 2017?

Ø Hvis (NAVN PÅ PARTI) ikke hadde eksistert hva hadde du stemt på da? o Hvorfor?

Ø Når bestemte du deg for hva du skulle stemme?

Ø Stod det mellom to eller flere partier? o Hvilke partier stod det mellom?

Ø Hvordan kom du frem til at du ville stemme på (NAVN PÅ PARTI)? o Var det noen politiske saker som var avgjørende for din stemmegivning? o Vil du si at du har en partitilhørighet? § Hvor sterk eller svak er denne tilhørigheter? o Er du medlem av et parti? o Deltok du i valgkampen på noen måte?

81 o Hvis de tok en Valgomat: Tror du at du ble påvirket av valgomaten? o Spilte det noen rolle for din stemmegivning hvem som var leder for de ulike partiene? o Var samtaler med venner i forkant av valget en viktig faktor? o Leste du noe på Facebook, Twitter eller Instagram eller lignende som påvirket deg? o Var det noe annet som du tenker spilte en avgjørende rolle for hva du ente opp med å stemme på?

Ressurser brukt på valget

Ø I forkant av valget, prøvde du å sette deg inn partienes politikk og deres ståsteder i ulike saker?

Hvis IO sier ja: Ø Hvordan gikk du frem for å finne ut av dette? o Valgomat o Snakket med familie eller venner? o Partienes hjemmesider o Fulgte med på partidebattene o Leste aviser o (Andre forslag?)

Ø Var det viktig for deg å sette deg ordentlig inn i valgkampen før du skulle stemme?

Hvis IO sier nei: Ø Hvorfor ikke? o Gadd ikke o Ikke viktig o Vanskelig o Slitsomt o Vet alt fra før

Ø Synes du at du lyktes i å få en oversikt? Ø Var det noe du spesielt la merke til i valgkampen, noe som gjorde deg mer engasjert?

82

Hvis IO sier nei: Ø Hvorfor ikke? o Gadd ikke o Ikke viktig o Vanskelig o Slitsomt o Vet alt fra før

Ø Helt til slutt, er det noe du ønsker å legge til, noe du tenker kan være relevant?

Hvis IO ikke virker altfor utålmodig: o Hvordan synes du at intervjuet var?

Avslutning Da var vi ferdig, igjen tusen takk for at du kunne stille opp. Hvis du ønsker kan jeg sende deg oppgaven når den er ferdig?

83