Chew on This: Investigating Public Perceptions of Lab Grown Meat by Kellie Boykin, B.S. a Thesis in Agricultural Communications

Home , White meat

Chew on this: Investigating public perceptions of lab grown meat

Kellie Boykin, B.S.

A Thesis

Agricultural Communications

Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCES

Approved

Dr. Courtney Meyers Chair of Committee

Dr. Nan Li

Dr. Lindsay Kennedy

Mark Sheridan Dean of the Graduate School

December, 2019

ACKNOWLEDGEMENTS

“Immeasurably more.” This phrase from a verse in the New Testament has been constantly on my heart through this whole process of being a graduate student because

God has truly blessed me more than I could have ever imagined in my time here at Texas

Tech. Looking back, there are so many people that have supported me, loved me, and encouraged me through this process.

First and foremost, I have to thank my parents. Dad and Mom, words don’t touch my gratitude to you. You were my first teachers and have always encouraged me to learn and pursue education. I owe so much of where I am today to my late grandpa, Dennis.

You got me hooked on books and always encouraged me to write – a young communicator in the making. Grandpa, you were the reason that I love agriculture. You showed me the importance of caring for the land and how important it is to feed people.

Now here I am pursuing a degree in agricultural communications.

Kaitlyn, you’re not just my sister, you’ve been my best friend since the day you were born. Although this journey to Texas Tech took me further away from you physically, you have continued to be my most faithful friend and closest confidant, letting me tell you about my thesis every single day.

When I moved to Texas, it was the furthest I had ever been away from home. To my dear Matthew, Cathy, Frank, Jason, Anthony, LaFawnda, Brennon, and Gage: Being

500 miles from my family and all things familiar has not been easy, but you accepted me into your family with open arms and provided me with so much love and stability. I am so thankful to be a part of your family!

Texas Tech University, Kellie Anne Boykin, December 2019

Although distance has separated us, Terra, Lilly, Michaela, and Celsey, you have remained so close to my heart. Thank you for listening to me, letting me bounce ideas off of you, encouraging me, making me laugh, and being my adopted sisters.

Dr. Meyers, when I met you for the first time in Fort Collins, CO, I was an undergraduate student that felt quite lost about what my future should look like. You were the most welcoming, energetic, and genuine person I had ever met, and I instantly knew I wanted to come to Tech and prayed I could be under you. Since coming to Tech, your enthusiasm and grace has never wavered. You are someone I look up to so much and aspire to be like “when I grow up.” You have pushed me to be better, encouraged me, and guided me through this crazy concept of research. Thank you so much for making this project possible and supporting me through the whole process!

To Dr. Li, you saw this project evolve from just an idea in your class to fruition.

Thank you for your help in designing this project and allowing it to come to fruition.

When I got stuck, you always looked at things from a different perspective and propelled me forward.

Dr. Kennedy, thank you so much for doing the nitty, gritty job of editing my stimuli and document! They say the design is in the details, and you took the time to look at all of the details of my thesis, while always meeting me with encouragement about the entire process.

This thesis process has grown me and taught me much more than the information contained in this document. I have been challenged in so many ways and grown more than I thought possible. Each graduate experience is unique, and my singular experience is something I will always cherish.

iii

Texas Tech University, Kellie Anne Boykin, December 2019

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

ABSTRACT ...... vii

LIST OF TABLES ...... viii

LIST OF FIGURES ...... x

CHAPTER I INTRODUCTION ...... 1

Background & Setting ...... 1

Social Media & Science Communication ...... 5

Need for the Study ...... 7

Purpose and Research Questions ...... 8

Conceptual Framework ...... 9

Assumptions ...... 11

Limitations ...... 11

CHAPTER II LITERATURE REVIEW ...... 12

Overview ...... 12

Consumer Acceptance of Food Technology ...... 12

Consumer Acceptance of Lab Grown Meat ...... 14

Science Communication ...... 17

Conceptual Framework ...... 18

Sentiment Analysis ...... 18 Opinion Leadership ...... 19 Risk and Benefit Perceptions ...... 21 Measures of Uncertainty ...... 22

CHAPTER III METHODOLOGY ...... 25

Overview ...... 25 iv

Texas Tech University, Kellie Anne Boykin, December 2019

Part I ...... 25

Purpose and Research Objectives ...... 25 Research Design ...... 26 Sample ...... 26 Data Collection ...... 27 Data Analysis ...... 28

Part II ...... 28

Purpose and Research Questions ...... 28 Research Design ...... 29 Sample ...... 30 Instrumentation ...... 31 Social Media Usage ...... 31 Food Technology Neophobia ...... 32 Attitudes Toward Lab Grown Meat ...... 33 Message Stimuli ...... 34 Message Evaluation ...... 35 Risk Perceptions ...... 35 Measures of Uncertainty ...... 36 Intention to Share Content ...... 37 Intention to Consume Lab Grown Meat ...... 37 Benefit Perceptions ...... 38

Procedure ...... 39

Message Testing ...... 40

Data Analysis ...... 41

Participant Description ...... 41

CHAPTER IV RESULTS ...... 44

Overview ...... 44

Research Objectives for Part I ...... 44 Research Questions for Part II ...... 44

Part I Results ...... 45

RO 1: Determine the potential social reach and volume of lab grown meat on Twitter ...... 45 RO 2: Establish the percent of positive, negative, and neutral messages about lab grown meat on Twitter ...... 47

Texas Tech University, Kellie Anne Boykin, December 2019

RO 3: Identify the trending themes about lab grown meat on Twitter ...... 47 RO 4: Determine the top posters on Twitter providing content about lab grown meat ...... 48

Part II Results ...... 48

RQ 1: How does the frequency of social media use affect intention to share content and message evaluation? ...... 49 RQ 2: How is pre-existing attitude toward lab grown meat related to risk and benefit perceptions? ...... 50 RQ 3: How is neophobia to new food technology related to risk and benefit perceptions of lab grown meat? ...... 50 RQ 4: What influence does the message theme of a blog post have on message evaluation and behavioral intention toward lab grown meat? ...... 50 RQ 5: What influence does the message theme of a blog post have on risk and benefit perceptions of lab grown meat? ...... 52 RQ 6: What influence does the message theme of a blog post have on perceptions of uncertainty regarding lab grown meat? ...... 54 RQ 7: How is neophobia to new food technology related to intention to consume? ...... 54 RQ 8: What influence does the message theme of a blog post have on intention to consume? ...... 54

CHAPTER V CONCLUSION, DISCUSSION, AND RECOMMENDATIONS ...... 56

Overview ...... 56

Research Purpose and Questions ...... 56

Part I Conclusions and Discussions ...... 57 RO 1: Determine the potential social reach and volume of lab grown meat on Twitter ...... 57 RO 2: Establish the percent of positive, negative, and neutral messages about lab grown meat on Twitter ...... 58 RO 3: Identify the trending themes about lab grown meat on Twitter ...... 59 RO 4: Determine the top posters on Twitter providing content about lab grown meat ...... 60

Part II Conclusions and Discussions ...... 60

RQ 1: How does the frequency of social media use affect intention to share content and message evaluation? ...... 60 RQ 2: How is pre-existing attitude toward lab grown meat related to risk and benefit perceptions? ...... 61

Texas Tech University, Kellie Anne Boykin, December 2019

RQ 3: How is neophobia to new food technology related to risk and benefit perceptions of lab grown meat? ...... 62 RQ 4: What influence does the message theme of a blog post have on message evaluation and behavioral intention toward lab grown meat? ...... 62 RQ 5: What influence does the message theme of a blog post have on risk and benefit perceptions of lab grown meat? ...... 63 RQ 6: What influence does the message theme of a blog post have on perceptions of uncertainty regarding lab grown meat? ...... 64 RQ 7: How is neophobia to new food technology related to intention to consume? ...... 64 RQ 8: What influence does the message theme of a blog post have on intention to consume? ...... 65

Recommendations ...... 65

Practice ...... 65 Research ...... 67

Summary ...... 69

REFERENCES ...... 70

APPENDIX A IRB APPROVAL ...... 78

APPENDIX B MESSAGE TESTING SURVEY ...... 80

APPENDIX C PARTICIPANT QUESTIONNAIRE ...... 89

vii

Texas Tech University, Kellie Anne Boykin, December 2019

ABSTRACT

Lab grown meat is a new technology being developed as a potential alternative protein source. Although some research has been done about public perception of lab grown meat, no studies to date have analyzed social media content regarding this topic.

Still yet, no studies have observed the effects of message themes on public perception of lab grown meat. This two-part study first sought to analyze the Twitter messages discussing lab grown meat using Meltwater, a social media monitoring software.

Secondly, the study sought to better understand measures of uncertainty and risk and benefit perceptions after viewing a themed blog post about lab grown meat.

In part one, a relevant keyword search from August 28, 2018 to February 28,

2019 collected over 11,000 Twitter messages. Sentiment of messages was analyzed with

47% of messages being neutral. Meltwater identified trending themes that were all closely tied to lab grown meat, and top content posters with the most amount of potential reach were identified. All top posters were found to be news entities or organizations instead of personal Twitter accounts.

In part two, participants were randomly assigned one of three themed blog posts against lab grown meat, neutral, or support lab grown meat. Perception questions were asked after viewing the blog post, and a total of 238 responses were collected. Results indicated message theme had a statistically significant effect on risk perception, benefit perception, and intention to share, but not on message evaluation or measures of uncertainty. Further discussion as well as suggestions for future research are included.

viii

Texas Tech University, Kellie Anne Boykin, December 2019

LIST OF TABLES

3.1 Social Media Use of Survey Respondents ...... 32

3.2 Descriptive Statistics for Food Technology Neophobia Scale (N = 238) ...... 33

3.3 Descriptive Statistics for Attitude Toward Lab Grown Meat (N = 238) ...... 34

3.4 Descriptive Statistics for Message Evaluation (N = 238) ...... 35

3.5 Descriptive Statistics for Risk Perceptions (N = 238) ...... 36

3.6 Original Items Used to Measure Uncertainty ...... 36

3.7 Descriptive Statistics for Measures of Uncertainty (N = 238) ...... 37

3.8 Descriptive Statistics for Intention to Share Content (N = 238) ...... 37

3.9 Descriptive Statistics for Behavioral Intention (N = 238) ...... 38

3.10 Descriptive Statistics for Benefit Perception (N = 238) ...... 39

3.11 Demographic Characteristics of Survey Respondents ...... 42

4.1 Top Seven Trending Themes and Number of Mentions on Twitter ...... 48

4.2 Top Twitter Accounts and Sum of Potential Reach ...... 48

4.3 Frequency of Assigned Stimuli Among Participants (N = 238) ...... 49

4.4 Means, Standard Deviations, and Intercorrelations for the Relationship Between Frequency of Social Media Use and Intention to Share Content and Message Evaluation ...... 49

4.5 Means, Standard Deviations, and Intercorrelations for the Relationship Between Attitude Toward Lab Grown Meat and Risk and Benefit Perceptions ...... 50

4.6 ANOVA of the Effects of Message Themes on Message Evaluation ...... 51

4.7 Means of Message Evaluation (N = 238) ...... 51

4.8 ANOVA of the Effects of Message Themes on Behavioral Intention ...... 51

4.9 Means of Behavioral Intention (N = 238) ...... 51

4.10 ANOVA of the Effects of Message Themes on Risk Perception ...... 52 ix

Texas Tech University, Kellie Anne Boykin, December 2019

4.11 Means of Risk Perception (N = 238) ...... 52

4.12 ANOVA of the Effects of Message Themes on Benefit Perception ...... 53

4.13 Means of Benefit Perception (N = 238) ...... 53

4.14 ANOVA of the Effects of Message Themes on Measures of Uncertainty ...... 54

4.15 Means of Measures of Uncertainty (N = 238) ...... 54

4.16 ANOVA of Effects of Message Theme on Intention to Consume ...... 55

4.17 Means of Intention to Consume (N = 238) ...... 55

Texas Tech University, Kellie Anne Boykin, December 2019

LIST OF FIGURES

3.1 Operational framework for survey experiment ...... 30

4.1 Social media volume and potential reach of lab grown meat on Twitter ...... 46

4.2 Positive, negative, and neutral sentiment of Twitter posts about lab grown meat ...... 47

Texas Tech University, Kellie Anne Boykin, December 2019

CHAPTER I

INTRODUCTION

Background & Setting

In the upcoming years, the world population is expected to grow and as it does, the demand for meat as a protein source is expected to grow with it (Lee, 2018). In the

U.S. alone, meat consumption rose 5% in 2015 (Wilks & Phillips, 2017). Scientists are looking outside of the realm of traditional agriculture to solve the higher demand for animal protein sources while simultaneously lowering the environmental impacts present in agricultural production (Shapiro, 2018).

Lab grown meat, an innovation in cellular agriculture and food biotechnology, has been proposed as an alternative protein source. No single name has been settled upon for this new technology with a variety of descriptors used in the media and literature: cultured meat, in vitro meat, lab grown meat, synthetic, artificial, and factory grown meat

(Verbeke, Sans & Van Loo, 2015). For consistency and lack of confusion, “lab grown meat” will be the term used throughout this study.

Lab grown meat is meat grown in a laboratory from a needle biopsy sample of skeletal muscle STEM cells from a live animal (Bhat, Kumar & Fayaz, 2015; Post, 2014).

The STEM cells are placed in a culture medium and proliferated in a bioreactor (Lee,

2018). In the culture medium, the number of cells increase, and the cells differentiate into muscle fibers (Post, 2014). The muscle fibers can then be assembled into a meat product just like any other traditional burger patty or meatball (Post, 2014).

Although this seems like a novel technology or something out of a sci-fi book, the idea of lab grown meat has been conceptualized for many years. In 1932, in an article

Texas Tech University, Kellie Anne Boykin, December 2019 published in the Strand Magazine entitled “Fifty Years Hence,” Winston Churchill made the statement, “With a greater knowledge of what are called hormones, i.e. the chemical messengers in our blood, it will be possible to control growth. We shall escape the absurdity of growing a whole chicken in order to eat the breast or wing, by growing these parts separately under a suitable medium" (Eschner, 2017, para. 2).

While it took more than 50 years to achieve, the first lab grown beef burger was cooked in 2013 by researcher Mark Post and his team from the University of Maastricht in the Netherlands (Bhat, Kumar, & Fayaz, 2015; Bohm, Ferrari, & Woll, 2018). Since

2013, there have been a multitude of start-up companies all over the world with the goal of producing lab grown meat that is marketable to the consumer (Shapiro, 2018).

Although this is a new technology, there is growing interest both in the public and in research. An online survey of U.S. citizens found two-thirds of the sample would try in vitro meat at least once (Wilks & Phillips, 2017).

This new technology has sparked interest in the eyes of many technology innovators who see the potential it could have on society. Lab grown meat start-ups are funded by people, such as Bill Gates, Jeff Bezos, Richard Branson, and former General

Electric CEO Jack Welch all looking to be a part of this revolution in how we think of food, specifically meat (Shapiro, 2018). Not to be left out, Tyson, the world’s largest producer of meat, has invested in two lab grown meat startups, Memphis Meats and

Future Meat Technologies (Min, 2019). Cargill has also been reported to have investments in lab grown meat start-ups (Campbell, Felix, Hines, & Chiles, 2019).

It has been touted that lab grown meat is a “win-win” because of the benefits of meat without negative drawbacks (Lee, 2018). Producing meat in a laboratory would be

Texas Tech University, Kellie Anne Boykin, December 2019 more energy efficient, more environmentally friendly, and more humane than conventional animal husbandry practices (Bhat, Kumar & Fayaz, 2015; Penn, 2018).

Animals in traditional production systems have historically low feed conversion ratios with cattle being around 15%, pigs at 30%, and chickens around 60% (Post, 2014).

Proponents state lab grown meat could be a more “resource efficient” way to produce meat as protein (Post, 2014). Fewer animals in production systems would decrease greenhouse gas emissions as well as decrease incidence of zoonotic diseases, such as bird flu and Spanish flu, as fewer people are in contact with fewer animals (Shapiro, 2018;

Post, 2014).

Lab grown meat requires no feed inputs, 43.6 gallons of water, and less than a square foot of land to produce one pound of meat (Penn, 2018). New Harvest, a research institute dedicated to funding research of lab grown meat stated, “Cellular agriculture could be how we safely and sustainably feed our growing global population” (New

Harvest, n.d., para. 18). This, plus the fact that animals would not have to be harvested, are clear benefits of this new technology (Bhat, Kumar, & Fayaz, 2015).

Proponents of lab grown meat claim the technology would significantly decrease the environmental footprint our current meat production system has on the environment, while others are not so sure. Some studies show lab grown beef would decrease energy consumption by 45% with 99% less land used (Shapiro, 2018). However, a BBC article published in February 2019, summarized a study showing how methane (greenhouse gas emitted from cows) is less harmful to the environment than carbon dioxide (greenhouse gas emitted from energy production used to grow lab grown meat) (McGrath, 2019).

Texas Tech University, Kellie Anne Boykin, December 2019

Another barrier to this new technology is cost. The one-pound burger produced in

2013 cost approximately $300,000 to create (Bhat, Kumar, & Fayaz, 2015). However, if lab grown meat could be produced cost effectively, scientists hope the issue of feeding a growing population and the increase in the demand for meat could be eliminated (Wilks

& Phillips, 2017).

With this new technology, new concerns and problems have arisen. There is much deliberation in previous literature about how lab grown meat would be regulated if it were to be sold to the public. However, this concern was addressed on March 7, 2019, when the U.S. Department of Agriculture announced the formal agreement for the

USDA, Food Safety Inspection Service, and the U.S. Department of Health and Human

Services’ Food and Drug Administration to jointly regulate and supervise all products derived from the cells of livestock and poultry (USDA, 2019). If lab grown meat becomes an item on the market for production and consumer consumption, the FDA will monitor the collection, banking, and growing of the cells needed for lab grown meat

(Campbell et al., 2019). With this agreement, the USDA will be in charge of production and labeling of subsequent products created via lab grown meat (Campbell et al., 2019).

Controversy still exists over how lab grown meat will be labeled. The Missouri

Department of Agriculture passed legislation on August 30, 2018, that stated “products must include a prominent statement on the front of the package, immediately before or immediately after the product name” if the product is plant-based, veggie, lab-grown, lab- created or comparable (Public Statement, 2018). Not long after Missouri passed its law on labeling, South Carolina followed suit. The House and Senate unanimously passed a law for South Carolina that stated lab grown meat could not be labeled as “meat”

Texas Tech University, Kellie Anne Boykin, December 2019

(Associated Press [AP], 2019). The lawmaker and member of South Carolina’s

Cattleman’s Association, Republican Rep. Randy Ligon stated he “doesn’t want to stop research into the alternative food, but he does want to make sure consumers understand what they are getting,” (AP, 2019, para. 3).

Social Media & Science Communication

Siegrist (2008) found perceived benefits, perceived risks, and perceived naturalness of food technologies all have an effect on whether a technology is accepted by the public. The influence of social media on consumer purchasing behavior has become a recent research focus, and research to date has shown social media content does affect public opinion and consumer behaviors by increasing trust and intention to buy

(Hajli, 2014). Determining how social media may affect consumer purchasing behavior of lab grown meat is one goal of this study.

For many people, media content is the only exposure to science they have once their formal education has been completed (Triese & Weigold, 2002). Media coverage can affect both consumer perception and consumer demand (McCluskey & Swinnen,

2011). Media is often someone’s first exposure to a new technology and its implications

(Malyska, Bolla, and Twardowski, 2016). As the public continues to gradually grow away from reading the newspaper every morning to learn about what is going on in the world, and more and more information is being disseminated to the public via social media, it is important to understand how science is being discussed on social media.

The rise and prevalence of social media have allowed scientists to educate people about their research rapidly without geographic boundaries (Eperen & Marincola, 2011).

It is now common to see scientists and notable scientific organizations such as Richard

Texas Tech University, Kellie Anne Boykin, December 2019

Dawkins, NASA, National Geographic, Stephen Hawking before his passing, and many others all posting on various social media platforms building their reputation and posting information about current scientific research (Eperen & Marincola, 2011).

Social media allow for increased visibility of ideas as well as provides an issue- based network for the exchange of these ideas (Xu et al., 2018). Darling, Shiffman, Cote, and Drew (2013) found social media, more specifically Twitter, has allowed scientists to actively disseminate research to non-governmental agencies, private industry, government agencies, and non-scientists with minimal effort. The researchers also found a little less than half of scientists’ followers on Twitter were non-scientists and members of the general public (Darling et al., 2013). “In a world that gets smaller with every new hashtag, social media has become a key communication channel between scientists, their collaborators and the public” (Winkless, 2013, p. 3).

Limited research has been done about lab grown meat and how this issue is presented on social media. Online comments on news articles about the topic have provided some understanding of public perception (Laestadius & Caldwell, 2015).

Laestadius and Caldwell (2015) conducted a content analysis of online comments about lab grown meat where only 18% of comments were found to be positive toward lab grown meat.

In a summary article about the launch of the first lab grown burger, O’Riordan,

Fotolopoulou, and Stephens (2016) observed through the Culturedbeef.net website that

Twitter and Reddit conversations increased significantly on the day of the launch. In a discourse analysis study conducted around the event of the highly publicized tasting of the first lab grown burger, Twitter and Reddit conversations following the launch showed

Texas Tech University, Kellie Anne Boykin, December 2019 attention from big names and large organizations including New Harvest, PETA, Peter

Singer, and Richard Dawkins (O’Riordan et al., 2016). Reddit conversation showed themes such as what the burger is and how it could be beneficial along with personal opinions of the public participating in the conversation and an identification of who the public investing in the conversation was (Riordan et al., 2016).

Need for the Study

Several factors - price, environmental affect, and animal welfare - play into the public’s perception of lab grown meat. It still remains unclear how consumers will accept this new technology (Verbeke, Sans, & Van Loo, 2017). In a survey conducted by Wilks and Phillips (2017), potential barriers of engagement with lab grown meat were taste/appeal, ethical concerns, religious reasons, environmental concerns, and environmental impact. In his report on the lab grown burger produced in 2013, Mark Post

(2014) stated “for cultured beef to become a viable alternative to livestock beef, its production needs to be resource efficient, sustainable, scalable, and lead to a product that is indistinguishable from current beef” (p. 29). Post (2014) also addressed the importance of consumer acceptance and preference to lab grown meat over traditional meat.

Concerns over “naturalness,” texture, and impact on traditional farming, have also risen surrounding the conversation of lab grown meat (Wilks & Phillips, 2017).

While new food technologies enable innovation, they are not always readily accepted by the public (Siegrist, 2008). Consumers often view new food technology with scrutiny due to the significance and essential nature of food in daily life (Lucht, 2015). It is therefore no surprise the success of new food technologies is heavily reliant on consumer acceptance (Dahabieh, Broring, & Maine, 2018).

Texas Tech University, Kellie Anne Boykin, December 2019

As this issue gains more public attention, there is a need to better understand what is being said and who is saying it. Social media monitoring is the “continuous systematic observation and analysis of social media networks and social communities” (Fensel,

Leiter, & Stavrakantonakis, 2012, p. 3). It is a tool used to “listen” to what people are saying about a topic on the web and across multiple social media platforms (Fensel et al.,

2012). Social media monitoring was chosen for the current study because, unlike traditional monitoring, social media monitoring allows for real time information with the most relevant issues to date (Bekkers, Edwards, & de Kool, 2013). Analysis of social media users’ preferences and priorities allows researchers to understand social networks and predict trends in offline behavior (Munro, Hartt, & Pohlkamp, 2015).

Purpose and Research Questions

This study was completed in two phases. The purpose for the initial quantitative analysis aspect of this study was to describe Twitter conversations about lab grown meat.

The specific research objectives were as follows:

RO 1: Determine the potential social reach and volume of lab grown meat

content on Twitter.

RO 2: Determine the percent of positive, negative, and neutral messages about

lab grown meat on Twitter.

RO 3: Determine the trending themes of lab grown meat on Twitter.

RO 4: Determine the top posters on Twitter providing content about lab grown

meat.

Equally as important as what is being said on Twitter about lab grown meat is how people perceive risks and benefits of lab grown meat when presented with a themed

Texas Tech University, Kellie Anne Boykin, December 2019 message. The purpose of this study was to examine the influence of pre-existing attitudes and themed messages on public perceptions of lab grown meat.

Therefore, the questions guiding the second phase of the study were as follows:

RQ 1: How does the frequency of social media use affect intention to share

content and message evaluation?

RQ 2: How is pre-existing attitude toward lab grown meat related to risk and

benefit perceptions?

RQ 3: How is neophobia to new food technology related to risk and benefit

perceptions of lab grown meat?

RQ 4: What influence does the message theme of a blog post have on message

evaluation and behavioral intention toward lab grown meat?

RQ 5: What influence does the message theme of a blog post have on risk and

benefit perceptions of lab grown meat?

RQ 6: What influence does the message theme of a blog post have on perceptions

of uncertainty regarding lab grown meat?

RQ 7: How is neophobia to new food technology related to intention to consume?

RQ 8: What influence does the message theme of a blog post have on intention to

consume?

Conceptual Framework

Due to the two-part nature of this study, a conceptual and framework guided the research questions and methodology of this study. The conceptual framework guiding the content analysis portion of the study included the concepts of opinion leadership and sentiment analysis. Sentiment refers to the judgement or emotional connection a person

Texas Tech University, Kellie Anne Boykin, December 2019 determines about a topic – often recorded as positive, negative, or neutral (Stieglitz &

Dang-Xuan, 2013). Social media is a platform that people often convey sentiment making it a communication outlet to be studied to observe sentiment towards lab grown meat

(Stieglitz & Dan-Xuan, 2013). The conceptual framework for the second part of the study embodied the concepts of risk and benefit perceptions and measures of uncertainty. Risk is a measure of hazard times exposure, and is inversely related to benefit perceptions

(Juanillo, 2001; Ueland et al., 2011). Uncertainty perceptions measures the amount of conflicting or confusing information that exists in a scientific issue (Zehr, 2000).

For this study, Twitter was the social media platform used to monitor social media. The openness of Twitter creates an ideal environment for social discourse (Park,

2013). Scientific communications are happening ever more frequently on Twitter allowing for a broad audience to access communication about science (Bombaci et al.,

2016). Because Twitter allows for such diverse interactions with endless amounts of others, opinion leaders on Twitter are often able to gather support and gain momentum rather quickly (Park, 2013).

Unlike traditional opinion leadership, being an opinion leader on Twitter is not dependent upon socio-economic status (Park, 2013). Proficiency in a topic is much more important, which allows nearly anyone to become an opinion leader on Twitter (Park,

2013). As long as a person can create content people find worthy of engagement, opinion leaders on Twitter could come from anywhere (Park, 2013). This prevents accurate prediction of potential opinion leaders on Twitter for this new technology. Identification of opinion leaders and their opinions through social media monitoring can allow for prediction and further analysis of public opinion. While several studies have focused on

Texas Tech University, Kellie Anne Boykin, December 2019 opinion leaders for plant biotechnologies (Xu et al., 2018), none have addressed Twitter influencers’ discussion of lab grown meat.

Assumptions

Several assumptions for this study have been identified. In the first phase of this research, the researcher assumed the Meltwater algorithm to code sentiment is the same as human coders and able to determine the human aspect of posts regarding the positive or negative perception of lab grown meat. Kim, Jang, Kim, and Wan (2018) identified an algorithm’s lack of sensitivity to conceptual boundaries of topics when using the big-data approach as a potential source of error. It is also assumed each of the posts Meltwater returned was relevant based on the customized keyword search.

In the second phase of the study, it was assumed Marketing Systems Group had access to a sample that was representative of the U.S. population. It was also assumed participants responding to the survey instrument were honest in their answers.

Limitations

A limiting factor in part one of the study was comments online may not reflect the perceptions of the general public because people who make comments online tend to hold stronger opinions than most (Wilks & Phillips, 2017). Therefore, people making comments about lab grown meat through Twitter may be the segment of the population who hold strong, polarizing opinions and may not reflect the true sentiment of the general public. The social media content was also limited to what Meltwater was able to identify using its search functionality. In the second part of the study, a budget for data collection limited the number of responses the researchers were able to request from MSG to 300.

Texas Tech University, Kellie Anne Boykin, December 2019

CHAPTER II

LITERATURE REVIEW

Overview

Chapter I provided background regarding lab grown meat and established the need for the investigation of Twitter conversations and risk and benefit perception of consumers. This chapter provides an overview of previous research on public perception of novel or risky food, communication involving new food technologies, previous studies done on public perception of lab grown meat, and the constructs used to formulate the conceptual framework used to guide the research objectives and design of the study. The constructs of risk and benefit perceptions combined with measures of uncertainty will be illuminated as the guide for the construction of the survey instrument.

Consumer Acceptance of Food Technology

Applying scientific knowledge to address societal problems with new technology is generally widely accepted when it is clear the new technology will benefit a consumer’s quality of life; however, new technology introduced into food is often received with much more skepticism from consumers (de Barcellos et al., 2010). One example is the application of biotechnology, which is often met with opposition from consumers despite scientific data proving its safety (Lin, Ortega, Caputo & Lusk, 2019).

“Successful adoption of biotechnology for animal agriculture will ultimately depend on having a thorough understanding of consumer preferences” (Lin et al., 2019, p. 10).

The term biotechnology originated in 1917 with the Organization for Economic

Co-operation and Development defining the term as the “application of scientific and engineering principles to the processing of materials by biological agents to provide

Texas Tech University, Kellie Anne Boykin, December 2019 goods and services” (Bud, 1993, p. 1). Its application dates back to alcohol production in the middle ages and has progressed to address agricultural challenges and even the development of drugs (Bud, 1993).

Prior to the evolution of lab grown meat, genetically modified organisms (GMOs) were the most significant advancement in food biotechnology and agricultural development (Chen & Chern, 2002). The results of the survey of public perception of

GMOs showed risk perception, environmental concern, religious and ethical concern, opinion of labeling, perceived difference between genetically modified and non- genetically modified foods, number of children within a household, and price were all the most significant factors influencing consumer willingness to purchase food containing

GMOs (Chen & Chern, 2002). Due to the large number of factors identified, the great complexity going into a consumer’s perception of and acceptance of agricultural biotechnology becomes evident.

Cardello, Shutz, and Lesher (2007) conducted a conjoint analysis surveying potential consumers of new food technologies. Their sample included a consumer panel of civilian lab employees, shoppers in a U.S. mall, and U.S. military troops on training exercises. Similar to other studies, Cardello et al. (2007) found risk was the most determining factor in consumer acceptance of a new food technology. Of the technologies used in the conjoint design (irradiation, pulsed electric fields, high pressure, heat pasteurization, ionizing energy, genetic modification, and cold preservation), genetic modification and irradiation showed the highest level of concern while other technologies evoked a much lower level of concern. The researchers also noted in their findings how

Texas Tech University, Kellie Anne Boykin, December 2019 critical taste and nutrition were to consumer perception and acceptance (Cardello et al.,

2007).

Different food technologies have been found to have different levels of consumer acceptance. Following a content analysis of the current literature on the consumer acceptance of seven food-related technologies – genetic modification, nutrigenomics, animal cloning, food irradiation, nanotechnology, high-pressure processing (HPP), and pulsed electric field processing (PEF) – Frewer et al., (2011) found GM foods and food irradiation to have a low acceptance level. Conversely, HPP, PEF, and nutrigenomics seemed to have no large level of consumer negativity (Frewer et al., 2011). While

“naturalness” is often a concern when addressing new food technologies, Frewer et al.

(2011) suggested it is not the unnaturalness of the technology itself which consumers dislike, rather it is the idea of “uncontained bioactivity” that raises concerns in consumers about a new technology. Consumers are more likely to accept a new technology if they have control over their consumption of the product as opposed to the product being untraceable (Frewer et al., 2011).

Consumer Acceptance of Lab Grown Meat

Due to the importance of consumer acceptance on the success of new food products, consumer attitudes, including naturalness and risk and benefit perceptions, should be evaluated at an early stage in the process of developing a new food technology

(Siegrist, 2008).

Wilks, Phillips, Fielding and Hornsey (2019) studied the psychological effects behind people’s willingness to eat lab grown meat, and found food neophobia, tested using the Food Technology Neophobia Scale (FTNS), was the strongest predictor of a

Texas Tech University, Kellie Anne Boykin, December 2019 person’s willingness to try lab grown meat. People who are more willing to try new food were more likely to eat lab grown meat as well as see the benefits of the technology

(Wilks et al., 2019).

In order to understand perceived naturalness and disgust, Siegrist, Sutterlin, and

Hartmann (2018) conducted an Internet study about consumer acceptance of lab grown meat. Each participant was given a description of either organic meat or lab grown meat.

Following the assigned description, each participant was asked how they perceived naturalness and how willing they were to consume either organic or lab grown meat.

Both groups were asked the same questions about traditional meat. The study showed that when comparing organic to lab grown meat, perceived naturalness affects a consumer’s willingness to consume the meat (Siegrist et al., 2018). The study also found participants given information about lab grown meat viewed traditional meat as more natural with a greater willingness to buy traditional meat than the organic meat participant group. They were able to conclude that perception and acceptance are influenced by how lab grown meat is described (Siegrist et al., 2018).

In their review of the current literature surrounding public perception of lab grown meat, Verbeke, Sans, and Van Loo (2015) identified three major issues to acceptance: unnaturalness, repulsion or “yuck factor,” and healthiness. The result of their survey showed that providing additional information about the benefits of lab grown meat resulted in a greater willingness to try lab grown meat as well as willingness to purchase lab grown meat. Although the study indicated only a small number of consumers completely rejecting the idea of trying lab grown meat, there is no proof of how likely

Texas Tech University, Kellie Anne Boykin, December 2019 consumers will be to repeatedly buy or replace traditional meat with lab grown meat

(Verbeke et al., 2015).

In a cross-country survey of consumer’s perception of plant and lab grown meat, consumers in India and China were more willing to consume lab grown meat than consumer counterparts in the U.S. (Bryant et al., 2019). Of the consumers in all three surveyed countries, meat-eaters and omnivores were more likely to purchase lab grown meat than pescatarians, vegetarians, or vegans. In the U.S., political leaning to the left and those more familiar with lab grown meat showed a higher intent to purchase lab grown meat given the scenario that it was on grocery store shelves. In each country, lower levels of food neophobia were more likely to consume and purchase lab grown meat. Overall, urban, well-educated and high-income consumers in India and China showed a much greater likelihood to purchase lab grown or plant-based meat than consumers in the U.S. (Bryant et al., 2019).

Throughout literature and news media, there is much inconsistency in the name used to describe lab grown meat. Everything from in vitro meat, clean meat, cultured meat, artificial meat, synthetic meat, and lab grown meat have all been used to refer to one product (Bryant & Barnett, 2019). In order to understand consumer perception of each of these names, Bryant and Barnett (2019) conducted an experiment with the four terms clean meat, cultured meat, animal free meat, and lab grown meat as the manipulated condition groups. The researchers found attitudes were much more positive when lab grown meat was referred to as “clean meat” or “animal free meat” (Bryant &

Barnett, 2019). The conditions where lab grown meat was referred to as “lab grown meat” were perceived much more negatively, but the term “clean meat” showed the

Texas Tech University, Kellie Anne Boykin, December 2019 greatest behavioral intention (Bryant & Barnett, 2019). This study found that attitudes and behavior toward lab grown meat can vary greatly according to the term used to describe the innovation.

Science Communication

Burns, O’Connor, and Stocklmayer (2003) reviewed the literature to date to find a cohesive definition of science communication. The definition agreed upon and which will be used for this study was “the media, activities, and dialogue to produce one or more of the following to science: awareness, enjoyment, interest, opinions, and/or understanding”

(Burns et al., 2003, p. 191). Public scientific awareness, understanding, and culture should the goal of science communication. In addition, science communication keeps people informed on new developments that may affect them as well as bringing a broader context of science to the public (Treise & Weigold, 2002).

Social media is a way for scientists to discuss their findings, share research, and act as a voice for science directly to the public without going through the news media

(Bik & Goldstein, 2013). Because of the user-controlled nature of social media, people are often able to take a “bottom-up” approach to scientific information seen on social media (Nisbet & Scheufele, 2009). This “bottom-up” logic is due to a person’s personal experience and lay knowledge used to understand science (Nisbet & Scheufele, 2009).

Social media is a way to gain “greater influence in the discursive contest that surround issues such as. . . biotechnology” (Nisbet & Scheufele, 2009, p. 1771).

Sharing functions on social media create an atmosphere of quick and extensive diffusion of information (Majmundar, Allem, Cruz, & Unger, 2018). In an effort to understand why people retweet, Majmundar et al. (2018) tested the Why We Retweet

Texas Tech University, Kellie Anne Boykin, December 2019

Scale. “Retweeters” were identified as being information creators, promoters, supporters, and consumers. The top four reasons people retweet were to show approval, argue, gain attention, and entertain. Because retweeters are often consumers and show approval of an issue via social media sharing practices, there arises a need to better understand intention to share content regarding lab grown meat.

The most effective science communication connects to the public through personal relevance – a value or priority (Nisbet & Scheufele, 2009). Food is one of few things people all over the world not only need, but value in special ways as part of daily life, celebrations, and culture. Because of the important nature food plays in daily life, consumers are increasingly seeking food information (Verbeke, 2008). Enjoying food more, having a healthier diet, avoiding allergens and knowing how food was grown/raised are all of upmost importance to today’s consumers. Communication and information affect both consumer food choices and dietary choices (Verbeke, 2008).

Conceptual Framework

Sentiment Analysis

Due to the two-part nature of this study, a conceptual framework guided the research questions and methodology of this study. The conceptual frameworks guiding the content analysis portion of the study included the concepts of sentiment analysis and opinion leadership. Sentiment is defined as suggesting “a settled opinion reflective of ones feelings” (Mejova, 2009, p. 4). Sentiment is most commonly characterized by its polarity – being either positive or negative toward an issue or product (Pang, Lee, &

Vaithyanathan, 2012; Pang & Lee, 2008). Sentiment is important because as humans,

Texas Tech University, Kellie Anne Boykin, December 2019 when making a decision, we care what other people think about the issue (Pang & Lee,

2008).

At one point in their lives, 81% of Internet users have researched a product online to determine their own opinion on the product (Pang & Lee, 2008). Humans want to know if their peers feel positively or negatively about topics they themselves are unsure about (Pang & Lee, 2008). Sentiment analysis allows us to answer questions and better understand how opinions are spread in discussions and how opinions are adopted by others (Mejova, 2009). Decisions and future debate topics surrounding new food technology can be better understood using sentiment analysis (Munro et al., 2015).

Negative stimuli tend to evoke stronger emotional and cognitive responses from people than neutral or positive stimuli (Stieglitz & Dang-Xuan, 2013). Likewise, negatively coded sentiment posts on social media garner more attention and feedback than positive sentiment posts. Stieglitz and Dang-Xuan (2013) found sentiment was associated with intention to retweet and the speed a message is retweeted ultimately affecting the diffusion of information through the Twitter platform.

Opinion Leadership

Previous research has shown key influencers (i.e. opinion leaders) play a significant role in conversations around controversial scientific issues because of their influence on social media (Xu, Yu, & Song, 2018). Opinion leaders may influence opinions of others through the two-step flow theory of information flowing from mass media to opinion leaders to the rest of the population (Xu et al., 2018). Opinion leaders are characterized by higher understanding and higher levels of knowledge about a topic than a non-opinion leader (Parks, 2013).

Texas Tech University, Kellie Anne Boykin, December 2019

Opinion leaders are referred to in Meltwater as “key influencers” which are defined as trusted experts within a specific field (Dods, 2018). These key influencers have a large following and have a high user engagement rate. Opinion leaders on social media are identified by the volume of their communication. Through the use of social listening, Meltwater identifies key influencers based on their rank as described below:

Rank analyzes the author of the post based on all discoverable social information.

Rank takes into account the size of the author’s community, the frequency at

which they participate in social environments, the channels on which they have a

presence and their level of engagement with you. Rank is on a scale of 0 to 10.

Views represents the total possible unique views for a specific post if everyone to

whom the post was directed was able to observe the post in their stream. This is

the number of impressions the post made in the world (Cruz, 2013, para. 8).

The “top posters” feature of Meltwater, specifically, measures who has created the most content about a topic within a specific time frame (Cruz, 2013). People are relying more on their social circle on social networks for news and opinions (Weeks,

Ardevol-Abreu & Zuniga, 2017). The more highly active and the more content an individual creates, the more influence an individual has on their followers on social media.

Opinion leaders on social media employ framing as they carefully choose what information to include or omit in order to emphasize a specific point of view (Xu et al.,

2018). The frames opinion leaders choose to use on social media indicate the information they have paid attention to and want their followers to know (Xu et al., 2018). Xu et al

Texas Tech University, Kellie Anne Boykin, December 2019

(2018) found that the message themes employed by opinion leaders causes variation in user engagement when discussing GMOs specifically.

Risk and Benefit Perceptions

Risk is assessed by the formula that states risk = hazard x exposure (Juanillo,

2001). Assessment of risk is essentially the process of answering the question of what is safe (Juanillo, 2001). In the context of risk communication, exposure, the amount of time and frequency exposed to a message, and the actual content of messages may impact risk perception (Binder, Scheufele, Brossard, & Gunther, 2011).

Binder et al. (2011) found increased exposure to a message amplified both benefit and risk perceptions. Discussion of the issue was shown to amplify risk, benefit, or neutrality positions rather than sway people closer to one side of the issue (Binder et al.,

2011). The results of this study are significant as we know that preexisting attitudes, whether a person assesses something as positive or negative, have been shown to affect behavior toward a new issue (Kim et al., 2014). Initial evaluation may determine the public’s perception of a new technology (Kim et al., 2014).

A clear example of this phenomenon is how negative risk perceptions of GMOs lead to a decrease in public support for the technology (Kim et al., 2014). Because certain applications of food technologies are seen differently than others, it is important to study each application independently (Magnusson & Hursti, 2002). Perceptions of GMOs may not translate to perceptions of lab grown meat. Since social media and blogs are a way for people to share preexisting opinions, be exposed to content, and discuss opinions, the researchers chose to use these concepts to design a survey using a blog post as a stimuli to test risk and benefit perception.

Texas Tech University, Kellie Anne Boykin, December 2019

When faced with new information, people typically use heuristic shortcuts to make judgements (Kim et al., 2014). Bringing all of these concepts together shows

“construction of attitudes and preferences are informed by perceptions of risks and benefits, which rely on cognitive heuristics to process information” (Kim et al., 2014, p.

967).

In regard to food technologies, consumers often see them as risky (Cavaliere &

Ventura, 2017). This is magnified in the eyes of consumers as marketers portray the exact opposite of food technology, “all-natural,” as the healthiest, most beneficial option when it comes to food (Biltekoff, 2010). Despite technologies role in keeping food safe and plentiful, there is an underlying connotation that things of nature or “natural” are inherently pure (Biltekoff, 2010).

In a literature review of risk and benefit perceptions of new food technologies,

Ueland et al. (2011) found risks and benefits were inversely correlated; when benefit perception is high, risk tends to be low. Consumers tend to be more cautious rather than adventurous toward new food (Ueland et al., 2011). Foods that are “traditional” and “well known” tend to align with perceptions of benefits, while new or highly processed food tends to be associated with higher risk perception (Ueland et al., 2011). Ueland et al.

(2011) urged communicating relevant information in order to perform successful benefit- risk analyses.

Measures of Uncertainty

Scientific uncertainty is an element of incompleteness in regard to something in nature or something that results in a dissonance in a scientific claim (Zehr, 2000). This is not always a negative thing, as it is often what pushes scientists to continue researching in

Texas Tech University, Kellie Anne Boykin, December 2019 order to clear up the uncertainty. However, this could pose an issue between the public and scientists as uncertainty can lead to mistrust of scientists among the public (Zehr,

2000).

Communicating scientific uncertainty is essential because all aspects of science contain some uncertainty (Fischhoff & Davis, 2014). If uncertainty is not communicated effectively, someone may put too much or too little faith into a technology and make an inaccurate decision regarding it. Scientific communication should uncover uncertainties and simplify uncertainties to a point where people can identify the best choice about a scientific innovation for themselves.

Uncertainty is increased when common words are used in uncommon ways regarding a technology (Fischhoff & Davis, 2014). Scientific communication should use clear terms that allow for no confusion, and should be “driven by what audiences need to know, not by what scientists want to say” (Fischhoff & Davis, 2014, p. 13668). Effective communication about uncertainties should result in clearer, more effective decisions about technologies.

Uncertainty is present in risk information and may affect the impact risk information has when it reaches the public (Han et al., 2008). Han and colleagues (2008) conducted a qualitative study of responses to uncertainty regarding cancer risk prediction where they found that perceptions of uncertainty had different effects on different people.

Some associated uncertainty with a greater risk perception, but some did not associate uncertainty with any heightened risk. Han et al. (2008) conclucded that uncertainty does matter to people, even if it has different effects on different people.

Texas Tech University, Kellie Anne Boykin, December 2019

Mass media often spotlight novel technologies and by doing so, can influence perceptions of uncertainty (Friedman, Dunwoody, & Rogers, 1986). Because lab grown meat is such a new innovation, not yet available to purchase, the uncertainty created by the media was explored.

Texas Tech University, Kellie Anne Boykin, December 2019

CHAPTER III

METHODOLOGY

Overview

This chapter focuses on the research methodology used to guide this two-part study. Part one sought to better understand the conversation about lab grown meat on

Twitter through social media monitoring. Part two set out to better understand how public opinion and willingness to accept lab grown meat was affected by how information about this topic was themed. This chapter provides the following details for each stage of this study: research questions, research design, instrumentation, sample, data collection, and data analysis.

Part I

Purpose and Research Objectives

The purpose of part one was to better understand the conversation about lab grown meat on Twitter. The specific research objectives for the social media monitoring portion of the study were as follows:

RO 1: Determine the potential social reach and volume of lab grown meat content

on Twitter.

RO 2: Determine the percent of positive, negative, and neutral messages about lab

grown meat on Twitter.

RO 3: Determine the trending themes of lab grown meat on Twitter.

RO 4: Determine the top posters on Twitter providing content about lab grown

meat.

Texas Tech University, Kellie Anne Boykin, December 2019

Research Design

In order to answer these research objectives, a descriptive, quantitative content analysis was conducted using Meltwater, a social media monitoring program. Meltwater has the capability of monitoring social media messages across multiple platforms, but for this study, only Twitter was monitored. Twitter is effective for social media studies due to the large number of brief messages, free-format of messages, and ease of accessibility

(Pak & Paroubek, 2010). The variables of interest for this study are reach (how many individuals potentially viewed the social media content), volume (content specifically about lab grown meat), trending themes (commonly used words or phrases surrounding the conversation of lab grown meat), and identification of key influencers (social media accounts that are most outspoken about the topic and have a large number of followers and shares of content).

Sample

Meltwater pulled all Twitter posts with the pre-determined keywords from August

28, 2018 – February 28, 2019. This time frame was selected after a search of news about lab grown meat was conducted from 2018-2019. Significant, news-worthy events regarding lab grown meat happened in August 2018 and February 2019 providing a six- month time frame to monitor. Upon research into news events in the past six months to a year, it was found that on August 30, 2018, the Missouri Department of Agriculture released a public statement announcing the enactment of the Missouri Meat Advertising

Law (“Public Statement,” 2018). This event was used as the beginning time anchor for the study. On February 19, 2019, the BBC published an article about cultured meat and

Texas Tech University, Kellie Anne Boykin, December 2019 its effect on climate change (McGrath, 2019). Using the BBC article as an ending anchor point, the monitor was run through February 28, 2019 for a six-month time period.

Using the researcher-established keywords, Meltwater searched all publicly- available Twitter content. Abilhoa and De Castro (2014) defined keyword identification as “the task of finding the words that best describe the subject of a text” (p. 309).

Indexing, summarization, topic detection and tracking are ways in which keywords should be identified to be used in keyword searches (Abilhoa & De Castro, 2014). Based on these principles, the keywords used to search for content were: “lab grown meat,”

“fake meat,” “in vitro meat,” “cell-cultured food,” and “cellular agriculture.”

Data Collection

The data for this part of the study were the tweets Meltwater identified using the selected keywords. Meltwater provides these tweets and additional descriptive data in the form of an Excel file and through data visualization tools called widgets. Each widget provided accompanying percentages and numbers to visually represent the data in the form of charts and graphs.

The variables of interest for this study were reach, volume, trending themes, and identification of key influencers. Meltwater widgets provided summaries of each of these items and the data were available for download to Microsoft Excel. Social reach is a measure of how many potential viewers a message received (Robinson, 2018). Social volume measured how many posts there were on a topic (Robinson, 2018).

Meltwater defines sentiment as “a research area examining people’s opinions, attitudes, and emotions in written language” (Robinson, 2017). Meltwater uses Natural

Language Processing (NLP) to analyze sentiment within social media posts (Robinson,

Texas Tech University, Kellie Anne Boykin, December 2019

2017). The Sentiment widget classified the tweets of lab grown meat into positive, negative, and neutral sentiment.

The keywords and phrases most commonly used in regard to lab grown meat were measured using the Trending Themes widget. The Top Posters by Volume widget showed who was posting the most about lab grown meat. This widget provides a summary of the top 10 posters of content regarding lab grown meat were ranked and displayed.

Data Analysis

Data analysis occurred within Meltwater. Each research objective was represented graphically by a Meltwater widget which was able to be exported as a .png or .jpeg file.

Raw data were exported to Microsoft Excel to view individual tweets, which assisted in the development of the message stimuli tested during Part II of the study.

Part II

Purpose and Research Questions

The purpose of this study was to examine the influence of pre-existing attitudes and themed messages on public perceptions of lab grown meat. The specific research questions guiding this portion of the study were as follows:

RQ 1: How does the frequency of social media use affect intention to share

content and message evaluation?

RQ 2: How is pre-existing attitude toward lab grown meat related to risk and

benefit perceptions?

RQ 3: How is neophobia to new food technology related to risk and benefit

perceptions of lab grown meat?

Texas Tech University, Kellie Anne Boykin, December 2019

RQ 4: What influence does the message theme of a blog post have on message

evaluation and behavioral intention toward lab grown meat?

RQ 5: What influence does the message theme of a blog post have on risk and

benefit perceptions of lab grown meat?

RQ 6: What influence does the message theme of a blog post have on perceptions

of uncertainty regarding lab grown meat?

RQ 7: How is neophobia to new food technology related to intention to consume?

RQ 8: What influence does the message theme of a blog post have on intention to

consume?

Research Design

In order to address the research questions of the second part of this study, a between-subjects experimental research design was used. The operational framework for this study is shown in Figure 3.1. A between-subjects experimental survey allows differing groups to participate in each experimental condition, allowing the researcher to view differences between the groups (Field, 2018). The experiment was embedded in a survey instrument. Institutional Review Board approval certification can be found in

Appendix A. The survey instrument used for data collection can be found in Appendix C.

Texas Tech University, Kellie Anne Boykin, December 2019

Assigned Message Stimuli

Dependent Variables Independent Variables • Message • Food Technology Positive Evaluation Neophobia • Risk Perception • General Social • Benefit Perception Media Use • Uncertainty • Attitudes About Negative • Lab Grown Meat Behavioral Intention

Neutral

Figure 3.1. Operational framework for survey experiment

Sample

Participants in the study were recruited using Marketing Systems Group (MSG).

MSG is an information systems company used to distribute survey instruments, compensate participants, and collect survey data. Because the survey was an experimental design with a nationally distributed sample, a minimum of 30 responses per message stimulus was needed according to the rule of thumb established by Roscoe

(1975). The experimental design had three message stimuli for a total of 150 required responses. In order to increase statistical power and account for online survey errors, 300 complete responses were requested. For completing the survey instrument, MSG compensated participants with points, which are redeemable for Amazon gift cards for their responses.

Two quality checks were embedded within the survey. The first question asked,

“Please select disagree.” The second question said, “Please select agree.” If both quality check questions were answered incorrectly, the response was removed from the data set.

Texas Tech University, Kellie Anne Boykin, December 2019

Sixty-one responses were removed leaving a total of 239 viable responses. One additional response was removed due to not meeting the minimum age requirement of 18 years of age. The final total of usable data was 238.

Instrumentation

The survey instrument was constructed in Qualtrics and distributed by MSG.

Participants answered 23 pre-treatment questions before being randomly assigned to one of the three message treatments. Thirty-five post-treatment questions were asked following exposure to the treatment.

Social media usage. Prior to viewing the message stimuli, three questions were asked about the participant’s social media usage. One question was asked regarding if participants had ever created a social media account. If the answer was “no,” skip logic sent participants past the next two questions. Two questions were adopted from a study about teacher use of Twitter (Nochumson, 2018). The first question, “how long have you had a social media account” had the following response options: less than 6 months, 6 months – 1 year, 1-2 years, and longer than 2 years. The second questions stated “how often do you use your social media account” with seven response options: daily, 2-3 times a week, once per week, 2-3 times per month, once per month, and less than once per month.

Social media habits were recorded with questions regarding social media account creation, how long participants have been on social media, and how frequent respondents use social media (Table 3.3). When asked whether they had ever created a social media account, 214 respondents (89.9%) reported yes with only 24 respondents (10.1%) reporting they had never created a social media account.

Texas Tech University, Kellie Anne Boykin, December 2019

Table 3.1 Social Media Use of Survey Respondents n % How Long on Social Media Less Than 6 Months 7 2.9 6 Months – 1 Year 9 3.8 1-2 Years 21 8.8 Longer Than 2 Years 177 74.4 Use of Social Media Daily 161 67.6 2-3 Times a Week 23 9.7 Once per Week 13 5.5 2-3 Times per Month 4 1.7 Once per Month 2 0.8 Less Than Once per Month 11 4.6

The majority of participants reported they had been on social media longer than two years (n = 177, 74.4%). Twenty-one respondents (8.8%) reported having social media account(s) for one to two years, nine (3.8%) respondents had social media account(s) for six months to a year, and only seven (2.9%) respondents had only been on social media for less than six months.

Respondents were asked how often they use social media. One hundred sixty-one of the participants (67.6%) indicated that they use social media every day. A significantly lower number of participants responded they use social media 2-3 times per week (n =

23, 9.7%), once per week (n = 13, 5.5%), once per month (n = 2, 0.8%), or less than once per month (n = 11, 4.6%).

Food technology neophobia. Pliner and Hobden (1992) developed the Food

Neophobia Scale (FNS) which sought to measure how willing a person was to try new foods. Cox and Evans (2008) revised the FNS to create the Food Technology Neophobia

Scale (FNTS), which took consumer perception of new technologies into account. The

FTNS questions used were adapted from Allen’s (2012) master thesis. Participants responded to 13 questions on a 7-point Likert-type scale (1 = strongly disagree, 7 =

Texas Tech University, Kellie Anne Boykin, December 2019 strongly agree). Reliability was calculated post hoc with Cronbach’s a = .85. All responses were summed and averaged to create the food technology neophobia scale

(Table 3.2). The grand mean was 4.02 (SD = 0.89) indicating neither a strong nor weak aversion to food technology overall.

Table 3.2 Descriptive Statistics for Food Technology Neophobia Scale (N = 238) M SD It can be risky to switch to new food technologies too quickly. 4.53 1.41 New food technologies are something I am uncertain about. 4.51 1.53 Society should not depend heavily on technologies to solve its 4.35 1.64 food problems. The media usually provides a balanced and unbiased view of 4.23 1.59 new food technologies. * New food technologies decrease the natural quality of food. 4.22 1.59 New food technologies may have long term negative 4.22 1.42 environmental effects. The benefits of new food technologies are often grossly 4.18 1.44 overstated. New foods are not healthier than traditional foods. 4.05 1.51 New food technologies are unlikely to have long-term negative 4.00 1.38 health effects. * There are plenty of tasty foods around, so we do not need to 3.84 1.60 use new food technologies to produce more. There is no sense trying out products developed through food 3.59 1.63 technology because the ones I eat are already good enough. New food technologies give people more control over their 3.29 1.35 food choices. * New products produced using new food technologies can help 3.24 1.29 people have a balanced diet.* Note: * indicates item was reverse coded. Scores were recorded on a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree).

Attitudes toward lab grown meat. Questions about attitudes toward lab grown meat were adapted from Bryant et al. (2019) where questions were modified from a semantic scale to statements where participants could select their level of agreement. Five questions were asked with responses recorded on a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree). Cronbach’s a = .68 was calculated post hoc. An example

Texas Tech University, Kellie Anne Boykin, December 2019 of the questions asked is “I believe lab grown meat is healthy.” Table 3.3 shows the summed and averaged scores for each item creating the Attitude Toward Lab Grown

Meat Scale with a calculated grand mean of 3.53 (SD = 1.06).

Table 3.3 Descriptive Statistics for Attitude Toward Lab Grown Meat (N = 238) M SD I believe lab grown meat is healthy. 3.22 1.58 I believe lab grown meat is bad for the environment.* 3.92 1.54 I believe lab grown meat is safe. 3.51 1.61 I believe lab grown meat is natural. 2.97 1.68 I believe lab grown meat is unsustainable as a long-term food 4.03 1.56 source.* Note: * indicates item was reverse coded. Scores were recorded on a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree).

Message stimuli. The researcher developed three blog posts as the message treatments (Appendix C). These messages were developed from online content found through Meltwater sentiment analysis and existing blog posts. One blog post was written in support of lab grown meat. One was written being opposed to lab grown meat, and the final blog post was written with a neutral and balanced view of lab grown meat. The positive blog post was adapted from Bloch (2019) and GrantTree (2018) and were edited to ensure a positive outlook toward lab grown meat. The neutral blog post was developed from Rabie’s (2019) blog post and edited to ensure a neutral viewpoint of lab grown meat. The negative blog post was adopted from Van Eenennaam (2019) and Condon’s

(2018) blog post and were adapted accordingly.

Messages were compared to tweets coded for sentiment to ensure the ideas and main points were for, neutral toward, and against lab grown meat. Blog post stimuli all had the same author, and each had approximately the same word count (within 50 words of each other). Word count remained around 400 words for each stimuli. 34

Texas Tech University, Kellie Anne Boykin, December 2019

Message evaluation. After viewing the randomly assigned stimulus, participants provided their evaluation of the message by answering eight questions using a 7-point

Likert-type scale (1 = strongly disagree to 7 = strongly agree). These measures were adopted from Steede’s (2018) study of trust of messages about animal antibiotics. A sample statement from the measure was “This blog post is reliable.” One quality check question was embedded in this question block to weed out any unusable or valueless data.

Steede (2018) reported Cronbach’s a = .839. In the current study, post hoc reliability was calculated with Cronbach’s a = .885. Descriptive statistics for message evaluation can be seen in Table 3.4. The responses were summed and averaged to create the message evaluation scale with a mean score of 4.60 (SD = 1.08) indicating overall agreement with the statements regarding the viewed blog post.

Table 3.4 Descriptive Statistics for Message Evaluation (N = 238) M SD The information in this blog post is easy to understand. 5.32 1.32 The blog post is credible. 4.50 1.26 I trust the information contained in this blog post. 4.45 4.42 The information in this blog post is accurate. 4.39 1.22 The information on this blog post is reliable. 4.34 1.31 Note: Scores were recorded on a 5-point Likert-type scale (1 = very unlikely, 5 = very likely).

Risk perceptions. Risk perception was measured with three questions using a 5- point Likert-type scale (1 = very unlikely, 5 = very likely). An example of the questions in the measure was “How likely is it that lab grown meat presents a serious health hazard?”

These questions were adopted from both Binder, Schuefele, Brossard, and Gunther

(2011) and Kim, Yeo, Scheufele, and Xenos (2014). Binder et al. (2014) reported a

Cronbach’s a = .86 and Kim et al. (2014) reported a Cronbach’s a = .91. Post hoc reliability was calculated on the three measures with a result of Cronbach’s a = .925. One 35

Texas Tech University, Kellie Anne Boykin, December 2019 additional question asked participants how the blog post made them worry about the impacts of lab grown meat with five response options (1 = doesn’t make me worry at all,

5 = makes me extremely worried). Responses were summed and averaged to create the risk perception scale with a mean score of 3.29 (SD = 1.14). This mean score indicates that overall, lab grown meat neither likely nor unlikely is perceived as a risk.

Table 3.5 Descriptive Statistics for Risk Perceptions (N = 238) M SD It is ___ that lab grown meat poses a serious danger to future 3.33 1.24 generations. It is ___ that lab grown meat presents a serious health threat. 3.30 1.22 It is ___ that lab grown meat is harmful to my health. 3.25 1.22 Note: Scores were recorded on a 5-point Likert-type scale (1 = very unlikely, 5 = very likely). Measures of uncertainty. To measure uncertainty, five questions adopted from

Li and Brossard (2012). Participants were asked their level of agreement with response options presented on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree).

The 5-item index had an alpha reliability of .595. Two items were removed as shown in

Table 3.6 and reliability increased to an acceptable level with Cronbach’s a = .733.

Table 3.6 Original Items Used to Measure Uncertainty Scientists agree with each other on the potential consequence of lab grown meat. Scientists are sure about the impact of lab grown meat on human health. There isn’t enough scientific evidence regarding the safety of lab grown meat.* There needs to be more scientific work done to examine whether lab grown meat is safe.* Scientists are uncertain about whether lab grown meat is dangerous to human beings.* Note: *Item is reverse coded. Strikethrough indicates item was removed in final analysis.

Descriptive statistics for these items are provided in Table 3.7. The summed and averaged responses resulted in a grand mean of 2.83 (SD = 1.08).

Texas Tech University, Kellie Anne Boykin, December 2019

Table 3.7 Descriptive Statistics for Measures of Uncertainty (N = 238) M SD There isn’t enough scientific evidence regarding the safety of lab 2.41 1.23 grown meat.* There needs to be more scientific work done to examine whether 3.32 1.45 lab grown meat is safe.* Scientists are uncertain about whether lab grown meat is 2.76 1.31 dangerous to human beings.* Note: * indicates item was reverse coded. Scores were recorded on a 5-point Likert-type scale (1 = very unlikely, 5 = very likely).

Intention to share content. Behavioral intention to seek out information about lab grown meat and share the blog post seen on social media were asked in a three- question item adopted from (Steede, 2018). Responses were reported on a 7-point Likert- type scale (1 = strongly disagree, 7 = strongly agree). A post hoc reliability analysis was calculated with Cronbach’s a = .853.

Table 3.8 Descriptive Statistics for Intention to Share Content (N = 238) M SD This blog post makes me want to seek out information about lab 4.57 1.81 grown meat. This blog post is something I would share on social media. 3.84 4.11 This blog post is something my social media followers would be 4.11 1.67 interested in.

Intention to consume lab grown meat. A three-item question set was asked to measure a consumer’s intention to consume and purchase lab grown meat. The questions were adopted from Wilks and Phillips (2016) survey of attitudes toward lab grown meat.

Respondents were asked questions such as, “How likely are you to try lab grown meat at least once?” The response options were Likert-type questions (1 = not at all Likely to 5 = extremely likely) with a reliability calculated post hoc with Cronbach’s a = .853. The results were summed and averaged resulting in a grand mean of M = 2.43 (SD = 1.32). 37

Texas Tech University, Kellie Anne Boykin, December 2019

Participants were then asked to imagine that lab grown meat was available on retail shelves alongside conventional meat and plant-based meat. Using the question from

Bryant’s et al. (2019) study regarding the likeliness to purchase and consumer lab grown meat versus conventional or plant-based meat, participants were asked: “You are at a dinner party with family and friends. The host has prepared a variety of meat options for the guests. Please select your preferred meat choice.” Options for response were

“conventional meat,” “lab grown meat,” “plant-based meat,” or “none of the above choice.”

The descriptive statistics for this construct can be seen in Table 3.9. The grand mean for this construct was 2.43 (SD = 1.32). This grand mean indicated overall respondents were somewhat unlikely to try, purchase, or consistently eat lab grown meat.

Table 3.9 Descriptive Statistics for Behavioral Intention (N = 238) M SD How likely are you to try lab grown meat at least once? 2.66 1.48 How likely are you to eat lab grown meat as a replacement for 2.32 1.42 conventional meat? How likely are you to purchase lab grown meat regularly? 2.30 1.37 Note: Scores were recorded on a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree).

Benefit perceptions. Risk and benefit perceptions were measured using “a set of application-specific statements” (Kim et al., 2014, p. 969). Kim et al. (2014) used multiple measures to measure risk and benefit perception of nanotechnology. Because lab grown meat is a new technology similar to nanotechnology, these questions were adapted for the current study.

Benefit perception was measured using three questions with a 7-point Likert-type scale (1 = strongly disagree and 7 = strongly agree). Participants indicated their 38

Texas Tech University, Kellie Anne Boykin, December 2019 agreement or disagreement to the following items: 1) I believe lab grown meat is good for the environment. 2) I believe lab grown meat is good for animals. 3) I believe lab grown meat is good for future generations of people. Questions were adapted from Kim et al. (2014) who reported a Cronbach’s a = .91. Reliability calculated post hoc resulted

Cronbach’s a = .878. Descriptive statistics for benefit perceptions are displayed in Table

3.10. The responses were summed and averaged for a grand mean of 3.85 (SD = 1.62), which indicated participants were somewhat neutral toward perceived benefits of lab grown meat.

Table 3.10 Descriptive Statistics for Benefit Perception (N = 238) M SD I believe lab grown meat is good for animals. 4.03 1.80 I believe lab grown meat is good for the environment. 3.79 1.77 I believe lab grown meat is good for future generations of 3.74 1.84 people. Note: Scores were recorded on a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree).

Procedure

The experimental design for this study was embedded in an online survey instrument created in Qualtrics (Appendix B). Upon completion of instrument design, the researcher worked with MSG to program redirect links into the Qualtrics survey for distribution. The redirect links were essential to record participants’ unique identification numbers so that they could receive compensation for participation.

When participants first accessed the study, they viewed a brief description of the research purpose and agreed to take part in the survey instrument by clicking “Accept.”

They then answered questions regarding social media use were asked first followed by

Texas Tech University, Kellie Anne Boykin, December 2019 measures of food technology neophobia. Attitudes toward lab grown meat were measured using five questions.

Participants then viewed one of the three randomly assigned themed message stimuli. After reading through the assigned blog post, a manipulation check question stated “I have read the blog and am ready to continue the survey” paired a forced response option.

Participants then answered questions to evaluate the blog post message and their uncertainty, risk perceptions, and benefit perceptions regarding lab grown meat.

Participants were asked how likely they were to try, purchase, and eat lab grown meat.

They were then asked one questions about their meat purchasing behaviors followed by demographic questions to report gender, age, education level, income, political view, and political party.

Message Testing

Before launching the full survey with a nationally distributed sample, message testing was conducted to ensure participants could distinguish between the three stimuli.

Graduate students participated in this stage of message testing. Participants were recruited via an e-mail sent to current graduate students and 29 responses were collected.

For the message testing phase, the three researcher-developed blog posts were embedded in a Qualtrics survey instrument. The blog posts were randomly shown to participants who were asked to read each blog and report whether the overall message was in support of lab grown meat, against lab grown meat, balanced/neutral toward lab grown meat, or whether they were unsure.

Texas Tech University, Kellie Anne Boykin, December 2019

Twenty-five participants correctly identified the support lab grown meat frame, and 23 of the participants correctly identified the against lab grown meat frame.

Responses identifying the neutral blog post were inconsistent. Only 18 participants correctly identified the neutral frame, with 10 respondents incorrectly identifying the neutral themed post. A sentence was added at the beginning highlighting a balanced viewpoint with no opinion for or against. This message was then shown to a select group of graduate students who deemed it was appropriate for subsequent testing.

Data Analysis

Participant description. Data analysis was carried out using IBMÒ SPSSÒ

Statistics version 25. The demographic characteristics collected to describe survey participants were: gender, age, education, meat consumption habits, income level, political views, political party (Table 3.11). Some demographic questions were missing several responses. This is likely because the demographic questions were asked at the end of the survey and the survey was not programmed to force responses.

A majority of respondents were female (n = 182, 76.2%). The age of respondents varied from 18-80 years old with the mean age of respondents being 45.7 years old.

Most survey respondents had an undergraduate degree (n = 83, 42.7%) or 33.5%

(n = 80) had completed high school. No participants reported “no education.” The majority of respondents were meat eating individuals (n = 184, 77.3%) with 7.1% eating only white meat (n = 17), 5.4% reported consumption habits outside of the listed options

(n = 13), 4.2% of respondents were vegetarian (n = 10), 4.2% reported being pescatarian

(n = 10), and only 2 respondents (0.8%) reported being vegan.

Texas Tech University, Kellie Anne Boykin, December 2019

The income level of respondents was primarily $20,000 to $39,999 (n = 55,

23.1%) and less than $20,000 (n = 53, 22.3%). The majority of participants responded they were moderate in their political views (n = 78, 32.8%). While 21 participants (8.8%) chose not to answer, the majority of participants identified their political party as democrat (n = 93, 39.1%). The rest of responses were split with 66 respondents (27.7%) being republican and 58 respondents (24.4%) identifying as an independent.

Table 3.11 Demographic Characteristics of Survey Respondents n % Gender Female 182 76.5 Male 56 23.5

Education Some High School 10 4.2 Completed High School 80 33.6 Technical Qualification/Trade 37 15.5 Certification College/Undergraduate Degree 83 34.9 Postgraduate Degree 28 11.8

Meat Consumption Habits Meat Eating 184 77.3 Eat White Meat Only 17 7.1 Other 13 5.4 Vegetarian 10 4.2 Pescatarian 10 4.2 Vegan 2 0.8

Income Less than $20,000 53 22.3 $20,000 to $39,999 55 23.1 $40,000 to $59,999 45 18.9 $60,000 to $79,999 37 15.5 $80,000 to $99,999 15 6.3 $100,000 or more 31 13.0

Political Views Very Conservative 34 14.3 Somewhat Conservative 37 15.5 Moderate 78 32.8 Somewhat Liberal 39 16.4 Very Liberal 31 13.0 Prefer Not to Answer 19 8.0

Texas Tech University, Kellie Anne Boykin, December 2019

Table 3.11 Continued Political Party Democrat 93 39.1 Republican 66 27.7 Independent 58 24.4 Prefer Not to Answer 21 8.8

Texas Tech University, Kellie Anne Boykin, December 2019

CHAPTER IV

RESULTS

Overview

The purpose of this study was to first, monitor the Twitter conversation surrounding lab grown meat using Meltwater, and second, to examine the influence of pre-existing attitudes and themed messages on public perceptions of lab grown meat.

This chapter contains the results of both parts of the study and answers the research objectives and questions presented below:

Research Objectives for Part I:

RO 1: Determine the potential social reach and volume of lab grown meat content

on Twitter.

RO 2: Establish the percent of positive, negative, and neutral messages about lab

grown meat on Twitter.

RO 3: Identify the trending themes about lab grown meat on Twitter.

RO 4: Determine the top posters on Twitter providing content about lab grown

meat.

Research Questions for Part II:

RQ 1: How does the frequency of social media use affect intention to share

content and message evaluation?

RQ 2: How is pre-existing attitude toward lab grown meat related to risk and

benefit perceptions?

RQ 3: How is neophobia to new food technology related to risk and benefit

perceptions of lab grown meat?

Texas Tech University, Kellie Anne Boykin, December 2019

RQ 4: What influence does the message theme of a blog post have on message

evaluation and behavioral intention toward lab grown meat?

RQ 5: What influence does the message theme of a blog post have on risk and

benefit perceptions of lab grown meat?

RQ 6: What influence does the message theme of a blog post have on perceptions

of uncertainty regarding lab grown meat?

RQ 7: How is neophobia to new food technology related to intention to consume?

RQ 8: What influence does the message theme of a blog post have on intention to

consume?

Part I Results

RO 1: Determine the potential social reach and volume of lab grown meat on Twitter.

Potential social reach was a measure of potential viewers of the Twitter content, and social media volume was the number of Twitter posts about lab grown meat. During the study’s time frame, potential reach remained relatively low except in several sporadic spikes, while social volume was much more variable. No pattern existed as to when there were spikes in potential social reach and social volume (Figure 4.1).

Texas Tech University, Kellie Anne Boykin, December 2019

Figure 4.1. Social media volume and potential reach of lab grown meat on Twitter

Texas Tech University, Kellie Anne Boykin, December 2019

RO 2: Establish the percent of positive, negative, and neutral messages about lab grown meat on Twitter.

Sentiment was a measure of attitude associated with each tweet. Meltwater uses natural language processing to label messages as positive, negative, or neutral. Of the

11,100 posts analyzed via our search parameters, 33.1% of posts were positive, 19.9% of posts were negative, and 47.0% of posts were neutral toward lab grown meat. Figure 4.2 depicts a Meltwater chart used to visually represent the division of sentiment scores.

Figure 4.2. Positive, negative, and neutral sentiment of Twitter posts about lab grown meat.

RO 3: Identify the trending themes about lab grown meat on Twitter.

Top trending themes were words most commonly found when the keyword search was completed. Meltwater identified seven top terms during the selected search dates

(Table 4.1). Each identified term was related to lab grown meat. “Grown meat,” “lab grown meat,” and “clean meat” made up 76% of the analyzed posts.

Texas Tech University, Kellie Anne Boykin, December 2019

Table 4.1 Top Seven Trending Themes and Number of Mentions on Twitter Trending Theme Number of Mentions Grown meat 4,400 Lab grown meat 1,400 Clean meat 1,100 Meat 709 Future 702 Environment 393 Plant 373

RO 4: Determine the top posters on Twitter providing content about lab grown meat. Meltwater considered the top posters to be the Twitter accounts that posted about lab grown meat and had the greatest potential social reach. Table 4.2 depicts the top 10 accounts posting about lab grown meat that have the greatest reach of content. Reach is defined as the number of potential viewers. The top 10 Twitter accounts were all news publications or publishers of educational content. The New York Times had the greatest potential reach, which was more than double the next highest-reaching Twitter account.

Table 4.2 Top Twitter Accounts and Sum of Potential Reach Twitter Account Sum of Potential Reach New York Times @nytimes 42,000,000 The Wall Street Journal @WSJ 16,200,000 Times of India @timesofindia 11,300,000 TED Talks @TEDTalks 11,100,000 WIRED @WIRED 10,300,000 The Guardian @guardian 7,430,000 Food & Wine @foodandwine 6,630,000 Sky News @SkyNews 4,650,000 Scientific American @sciam 3,670,000 Newsweek @Newsweek 3,300,000

Part II Results

Each participant was randomly assigned a themed blog post (Table 4.3). There were slightly more participants (n = 88, 36.8%) who saw the positive stimuli. The 48

Texas Tech University, Kellie Anne Boykin, December 2019 remaining participants saw either the negatively themed blog (n = 74, 31.0%) or the neutral themed blog (n = 77, 32.2%). The inconsistency in how many saw each stimulus is due to 61 responses being removed due to not meeting the quality check questions.

Table 4.3 Frequency of Assigned Stimuli Among Participants (N = 238) n % Against Lab Grown Meat 74 31.0 Neutral 77 32.2 Support Lab Grown Meat 88 36.8

RQ 1: How does the frequency of social media use affect intention to share content and message evaluation?

A point biserial correlation was calculated to observe the relationship between social media use, frequency, and intention to share content and message evaluation

(Table 4.4). The “social media use” construct was operationalized into 1 = daily and 2 = not daily. A low, negative association was found between message evaluation and social media use (rpb = -.12) and between intention to share and social media use (rpb = -.22)

(Davis, 1971). According to Davis (1971), a substantial, positive relationship was found between message evaluation and intention to share (rpb = .57).

Table 4.4 Means, Standard Deviations, and Intercorrelations for the Relationship Between Frequency of Social Media Use and Intention to Share Content and Message Evaluation Measure M SD 1 2 3 1. Social Media Use 1.32 .47 - 2. Message Evaluation 4.60 1.08 -.12 - 3. Intention to Share 4.17 1.56 -.22* .57* - Note: * indicates correlation is significant at p < .01.

Texas Tech University, Kellie Anne Boykin, December 2019

RQ 2: How is pre-existing attitude toward lab grown meat related to risk and benefit perceptions?

To answer RQ2, a Pearson product moment correlation was calculated between attitude toward lab grown meat, risk perception, and benefit perception (Table 4.5).

According to Davis (1971), substantial association was found between attitude and benefit perception (r = .61), attitude and risk perception (r = -.53), and risk and benefit perceptions (r = -.51).

Table 4.5 Means, Standard Deviations, and Intercorrelations for the Relationship Between Attitude Toward Lab Grown Meat and Risk and Benefit Perceptions Measure M SD 1 2 3 1. Attitude Toward Lab 3.53 1.06 - Grown Meat 2. Benefit Perception 3.85 1.62 .61 - 3. Risk Perception 3.29 1.14 -.53 -.51 - Note: All correlations were significant at p < .01.

RQ 3: How is neophobia to new food technology related to risk and benefit perceptions of lab grown meat?

In order to address the relationship between neophobia to new food technology and risk and benefit perceptions, a Pearson product moment correlation was run with an alpha level set at .01 a priori (p < .01). The result of the correlation between food neophobia and risk perception was r = .52, which according to Davis (1971) equated to a moderate correlation. The result of the correlation between food neophobia and benefit perception was r = -.31. All correlations were significant at p < .01.

RQ 4: What influence does the message theme of a blog post have on message evaluation and behavioral intention toward lab grown meat?

To answer RQ4, a one-way ANOVA was conducted for each dependent variable.

Table 4.6 illustrates that no significant difference was found between the message theme that was viewed and how participants evaluated the message (F = .59, p = .55). The group

Texas Tech University, Kellie Anne Boykin, December 2019 means (Table 4.7) showed that those who viewed the neutral themed blog post reported slightly more agreement regarding trust, reliability, ease of understanding, and accurateness (M = 4.67, SD = 1.09), but this was not statistically significant when compared to the evaluation of the other message themes.

Table 4.6 ANOVA of the Effects of Message Themes on Message Evaluation Source SS df MS F p Between Groups 1.39 2 .69 .59 .55 Within Groups 275.45 235 1.172

Table 4.7 Means of Message Evaluation (N = 238) Message Theme M SD Support Lab Grown Meat 4.50 1.09 Neutral 4.67 1.09 Against Lab Grown Meat 4.64 1.07

Another one-way ANOVA (Table 4.8) reported statistically significant differences between the blog post theme viewed and behavioral intention (F = 5.02, p =

.01). The group means (Table 4.9) showed that those who viewed the blog post themed against lab grown meat were less likely to seek out information about lab grown meat, share the blog post on social media, or think the blog post would interest their followers on social media (M = 2.03, SD = 1.21).

Table 4.8 ANOVA of the Effects of Message Themes on Behavioral Intention Source SS df MS F p Between Groups 16.99 2 8.49 5.02 .01 Within Groups 397.42 235

Table 4.9 Means of Behavioral Intention (N = 238) Message Theme M SD Support Lab Grown Meat 2.64 1.27 Neutral 2.56 1.41 Against Lab Grown Meat 2.03 1.21 51

Texas Tech University, Kellie Anne Boykin, December 2019

A post hoc analysis using Bonferroni comparison was calculated to understand where the significance occurred between the message themes. Statistically significant difference was found between support lab grown meat and against lab grown meat themed (p = .01) and against lab grown meat and neutral themes (p = .04). There was no statistically significant difference found between support lab grown meat and neutral blog posts (p = 1.00).

RQ 5: What influence does the message theme of a blog post have on risk and benefit perceptions of lab grown meat?

Two one-way ANOVA calculations were calculated to answer RQ5. The first comparison was between the viewed blog post theme and risk perception. As Table 4.10 displays a statistically significant difference in risk perception was found between the three message themes (F = 3.44, p = .03). The group means (Table 4.11) showed that those who viewed the blog themed against lab grown meat were somewhat more likely to perceive risk (M = 3.58, SD = 1.07) than those that viewed the neutral (M = 3.19, SD =

1.20) or the support lab grown meat (M = 3.13, SD = 1.12) themed blog post.

Table 4.10 ANOVA of the Effects of Message Themes on Risk Perception Source SS df MS F p Between Groups 8.79 2 4.39 3.44 .03 Within Groups 300.27 235 1.28

Table 4.11 Means of Risk Perception (N = 238) Message Theme M SD Support Lab Grown Meat 3.13 1.12 Neutral 3.19 1.20 Against Lab Grown Meat 3.58 1.07

Texas Tech University, Kellie Anne Boykin, December 2019

In order to identify where the significance existed, a Bonferroni comparison was calculated, which found a statistically significant difference existed between the support lab grown meat theme and the against lab grown meat theme (p = .045).

Table 4.12 shows the one-way ANOVA calculated to compare the viewed message theme on benefit perception showed a significant difference between the two variables (F = 7.08, p < .001). The group means (Table 4.13) revealed that participants who viewed the blog post themed against lab grown meat somewhat disagreed with the statements of benefits of lab grown meat (M = 3.28, SD = 1.60). However, those who viewed the support lab grown meat themed blog post on average neither agreed nor disagreed with benefit statements about lab grown meat (M = 4.14, SD = 1.57).

Table 4.12 ANOVA of the Effects of Message Themes on Benefit Perception Source SS df MS F p Between Groups 35.26 2 17.63 7.08 .00 Within Groups 585.06 235 2.49

Table 4.13 Means of Benefit Perception (N = 238) Message Theme M SD Support Lab Grown Meat 4.14 1.57 Neutral 4.06 1.56 Against Lab Grown Meat 3.28 1.60

In order to shed light on where the significant difference existed, a post hoc

Bonferroni comparison was run. The comparison showed a statistically significant difference between against lab grown meat and support lab grown meat (p = .002) and between against lab grown meat and neutral (p = .007) themed blog posts.

Texas Tech University, Kellie Anne Boykin, December 2019

RQ 6: What influence does the message theme of a blog post have on perceptions of uncertainty regarding lab grown meat?

To address RQ6, a one-way ANOVA was calculated to assess the interaction between the message theme viewed and perceptions of uncertainty toward lab grown meat. As shown in Table 4.14, no significant difference between message themes and measures of uncertainty was observed (F = 1.46, p = .23). The group means (Table 4.15) showed that those who viewed the against lab grown meat (M = 2.68, SD = 1.19) themed blog more strongly disagreed with statements of certainty than those who viewed the other themed blogs.

Table 4.14 ANOVA of the Effects of Message Themes on Measures of Uncertainty SS df MS F p Between Groups 3.38 2 1.69 1.46 .23 Within Groups 271.45 235 1.16

Table 4.15 Means of Measures of Uncertainty (N = 238) Message Theme M SD Support Lab Grown Meat 2.97 .95 Neutral 2.81 1.09 Against Lab Grown Meat 2.68 1.19

RQ 7: How is neophobia to new food technology related to intention to consume?

The relationship between new food technology neophobia and intention to consume lab grown meat was measured with a Pearson product moment correlation. The result of the correlation was r = -.39 and was significant at p < .01. According to Davis

(1971) this indicated a negative, moderate correlation between the two constructs.

RQ 8: What influence does the message theme of a blog post have on intention to consume?

The interaction between message theme of a blog post and intention to consume lab grown meat was measured with a one-way ANOVA. Table 4.16 shows there was a 54

Texas Tech University, Kellie Anne Boykin, December 2019 statistically significant difference between message themes and intention to consume (F =

5.02, p = .007).

Table 4.16 ANOVA of Effects of Message Theme on Intention to Consume Source SS df MS F p Between Groups 16.99 2 8.49 5.02 .007 Within Groups 397.42 235 1.69

A calculation of the group means, shown in Table 4.17, indicated that those who viewed the message theme against lab grown meat (M = 2.03, SD = 1.21) more strongly disagreed that they intended to consume lab grown meat than those who viewed the other two message themes. A Bonferroni post hoc analysis showed the significant differences existed between against lab grown meat and neutral (p = .04) and between against lab grown meat and support lab grown meat (p = .01) themed blog posts.

Table 4.17 Means of Intention to Consume (N = 238) Message Theme M SD Support Lab Grown Meat 2.64 1.27 Neutral 2.55 1.41 Against Lab Grown Meat 2.03 1.21

Texas Tech University, Kellie Anne Boykin, December 2019

CHAPTER V

CONCLUSION, DISCUSSION, AND RECOMMENDATIONS

Overview

This two-part study monitored the social media conversation about lab grown meat on Twitter, then utilized a between-subjects experimental research design to examine the influence of pre-existing attitudes and themed messages on public perceptions of lab grown meat. This chapter provides the conclusions and discussions of the results provided in Chapter IV. Recommendations for future research and practice are also provided.

Research Purpose and Questions

The purpose for the initial quantitative analysis aspect of this study was to describe Twitter conversations about lab grown meat. The specific research objectives were as follows:

RO 1: Determine the potential social reach and volume of lab grown meat

content on Twitter.

RO 2: Determine the percent of positive, negative, and neutral messages about

lab grown meat on Twitter.

RO 3: Determine the trending themes of lab grown meat on Twitter.

RO 4: Determine the top posters on Twitter providing content about lab grown

meat.

The purpose of the second part of the study was to examine the influence of pre- existing attitudes and themed messages on public perceptions of lab grown meat.