Content Analysis of Video Game Loot Boxes in the Media a Thesis

Content Analysis of Video Game Loot Boxes in the Media

A thesis presented to

the faculty of the Scripps College of Communication of Ohio University

In partial fulfillment

of the requirements for the degree

Master of Science

Raymond T. Humienny

May 2019

© 2019 Raymond T. Humienny. All Rights Reserved. This thesis titled

Content Analysis of Video Game Loot Boxes in the Media

by

RAYMOND T. HUMIENNY

has been approved for

the E.W. Scripps School of Journalism

and the Scripps College of Communication by

Hans Meyer

Associate Professor of Journalism

Scott Titsworth

Dean, Scripps College of Communication

ii Abstract

HUMIENNY, RAYMOND T., M.S., May 2019, Journalism

Content Analysis of Video Game Loot Boxes in the Media

Director of Thesis: Hans Meyer

Throughout the relatively nascent course of games media scholarship, representation of video games within popular mainstream media tends to suggest an antagonistic relationship between those familiar and unfamiliar with video games respectively. Yet, this outlook fails to acknowledge the content of popular gaming media that can be equally critical of the representation of games in reporting. For instance, within the past two years, reports pertaining to video game “loot boxes” have not only shown that reward systems in certain games can structurally mimic online gambling, but games and mainstream media can cohabitate in this reporting arena. Given our nascent understanding of gaming media from a mainstream perspective, this study examines how gaming and mainstream news outlets comparatively frame this “loot box” rewards practice. A content analysis of 274 articles containing the term “loot box(es)” revealed similarities wherein both types of media outlets framed “loot boxes” with political messages, references to gambling and cast some form of normative judgment. Traditional news writing provided fewer overall frames than more opinionated types of writing.

Political intervention was the greatest predictor of frames assigned by both media.

Overall, the internal regulation of “loot boxes” and games industry’s opposition to government-assisted regulation are the strongest implications that warrant future study of this controversy.

iii Dedication

For my family, Angela and Newswire.

iv Acknowledgments

I would like to thank Amrik Singh and Kyle Linville for their work as coders on this project. As friends I game with frequently, their input for the codebook was invaluable. My thanks also go to my committee members Dr. Michael Sweeney, Dr.

Aimee Edmondson and committee chair Dr. Hans Meyer for the support in guiding this thesis.

The E.W. Scripps School of Journalism graduate program provided the basis for this endeavor. Additionally, IBM’s SPSS was used to generate all statistical analyses within this study.

v Table of Contents

Page

Abstract ...... iii Dedication ...... iv Acknowledgments ...... v Introduction ...... 1 Literature Review ...... 5 RQ1a: What themes generally define the “loot box” controversy?...... 7 RQ1b: How are they related? ...... 7 Legal History on Video Games ...... 7 Framing Effects ...... 19 Games Media Research: Uses and Gratifications...... 24 Gaming Pathology ...... 32 Methods ...... 46 Results ...... 48 Discussion ...... 65 RQ1a: What themes generally define the “loot box” controversy?...... 66 RQ1b: How are they related? ...... 67 RQ2: What comparisons can be made between MSM and GM reports on “loot boxes?” ...... 68 Political Intervention: An Unforeseen Conclusion ...... 70 Conclusions ...... 73 References ...... 76 Appendix: Loot Box Codebook ...... 87

vi Introduction

Mainstream media (MSM) reports about video games and their effects tend to imply disaster, but this is only one half of the story. While the MSM have kept a close view on the controversial aspects of gaming for a variety of framing and agenda setting purposes (McKernan, 2013), they have not so without a reasonable warrant for such fears

(Entm’t Software Ass’n v. Swanson, 2008). By that same citation, MSM coverage of games also highlight the socially beneficial characteristics of the virtual platform with references to exercise games that utilize motion controls (McKernan, 2013); negative frames of the gaming industry are frequented due to their controversial tone and the issues they evoke from critics (e.g. underage gambling, sexual objectification of women and girls, aggressive attitudes among children; Funk et al., 2017). While games scholarship continues to examine video game culture and its many traits from the inside- out (Shaw, 2010), the representation of games in popular media will—inevitably— continue to fluctuate between good and bad press (Copenhaver et al., 2017).

Yet recent events regarding Star Wars Battlefront II, “a ‘Star Wars-themed online casino’” (Kim, 2017; Dingman, 2017) have brought these issues to a head while also shedding light on a taken-for-granted virtual economy within video games. In short, one qualifier of a video game is a reward upon completion of a task, whether it be the game in its entirety or a single assignment. The content of these rewards varies from game to game, but a commonly practiced, contemporary method of rewarding the player in online multiplayer games involves randomly generating a single prize from a small pool of potential winnings. These pools of rewards are referred to as “loot boxes,” “loot crates”

1 or some genre-specific variation of the name (e.g. Tom Clancy’s Rainbow Six Siege, a tactical shooter in which the playable characters are “Operators” of world-wide counterterrorism units, names their loot boxes “Alpha Packs”). These loot boxes can be purchased in one of two ways: (1) by earning in-game currency through the completion of tasks, challenges and other activities; or (2) by purchasing in-game currency using real-world currency.

Specifically, the more predatory loot box practices come into question here and subsequently fit the criteria for this study. Whether players of Battlefront II spend real- world or in-game currency to obtain their loot boxes, the rewards dealt by such loot crates bolster players’ performances in online multiplayer—pitting statistically-enhanced players against those who did not reap the benefits of obtaining Battlefront II loot box rewards (Good, 2018). Coincidentally, the set of rewards given by a loot box cannot be locked into a selection; the odds of obtaining a single, desired item are assigned by the game’s developers, which has shown to result in difficult consumer relationships

(Lawrence, 2017), as detailed below.

The emergence of randomized prizes known as “loot boxes” begins with

Blizzard’s Overwatch, an objective-based first-person shooter with heavy emphasis on teamwork released on May 24, 2016. Overwatch’s loot box rewards are exclusively aesthetic customizations the player can apply to various playable characters; in no way do these rewards influence the outcome or individual performance of a game. However,

EA’s Star Wars Battlefront II, the aforementioned virtual casino disguised as a third- person shooter, drew not only consumer backlash but also political counteraction after

2 including loot box rewards that directly affected gameplay. As a result, players reported an unbalanced playing field due to the advantages of a select few who obtained statistical advantages from loot boxes (Good, 2018).

The intersubjective perception among gamers of corporate deception is the primary complaint of these “loot box” systems: on top of the base purchase of a game, more contemporary developers incentivize players to make small, additional purchases

(e.g. typically $0.99 to $5.00 USD) within their game, in what are known as

“microtransactions.” These small purchases can vary from character aesthetic customizations to in-game currency and booster packs for experience points. While this practice is common—and typically accepted—for games that do not require a base purchase to generate revenue (i.e. free-to-play games), their implication along with base purchases outwardly demonstrates profit-maximization. Thus, considering that incentives already exist for players to purchase more for games they have already bought, popular opinion even within gaming media (GM) outlets is heavy opposition to loot boxes, categorizing them as “deceptive,” “hostile to consumers” and “downright manipulative”

(Yin-Poole, 2018; Kim, 2018; Alexandra, 2017).

In closing, I would fail my audience if I did not mention that gaming is, has been and will be my number one hobby. Thus, this project strikes an introspective chord: it would be naïve to presume that the MSM have had it out for gaming over the past few decades—and that, somehow, this study would act as a final nail in the coffin of unruly games criticism from the outside-in. Such a widespread vendetta against a profitable form of entertainment simply because of unfamiliarity would misrepresent the arguments

3 within this study. Yet, I should not presume that games are totally exempt from critical scrutiny, knowing full well that games are communicative texts with pervasive messages

(Brown v. Entm't Merchs. Ass'n, 2011), even if this ability is not utilized to its fullest

(Bogost, 2015). Instead of polarizing these two ends further from each other, this research aims to highlight where there is agreement and what can be learned from the differences between MSM and GM.

4 Literature Review

Although the following literature review refers heavily to different genres of gaming, it must be understood that the games addressed explicitly by this study are no more than any game containing “loot boxes,” reported by the media or otherwise.

However superficially accurate this depiction may be, nonetheless, the lay-reader comes to mind when imagining what exactly a “video game” is. In some sense, the simplest explanation is the best: a video game is an interactive piece of entertainment consisting of frontend (e.g. user interfaces, overlays, gameplay) and backend media (e.g. operating systems, game engines, GPU drivers) that can connect its user to the internet to play with other. As referenced later though, a “video game” may qualify as a cabinet arcade machine repurposed for the functions of gambling (Helton v. Hunt, 2003), but this definition is disingenuous for loot box concerns. The video games examined by this study are not designed with a primary gambling purpose in mind such that their only function is to entice players to spend more money on a single game; that contention is what lawmakers in Hawaii, California, Washington, Indiana and Minnesota argue is moving from the peripheral to the purview of games with loot boxes (Rosenberg, 2017). Thus, working a way backward from loot boxes and other microtransactions more accurately defines the general category—not genre—of games appropriate for this study.

As reported by Syracuse Law Review, loot boxes, paired with several other micro-transactions within games in 2017, have earned the games industry $30 billion in total profit (Jerauld, 2018). This number refers more specifically to larger developers such as Blizzard Activision with their popular release titles Overwatch and Destiny 2,

5 first-person shooter games that emphasize two differing functions: team-oriented and fast-paced gameplay versus role-playing in massive multiplayer and grinding for loot respectively. Upon further examination, some limitations of this study become clearer:

Overwatch’s loot boxes are designed to reward players only four aesthetic prizes per box, while Destiny 2’s “engram” system (i.e. loot box equivalent) rewards only one item per engram that could potentially gameplay in limited circumstances (Destiny 2, 2018). Thus, how could regulators qualify the amount of attention paid to loot boxes by their developers? Congruently, Facebook and Apple games, many of which consisting of free- to-play models of puzzle games, sustain their revenue stream by microtransactions within their games (Sinclair, 2018), rendering a reasonable attentive to loot boxes a warranted concern for developers to produce profits. Such microtransactions may consist of randomized reward pools refer to as “loot boxes” or some variant of the terms.

Nonetheless, unless the model of a game is that of a free service to the player, there is no limited genre of games that are defined by their loot boxes. Rather, developers are keen to define their loot boxes themselves and what purpose they serve in-game.

Coincidentally, a proxy search using the Ohio University ArticlesPlus peer-review browser returns relatively few explicit references to “loot boxes” in video games and gaming media scholarship, even within journals such as Computers in Human Behavior,

Games and Culture and The Journal of Games Criticism. This lack of research, however, was expected because video game “loot box” reward systems – whether they have been referred to as “loot boxes” or randomized rewards – have only recently been scrutinized by the MSM and gaming media with considerable warrant. The only existing research

6 containing the term “loot boxes” or any mention of randomized rewards in video games examines the relationships between spectating eSports, playing games and vulnerability to online gambling and betting (Macey & Hamari, 2017). Although this study mentions the influence of “loot box” reward systems in contemporary games, there has not been an extensive study of their representation in the media. A re-examination of how video games are perceived by the MSM and gaming media would shed light on how media reports frame “loot boxes” in video games. Thus, the author posits two broad research questions as the basis for this study:

RQ1a: What themes generally define the “loot box” controversy?

RQ1b: How are they related?

To further contextualize these broad research questions, this study will apply the legal history of video games, framing effects literature and research on uses and gratifications to understand how the MSM, GM and various types of games players interact with and represent video games.

Legal History on Video Games

As referenced by Kim (2017) and Alexandra (2017), political actors scattered worldwide are preemptively combatting loot boxes with the allegation that such a reward system is structurally and functionally like online gambling. Although loot box rewards—unlike gambling—never present the player with a true tie or loss (i.e. players are compensated for duplicate rewards already received), extended systems such as skin trading (Grayson, 2018) provide users incentive to purchase loot boxes with actual currency in order to turn a small profit. Despite this line of argumentation, the FTC has

7 recently warranted an investigation of loot boxes and whether games containing them must denote a rating’s category that marks their presence (Gach, 2018).

Thus, the concurrent issue with loot boxes deals primarily with who ought to regulate the content within games. The views that align with this study argue more in favor of inside regulation performed by the games industry itself, rather than an outside, political entity (Rosenberg, 2017; Willett, 2018). For example, the Entertainment

Software Rating Board (ESRB) is a non-profit, self-regulatory organization that was created in 1994 to coax political concern over the ratings of violent and sexually-explicit video games (Sinclair, 2018). Specifically, the ESRB provides ratings for video games using five categories: (1) “E” for everyone, or appropriate for all ages; (2) “E 10+” for everyone ages 10 and up; (3) “T” for teen, or ages 13 and up; (4) “M” for mature, or ages

17 and up; and (5) “AO” for adults only, or ages 18 and up. The criteria that defines each of these categories are rather extensive, generally building up from minimal cartoon depictions of violence (e.g. E-E10+) to uncanny representations of explicit behavior (e.g.

M-AO). Nonetheless, the state’s challenge to these lattermost ratings, particularly “M” and “AO,” typically involve adding barriers for the purchase of such video games.

Self-regulations on free speech by the ESRB can be seen in the most noteworthy video game case Brown v. The Entertainment Merchants Association (EMA; 2011). This case regarded a 2005 law put forth by former California Governor Arnold

Schwarzenegger that prohibited the sale of violent video games to minors throughout the state. Under subsequent Governor Edmund “Jerry” Brown Jr., the appeals for the case reached the Supreme Court, in which a 7-2 decision ruled in favor of the EMA that

8 challenged this legislation. The most commonly cited reason for this ruling is video games’ expressive protection according to the First Amendment; much like books, movies and television, video game messages could contain pervasive meanings and interpretations pertaining to politics, social issues and creative expressions. Additionally, the late Justice Antonin Scalia oversaw the ruling opinion that the ESRB’s existence in the games industry rightfully fulfilled any speculative restrictions that California law would have applied (Brown v. Entm’t Merchs. Ass’n, 2011).

Although, it may be of interest for the reader to consider the opinion of Justice

Samuel Alito in the case of Brown v. EMA (2011), noting that a law worded more precisely would have eluded rejection (Willett, 2018). Albeit speculative, Willett (2018) notes that this consideration could apply to the ESRB’s response to the inclusion of loot boxes within video games. In short, the ESRB does not consider loot boxes to be or to contain any explicit or implicit form of gambling, likening it instead to card collection

(Rosenberg, 2017; Sinclair, 2018; Willett, 2018). Yet, this analogy ignores the premise that particular values are assigned to specific cards in a pack and that spending money for a chance at a desired card is a considerable expenditure for the consumer (Rosenberg,

2017). Arguably, these expenditures could be risks, given that a single consumer could spend egregious sums of cash to obtain one particular item in a game of chance (Jerauld,

2018). Consequently, the ESRB has followed-up with their response toward loot boxes with the denotation of “In-Game Purchases” on games that feature such microtransactions (Willett, 2018). Yet, such an action does not seek to qualitatively parse through the different functionalities of loot box reward systems and those that abuse the

9 ease of consumer manipulation (Rosenberg, 2017). The latter option of these two systems is often rightfully met with consumer backlash and back-peddling on part of the developer—patches and updates correcting the issue to some degree by appeasing the games’ players, which Rosenberg would say is the “things working as intended” model

(2017). Although this depiction of a consumer protest is rather simplified, Rosenberg’s

(2017) emphasis is placed on correcting loot boxes systems that intersubjectively exploit, deceive and manipulate rather than those that simply work with a restricted purpose in mind (e.g. aesthetic character customizations that in no way affect gameplay). However, political fixation on the games industry’s golden egg of a loot box would also suggest that legislative power should rectify how the ESRB treats the contents of loot boxes; namely, disclosing the odds of potential winning at the very least would elaborate on the nebulous denotation of in-game purchases (Jerauld, 2018; Willett, 2018).

The practicality of disclosing specific odds would best be left for the game’s developers and the ESRB rather than their retailers. For example, in the case of the

Entertainment Software Association (ESA) v. then-Illinois Governor Rod Blagojevich

(Entm’t Software Ass’n v. Blagojevich, 2006), the ESA challenged two statutes banning the sale of violent and sexually-explicit games to minors because of the burdens placed on commercial speech. Specifically, games fitting either of these two categories received a two-by two-inch (2”x2”) sticker reading “18.” All retailers responsible for the sale of video games in Illinois were to ensure that such games presented the sticker on its front cover. However, the Seventh Circuit Court of Appeals did not find enough reason or argument that this measure would be necessary given the ESRB rating given to video

10 games (Entm’t Software Ass’n v. Blagojevich, 2006). Additionally, the decision to determine whether a game received the 2”x2” sticker was the retailer’s responsibility, even if games were not the primary goods being sold within said establishment (e.g.

Walmart, Target). This process required a subjective examination of the games’ content by the retailer. Therefore, the impracticality of all state-wide retailers to interact with their games and determine the need to highlight explicit content defeats the purpose of the ESRB. The Seventh Circuit ruled the two statutes unconstitutional.

Unsurprisingly enough, the ESA is a frequent flyer with the U.S. judiciary branch when it comes to rectifying the treatment of games with legislation. Another example comes from a 2009 case in Chicago, wherein the Chicago Transit Authority (CTA) selectively censored advertisements of M- and AO-rated video games on their transportation vehicles. The ESA contested that this ordinance was unconstitutional arguing that the advertisements inhabited a public forum. The CTA had agreed prior to the ordinance that gaming advertisements could be allowed on their trams and subways, but they later amended the ordinance with the ban on M- and AO-rated video games. As a result, the ESA proved successful in its argument that the CTA ordinance could not withstand precise scrutiny because of the unwarranted amendment to the pair’s agreement (Entm’t Software Ass’n v. Chi. Transit Auth., 2010). Therefore, the district court ruled in favor of the ESA, noting that unlike advertisements for tobacco and alcohol, mature video games are protected media under the First Amendment.

It must be reiterated that the self-regulation of games’ ratings by the ESRB was designed in-part to combat political and public pressures of violent and sexually-explicit

11 video games (Willett, 2018; Rosenberg, 2017). To recap, games’ first amendment protections were deemed valid because of their artistic potential and the expressiveness as persuasive, communicative texts (Rosenberg, 2017; Brown v. Entm’t Merchs. Ass’n,

2011). Yet, this protection only holds up rhetorically so long as games have something to say (Bogost, 2015; Willett, 2018). As Bogost (2015) highlights, games are art as much as they are appliances; the inclusion of a predatory loot box system then only facilitates an insidious view of games as devices that enable gambling-like behavior. Thus, in line with

Willett (2018), the ESRB’s response to loot boxes ought to ensure consumers that self- regulation is possible, lest they see federal intervention as preferable. Although he appeals to a rhetorical slippery slope scenario, Willett (2018) reminds the reader that it would not be farfetched to believe that “those in power (would) limit themselves to regulating supposed video game gambling when some instinctively connect mass shooting and video games” (Willett, 2018; italics in original copy).

Likewise, dissenting and nuanced opinions in cases ruling in favor of the ESA may hint at the difficulties with assessing games’ correlations to aggressive behavior.

Unlike other state-wide ordinances imposed on M- and AO-rated video games, a

Minnesota law once imposed a fine of $25 to any minor caught in a transaction to rent or own banned games. As expected from preceding cases involving the prohibition of M- and AO-rated video games involving the ESA, the court ruled in favor of the plaintiffs with a notable opinion among the circuit court judges: the reaction against violent and sexually-explicit video games are, however superficial, reasonably warranted in the case of minors’ psychological wellbeing (Entm’t Software Ass’n v. Swanson, 2008). The

12 burden of proof that rests with the state to provide enough evidence of a causal link between exposure to violent media and psychological problems, according to the circuit judges, may in fact highlight “a refined estrangement from reality, but apply it (they) must” (Entm’t Software Ass’n v. Swanson, 2008). In other words, this case reveals that although games receive First Amendment protections, they are not simultaneously free from the scrutiny of their purpose and function. For example, one criticism implies violent video games serve an artistic, political or social purpose that outweighs simulated violence. An opposing, yet parallel, critique might argue that the judicial system, in defense of video games, demands an unrealistically high standard of evidence by asking for causal links, instead of inferences using strong correlations, between exposure to explicit media and psychological dysfunction or insensitivities. As speculative as these claims are, the point remains that First Amendment protections for video games comes with reasonable scrutiny and criticism. For extensive literature on this topic, look no further than growing critical studies in video games by Shaw (2017).

Nonetheless, consider the possible implications from ESA v. Swanson (2008) and

ESA v. Blagojevich (2006) had the ESRB been unable to sufficiently self-assess the content of the explicit games in question. With emphasis on the former case, what

“estrangement from reality” (Entm’t Software Ass’n v. Swanson, 2008) would rationalize banning—or, to some degree, restricting—violent and sexually-explicit video games?

Causal links between violent displays of behavior and video game consumption have been debunked before (Jenkins, 2004; Ferguson, 2013; Scutti, 2018), yet our nascent understanding of video game media lacks a definitive consensus on peripheral issues such

13 as video game addiction and social learning through games (Copenhaver et al., 2017

World Health Organization, 2018). In a sense, this lack of consensus reels not only political actors and parental watchdogs back into the conversation of games’ more negative effects but also researchers. Subsequently, this results in studies of gaming that often conflict with addiction counterarguments, however popular or nuanced. Some valid studies indicate definitive video game addiction prevalence rates in countries other than the United States (Seok & Decosta, 2012; Lopez-Fernandez et al., 2014). Hence, what are we to do with correlations and other indicators of gaming’s problems within the peripheral view of the rhetorical “violent video games” argument? The ESRB has curtailed these concerns of political actors and parental watchdogs for the time being, yet self-regulating the content of games containing loot boxes may prove to be more challenging to avoid government intervention (Sinclair, 2018).

As expected, a recurring theme within video game’s legal history is the restriction of access to violent and/or otherwise sexualized content within games. For example,

Eugene Brown, a sex-offender detained to the Rushville, Illinois’ Treatment and

Detention Center, sued the institution—joined by 17 fellow inmates—claiming that their restrictive access to movies and video games was unreasonably upheld and violated the

First Amendment (Brown v. Phillips, 2015). However, the Seventh Circuit Court could not determine enough argumentation to overrule the restriction. Although professional testimony from therapists and security of the detention center argued – with appropriate data – that it was “common sense” to restrict sex-offenders’ access to legal sexuality- explicit media, there lacked evidence to suggest that such a ban would reduce recidivism

14 in criminals (Brown v. Phillips, 2015). Appropriately, the court affirmed and vacated the case simultaneously, referring primarily to each detainee’s specified treatment and how best to approach regulation on an individual basis.

Perhaps the most heavily contextualized type of case associated with video game legality is the unauthorized use of likenesses. These cases often refer to a transformative test which determines whether a product is the artist’s own creative expression or comprised mainly of references to celebrity. This test originated from a California

Supreme Court case (Comedy III Productions, Inc. v. Gary Saderup, Inc., 2001), wherein a work containing likenesses to celebrities must transform such expressions to become more than just mere imitation. Five categories constitute the transformative elements of a product:

(1) the use of celebrity as inspiration rather than pure substance; if the celebrity

representation in question is determined to be more inspiring than pure substance,

the work is considered transformative.

(2) the motivation of a consumer to purchase the good based on celebrity

reproduction or creator’s personal expression; if it is reasoned that the consumer

would purchase the product primarily because of the celebrity reproduction, the

product is not considered transformative.

(3) the number of original versus imitative elements in the work; if there are more

original elements than imitative ones, the work is deemed transformative.

(4) the economic value derived from the imitation; and

(5) commercial replication of celebrity.

15 This transformation test resurfaced in the case of the American rock band No

Doubt v. game developer Activision (2011), wherein the band members of No Doubt sued the developer for overextending their likenesses in Band Hero, a game where player could simulate the life of a rock star by playing songs with a guitar controller. Despite having signed off on the use of their likenesses, No Doubt sued claiming their characters could be used as avatars for any song performed in the game. Not only did this fall under the scrutiny of the transformative test, but band members claimed that such manipulation could be considered character defamation (e.g. one article cites the fact that No Doubt’s lead singer Gwen Stanfani could perform Rolling Stone’s “Honky Tonk Woman” with

Mick Jagger’s voice; Gardner, 2012). Therefore, the character depictions of No Doubt’s band members were found to be non-transformative, and the court ruled in favor of avoiding customer confusion (No Doubt v. Activision Publishing, Inc., 2011). In other words, the developer’s work was not found to be their complete, original artistic expression.

In another example of infringing on likenesses, the satirical Grant Theft Auto

(GTA) series is no stranger to evoking the ire of the its detractors. Rockstar Games, Inc. once attracted the attention of strip club owners E.S.S. Entertainment 2000, Inc. (ESS) for loosely parodying their Play Pen business in the virtual world as the Pig Pen (E.S.S.

Entm’t 2000, Inc. v. Rock Star Videos, Inc., 2008). Specifically, ESS believed that players who purchased GTA: San Andreas would be led to believe that ESS endorsed the title, warranting a transformative balancing test to see if public interest and consumer confusion outweighed that of creative freedom. According to this test’s results under The

16 Ninth Circuit Courts, the Pig Pen was not a substantial feature within San Andreas that would confuse the player to believe ESS endorsed the game; it was not reasonable to believe that a player would realistically spend most of their time at the Pig Pen in the virtual world (E.S.S. Entm’t 2000, Inc. v. Rock Star Videos, Inc., 2008). Therefore, the courts ruled in favor of Rock Star Videos, and its virtual parody of a gentleman’s club was protected by the First Amendment.

Perhaps the most notable string of court cases regarding the use of likenesses involve the use of real-world athletes in video games from Electronic Arts (e.g. Madden

NFL, NCAA Football). In January 2015, former NFL athlete Michael Davis, along with three other former players, filed a class-action lawsuit against EA for unauthorized use of their identities in their Madden NFL franchise series (Davis v. Elec. Arts, Inc., 2015).

Despite contextual cues that highlighted each player’s respective in-game character (e.g. height, weight, name), the court found no infringement of the players’ likenesses because

EA’s product presented the characters in a way “related to the main purpose and subject of the game” (Davis v. Elec. Arts, Inc., 2015). Additionally, the court determined that despite some information about the players, EA’s game was not a factual reference source. In other words, the likenesses shared between the former NFL players and their respective virtual characters served the game’s purpose: to simulate an NFL football game with realistic representations of teams from different eras.

Although this study is concerned with the intersection between video games and gambling, imprecise definitions of the term “video game” blend together ideas of home entertainment console systems and arcade cabinet machines when adding “gambling” to a

17 legal search on LexisNexis. In the case of Helton v. Hunt (2003), the plaintiff, owner of video game cabinet machines, sued the county sheriff for unrightfully obtaining and destroying his property due to a prior contraband ordinance against gambling devices. It was ruled that Helton had not been grandfathered into this clause, and the county ordinance constitutionally enabled officers to confiscate the equipment.

However imprecise the arcade cabinets of Helton v. Hunt (2003) suit this study’s definition of video games, the gambling function of these particular machines might foreshadow how future judiciaries look at gaming. In fact, the gambling metaphor has already led to political scrutiny of games containing loot boxes (Kim, 2017), so a reasonable concern for prosecutors would be to identify ways in which a given game may have elements purposed to emulate gambling and its functions. Thus, gambling may soon appear to be included in the growing list of concerns with interactive media, and this addition is more likely because a tangible transaction is being made.

Unlike virtual representations of violence and sex in video games, depictions of gambling can extend beyond a game’s functionality with the user alone. Although this study only mentions the broader virtual economy of skin-trading and gaming marketplaces, it would be naïve to conclude that loot boxes are not a gamer’s concern. As

Sinclair (2017) mentions, self-regulation of loot boxes may prove more difficult than the

ESRB’s successful attempts at rating violent and sexually-explicit material in games.

Essentially, the issue at-hand that Sinclair (2018) describes is threefold: (1) from a governmental perspective, gaming companies can easily rewrite the way loot boxes work from a gameplay outlook, so loot box law over time requires renewed implementation;

18 (2) from a game’s player perspective, loot boxes are not holistically a bad thing dependent on the game itself, considering that there is a fundamental difference between predatory cash-grabs and microtransactions that ethically fuel a game’s business; and (3) from a game’s developer perspective, it would take serious convincing to abandon a highly profitable mechanism—a proven one, moreover—within one’s creation after rationalizing it to be permissible even in the wake of controversy. Despite Willett’s

(2018) outlook that the ESRB has not done enough to quell the concerns of U.S.

Representatives such as Hawaii’s Chris Lee, an “In-Game Purchases” label could placate loot box concerns in the meanwhile. Nonetheless, the current threat of federal regulation is paramount to the ESRB retaining future control over the regulation of games completely, not just limited to loot boxes and gambling.

Framing Effects

According to Entman (1993), framing is the act of structuring parts of a perceived reality to various degrees of prominence in a communicating text. However, the act of framing media is a multi-dimensional process that involves more than configuring the physical structure of a news product. Framing is also a component of agenda setting, which are the parallel conclusions drawn from public perceptions of a given order of prominence assigned by the news (DeFleur, 2010). Framing and agenda setting together yield the relationship between what is perceived as important by the public, the media and those who in positions of power.

For instance, McKernan’s (2013) construction of video game news reports in The

New York Times over the span of 30 years (1980-2010) demonstrates how informational

19 media develops an unfolding narrative of video games and violence, health issues and the dumbing-down of American youth. Namely, the compilation of reports builds a perception of reality affected negatively by video games, despite later stories about the social benefits of games. Part of this stagnant period of similar reports on video games and violent, anti-social behavior come from the tragedies of late-20th century school shootings, such as Columbine, which pair games condemningly with aggressive social learning (McKernan, 2013). During this same period of reports, journalists were keen to reference psychological studies on correlations between video games and aggression, tolerance of violence and childhood obesity, not to mention political figures that have actively campaigned for video game regulation. Nonetheless, these depictions of video games provide the reader with a grim perception of reality that games function as a perverse tool to desensitize children to violence, lethargy and taboo behavior. Yet, these articles also admittedly refer to video game violence as seen particularly in controversial titles such as the Grand Theft Auto or Call of Duty series (McKernan, 2013), the former being the archetype for media and academic study on pervasive games (Garrelts, 2006).

Thus, framing is an interaction between the receiver of said text and their respective culture: the concept of framing effectively portrays how power is emitted through the written word (Entman, 1993). Bias toward the weight of risk and negative effects of video games tends to result in a more cautionary framing of games and the behaviors of their respective players (Copenhaver et al., 2017; Kumpel & Haas, 2016).

Additionally, academic papers published regarding the negative effects of video games tend to receive more scholarly citations than their positive counterparts

20 (Copenhaver et al., 2017). The authors suggest that negativity bias would subsequently lead these negative findings to greater press coverage. Indeed, psychologists have grappled with the argument that such negative findings regarding video game effects have “exceeded the data or ignored conflicting data” (Ferguson, 2013). Arguably, these exaggerations can be explained by controversial games released only in niche markets to small audiences, yet popular titles such as Mortal Kombat or Doom evoke criticism from parental groups and politicians concerned with the association between video games and deviant behavior (Ferguson, 2013).

Subsequently, the reader cannot possibly access all the information available to evaluate a given issue when portrayed by media. The selection of certain information by the media implies that some degree of information is lost or otherwise ignored in the news making process, priming the audience to only the information accessible to them

(Iyengar & Kinder, 1987). In other words, media influence how an audience applies and accesses information when dealing with issue awareness because a single communicating text is not comprehensive of the matter at hand. As given by Copenhaver et al. (2017) and

Kumpel & Haas (2016), negative press for the video game industry only reinstates negative priming effects. In other words, attitudes toward gaming tend to reflect the frames provided by the media. As stated by Perloff (2013) in response to Lang’s (2013) work on media effects research:

If aggression that follows playing violent video games is due to associative network priming, then the influences may be temporally limited, and concern about instigation of aggression should most appropriately focus on the locale where the games are played. (323)

21 In other words, trait aggressiveness should be contextualized to the game being studied, and if subsequent media frames were to follow such research, it is possible that individuals high in trait-aggressiveness would receive a heavier impact from such frames

(Perloff, 2013).

Nonetheless, this study considers a multidimensional approach to how video games are generally framed by the media. The context of “loot boxes” in video games restricts the author’s understanding of framing to gambling, addiction and more problematic references. By failing to recognize how the media frames video games by other categories, this study loses different avenues to approach the present research. For example, two of the commonly cited depictions of contemporary games mention either education (Barab et al., 2009; Foster, 2008; Foster & Shah, 2015) or physical exercise

(Lwin & Malik, 2014). Contrary to uses and gratifications, media frames of these depictions emphasis the social benefit of such functionalities.

Framing has additionally moved into the realm of video games themselves when considering the cinematic structure of certain narrative-style games (Valeriano & Habel,

2016; Chang & Hsieh, 2017). Hence, there is a multi-dimensional layer to framing effects that certain media channels exclude from their reports (i.e. MSM reports of video games v. an in-depth look at a single video game and its programmed reality). This study proposes that MSM narratives about video games are shaped largely by empirical literature on video game effects such as addiction (IGD), relationships with online/offline gambling and other risks posed to youth.

22 Perhaps the closest empirical link between video games and problematic gambling antecedents comes from neurological research about such a connection. Bailey

& West (2016) examined how action gamers’ minds (playing shooters for 10 or more hours per week) react when playing a digital blackjack game. This study focused on event-related brain potentials (ERPs), or the brain’s immediate response to stimuli (in this case, a win, a tie, a loss and a bust). When compared with low gamers (playing games for

2 or less hours per week), action gamers were found to possess higher insensitivity to blackjack outcomes. In other words, the brain had lower changes in ERP between wins, ties, losses and busts then that of low gamers (Bailey & West, 2016). Thus, Bailey and

West’s (2016) findings imply that action gamers are less sensitive to the neurological effects that follow the outcome of a digital gambling game than those who game less frequently.

This study will refer to Kumpel & Haas’ (2016) approach toward framing video games as provided by four media frames:

(1) equivalency frames that present information in a logically equivalent way, yet

using a different line of reasoning;

(2) emphasis/issue frames that highlight multiple relevant considerations on a

single topic;

(3) valence frames that list normative judgements about a topic, i.e. good or

bad/positive or negative; and

(4) value frames that carry the implication that a single position is right/correct

and its opposite is wrong/incorrect.

23 Bailey and West’s (2016) findings emphasize the greater risks of growing insensitivity among heavy gamers. Additionally, it is possible to suggest that such insensitivities would enable more references to gambling and addiction in a media frame on “loot boxes” in video games. Therefore, the author posits the first set of working hypotheses:

H1a: The MSM will report more normative judgment frames about “loot boxes”

than gaming media.

H1b: The MSM will use more gambling references (i.e. gambling frames) than

gaming media.

H1c: The MSM will use more addiction references (i.e. addiction frames) than

gaming media.

Games Media Research: Uses and Gratifications

Indeed, it would be oxymoronic to suggest that gamers are a passive consumer group. Like all other media audiences, gamers seek individual uses and subsequent gratifications for the games they play (Katz et al., 1974). In a general sense, video game uses and gratifications can be defined by affective (e.g. aesthetic satisfaction; game environment, maximizing graphics), cognitive (e.g. comprehension and learning; puzzle solving, advanced mechanics), personal integrative (e.g. improving credibility; leading teams and raids, streaming platforms), social integrative (e.g. improving connections; online friends, teamwork and cooperation) and tension release variables (e.g. escapism; role-playing, world building; Sjoblom & Hamari, 2017). Like movies and television, video games themselves can be designed to fit one or more of these variables by design

24 (i.e. massively multiplayer online role-playing game, MMORPG; real-time strategy,

RTS; first- and third-person shooters, FPS/TPS; horror, etc.). Nonetheless, the variety of uses and gratifications provided by video games has additionally shown the merits of contextualizing who the “gamers” in gaming studies are precisely (Carras et al., 2017;

Braun et al., 2015). For the purposes of this study, the “gamers” trope refers to those who play online video games that feature “loot box” rewards. Thus, because “loot box” rewards provide different gratifications for the player, how does the MSM define these uses and gratifications versus gaming media? Although this study is quantitative by design, the following descriptions of uses and gratifications act as a form of gaming literacy for those familiar and unfamiliar with the medium itself. As mentioned, the uses and gratifications of video games defined by scholars is by no means exhaustive; like

Shaw (2017), this section is aimed to demystify gaming stereotypes and inform the reader of how various types of gamers use various types of games.

Uses and gratifications theory stems from research conducted in the early 1940s comparing women who regularly listened to soap operas to those who did not listen but altogether demonstrated similar lifestyles to each other. The results uncovered major gratification-seeking behaviors of listeners such as emotional release, wishful thinking, receiving advice, exercising sympathy, voyeurism and escapism (DeFleur, 2010). Since then, a cascade of research followed with the purpose of defining basic needs that can be satisfied via the mass media and identifying audience activity. One of the first theoretical studies on uses and gratifications discussed practicality of the theory itself, noting the difficulty of systematizing the full spectrum of an individual’s needs. As a result, the

25 discussion concluded that research pertaining to human needs should focus on the content media delivers and to what degree said content helps satisfy or does not help satisfy personal gratification (Katz et al., 1974).

Advancements made in social media technology have also influenced U&G theory interpretation, insofar as new gratifications for users consist of modality, agency, interactivity and navigability (Sundar & Limperos, 2013). In other words, technology has eased of the acquisition of gratification via the mass media. For the sake of this study, similar research in the field of computers and human behavior reveal exhaustive uses and gratifications for video games. Hence, how this study will categorize human needs and satisfaction as dependent on video game functionality.

An example of uses and gratification within gaming media not only includes games themselves but spectatorship. Sjoblom et al. (2017) measured five high-level needs among 1,097 online survey respondents: (1) cognition, (2) affective motivation, (3) personal integration, (4) social integration and (5) tension release. Social integration was found to be the primary predictor of user subscription to a channel, while the content structure outweighed the topic (Sjoblom et al., 2017). Their findings reveal a new variety of uses and gratifications for video game spectatorship. Consequently, the role of host carries some profession implications, whether the channel used is a video sharing or streaming website (e.g. YouTube, Twitch). The “professional gamer” requires a rhetorical analysis to reveal the reasons behind such a concept. Seo (2015) specifically traveled to local tournaments with interviews prepared to understand what constitutes the role. Most prominently, Seo’s (2015) findings highlighted a unique ethos of gamers

26 willing to separate casual gameplay from work, a reconciliation from other parts of life.

Like professional sports’ teamplay, several characteristics that separate professional gamers from their opposites include mutual respect, fairness, and serious play. Although

Seo’s (2015) findings are not exhaustive, his work is beneficial to market research to gamers that treat gameplay as a career.

Here, the author speculates that the distinction between a “professional gamer” and a behavioral addiction disorder may contain substantial overlap. Yet, such a claim does not accommodate the merits of actual professional gamers or eSports participants

(e.g. scheduling hours to stream, practicing with full awareness to “work,” strategizing with other teammates, etc.). Nor does it include the psychoanalytical research to support such a claim that “professional gamers” exhibit behavioral addiction disorders (i.e. are professional athletes whose training equates to a full-time job exhibiting exercise addiction? Here, the colloquial meaning of addiction is conflated with clinical terminology and prognoses). Mainly, critical theory with respect to video games has pointed out that positive aspects of video games are not as thoroughly scrutinized as their negative effects (Shaw, 2010). In other words, this study does not consider the professional capacity for games as a sport due to reasonable concerns that such an approach asks more questions than provides answers (e.g. what other research is there on

“professional gamers” distinguishing themselves from other gamers? How do

“professional gamers” define their passion for video games? How do “professional gamers” relate to their audiences?). However, such a conclusion is speculative given the growing popularity of esports and video game streaming (Funk et al., 2017; Hallman &

27 Giel, 2017). Instead, future studies may consider adding a framing variable that properly defines “legitimacy” of video games outside their comfort zone media channel and audience.

Scholars have subsequently recognized the economic growth of the video game industry over the last several decades, despite negative press (Funk et al., 2017).

Specifically, the streaming platform Twitch and accompanying eSports outgrowth have warranted communication and economic study. Funk et al. (2017) put forth a scholarly brochure outlining the eSports environment, in particular: (1) debates regarding eSports’ status as an actual sport, (2) how individuals can view eSports as its own competitive platform, (3) what differences there are for implications of eSports in industry study of video games versus academia; and (4) the overall commercialization of the platform.

Hallman & Giel (2017) provide a supplemental argument in favor of eSports as an official sport, offering positive implications for communication of the digital media through credible platforms. Citing similar indications of eSports’ growth, Hallman &

Giel (2017) explain that spectator motivations can contribute heavily to future expansion.

Nonetheless, video game spectatorship continues to provide new findings on various uses and gratifications of the media itself. Sjoblom & Hamari (2017) conducted a direct follow-up study to Sjoblom et al. (2017), adding consumer motivation theory to the approach on Twitch spectatorship. Their findings highlighted a positive correlation between information seeking behavior and number of streams watched. Sjoblom &

Hamari (2017) subsequently found added context for audience-host interaction and tension release as an impactful component. Likewise, studies regarding Twitch have

28 largely been concerned with audience-host effects and semiotics (Jacques et al., 2015; Hu et al., 2017). Hu et al. (2017) approached two livestreams in China with respect to social identity theory and more than nine working hypotheses. In short, Hu et al.’s (2017) findings highlight audience identity’s congruency with that of their respective “hosts.”

They conclude with three main points: (1) the need for social-psychological research that does not assume social identity to be “forceful behavior guidance” (Hu et al., 2017), (2) the salience of para-social interactions between audience and host; and (3) broader effects on audience interaction. Thus, Twitch channels themselves present ample opportunity to demonstrate how hosts become the agenda setters of their own creative works beyond streaming.

In terms of semiotic literature on Twitch, Reckenwald (2017) provides a multi- column transcript scheme of a gameplay livestream to describe the structure of discourse between a host and their audience. One noteworthy observation of Reckenwald’s (2017) study is the emphasis of “pivoting” as a novel communication behavior because the host is primarily eliciting gameplay in addition to conversation. Thus, Twitch users provide highly contextualized pockets of discourse for analysis because the visual cues of game events deter the actors from conversation alone. Here, the author considers the social network of streaming platforms, such as Twitch, as a viable system of video game uses and gratifications; specifically, while the host provides the basis for entertainment, the respective audience engages in a para-social relationship with their host via chat log. This duality is further emphasized in the accompanying codebook to account for any references made to Twitch by the media.

29 Critical theory on video games also reveals more functions games serve to their players. Shaw (2014) highlights three affordances digital media offers its users, equivalent to Hall’s (1980) encoding and decoding for television: (1) perceptible, or factors that are visible to the player; (2) hidden, or factors that are intentionally invisible to the player and can be exploited (e.g. cheat codes programmed into the game); and (3) false, or factors invisible to the player that can be exploited but are unintentional to the game’s functionality (e.g. glitches that provide the player advantage). As games offer mods and cheats, these categories often become heavily contextualized due to the uncertainty of effects and consumer negotiations. All in all, Shaw’s (2017) argument leads the reader to see that computer algorithm, although literally code, is a fluid input provided by human imagination. At the same time, this input can be manipulated – whether intentionally or not – by the end user, resulting in a variety of consequences and/or benefits.

Other less apparent uses and gratifications for games include cognitive evaluation.

Because uses and gratifications theory disputes the passive audience assumption provided by the former hypodermic needle effect (Shaw, 1979), audience response to violent content comes with more reflection and evaluation than mimicry. In a study of violence not only portrayed by video game media alone, Tamborini et al. (2013) discuss how audiences assess representations of violence in film entertainment according to how graphic the content was, how realistic it was and if it was justifiable in their opinion. For the two media together, depictions of violence were determined more by audience justification than the other two variables, suggesting that violent media is not altogether

30 crude for its own sake. In other words, audiences are active in their evaluation of whether violence in a particular media is justifiable (e.g. punishing an evil villain or social taboo).

Nonetheless, the body of literature dedicated to justifying the popularity of video game media serves a secondary objective for the presence of gaming news outlets.

Regarded also as “enthusiast journalism” with more of an emphasis on public relations and entertainment (Foxman & Nieborg, 2016), gaming news outlets are essential community pieces of gaming culture (e.g. review channels, streamers, “Let’s Plays,” esports commentary). While larger channels of gaming infotainment have become popularized over the years (e.g. IGN, Rooster Teeth, PC Gamer), the blogosphere and internet forums sustain a larger part of the gaming community. Foxman & Nieborg

(2016) additionally highlight the ambivalent attitudes gaming journalists hold about their position: being able to make a profession out of gaming appeared to convey a contradicting sense of legitimacy.

By no means does this imply that gaming media professionals take their jobs less seriously as their mainstream media counterparts; take, for example, more generalized gaming media outlets such as IGN, Polygon and Kotaku which cater to a wider array of gaming genres and feature reporting. Each of these websites boasts a gender-diverse audience, with IGN reaching more than 154 million viewers on their digital network (e.g.

YouTube, Podcasts; IGN, 2018). Specific to physical publications, another recognizable name in among mainstream gaming media is Game Informer, owned by the GameStop corporation. As one of the longstanding publications of gaming media aside from

Nintendo Life, Game Informer’s subscription base is predominately young adult and

31 male, estimated at 7.6 million individuals (Brewer, 2018). Despite their differences, each of these gaming media outlets exhibit a similar creative expression in which they conduct gaming reviews, trade show reports and general infotainment; video game media professionals allow their own personalities shine through their work while combatting influence from game developers, marketers and advertisers (Foxman & Nieborg, 2016).

While this study does not explore to great length the hierarchies and interconnectedness of gaming culture as a whole, it does recognize that despite the lack of journalistic traits, gaming news outlets are no less critical of video game representations than the MSM. In other words, the author anticipates that while gaming news outlets will have reports on loot boxes congruent with other channels of information

(i.e. value frames depicting loot boxes as detrimental), the gaming press will continue to defend video games on a larger scale (i.e. altogether, games are not harmful as given by references to how people use games to fulfill wants and needs). Thus, the second hypotheses are stated:

H2a: Gaming media provides more commentary (summation of frames and

critical variables) on loot boxes than the MSM.

H2b: Fact-based reporting provides less commentary on “loot boxes” than

alternative forms of writing (e.g. Op-Ed, Review, Other), regardless of media

outlet.

Gaming Pathology

However antithetical to this study the present section appears, pathological gaming is nonetheless an existing criterion that justifies the exploration of this media

32 research. As mentioned, this study does not seek to discredit the psychological findings or make superfluous claims that a lack of consensus on IGD disavows its investigation.

Indeed, the scientific method begins with a question and results quite frequently with more questions than concrete answers, so the author recognizes the merit of empirical literature that supports and combats his current position. Therefore, this section is meant to represent the empirical literature that supports why the media leans more negatively toward video games and their subsequent effects.

While the similarities and differences between each media channel’s respective agenda is pertinent to this research question, the author would like to emphasize the qualitative expectations of whatever answer will be provided. As such, this section regarding pathological gaming will suffice for the future implications regarding negative depictions of video games.

The superficial linkage between video games and violent behavior has been scrutinized by many scholars in both the communication and psychological fields, yet the lack of consensus between research findings on this matter, as well as IGD, only add speculation to an otherwise frenzied discourse on video game effects (Copenhaver et al.,

2017 World Health Organization, 2018). Hence, the American Psychological Association

(APA) has taken steps in the past decade to address this with more precision, adding IGD in their “Conditions for Further Study” section of the fifth edition Diagnostic and

Statistical Manual of Mental Disorders (DSM-V):

“(IGD) is a pattern of excessive and prolonged Internet gaming that results in a cluster of cognitive and behavioral symptoms, including progressive loss of control

33 over gaming, tolerance, and withdrawal symptoms, analogous to the symptoms of substance use disorders” (APA, 2013).

Prior to IGD, psychological research on non-substance addiction focused on the effects of excessive internet use (Internet Addiction Disorder, IAD; Young, 1996) and online/offline gambling. Indeed, this research suggests, and often finds, significant relationships between combined IGD, IAD and gambling disorders (Gunuc, 2015; Macey

& Hamari, 2017). Thus, the claims propelled by the MSM regarding the negative effects of video games is not entirely unfounded. Instead, this study suggests that such claims are exaggerated for the purposes of framing and/or priming an issue (e.g. school shootings, deviant behavior in children, video game addiction).

As hinted, antecedents to IGD – as opposed to IAD or gambling – are usually found to contain more than one criterion (Toker & Baturay, 2016) for prognosis. In other words, IGD by itself is unfit to warrant a diagnosis; comorbidity with cognitive impairments, such as obsessive-compulsive disorder and attention deficit hyperactivity disorder, is frequently found to be the strongest factor in sustaining IGD symptoms

(APA, 2013; Toker & Baturay, 2016). This has subsequently led researchers to conduct meta-analyses on what constitutes an appropriate IGD scale (Cronbach > 0.70) for contextualized research.

According to the DSM-V from the APA, IGD is generally categorized as a behavioral, or non-substance (e.g. omitting drugs and alcohol), addiction to video games

(APA, 2013). However, the DSM-V also recommends further examination of IGD before recognizing it as a formal, clinical condition due to vague definitions of withdrawal

34 effects from abstained video game play and the denotation of addiction itself. In other words, how are behavioral addictions defined such that other behavioral addictions are independent from one another (e.g. addiction to tanning beds, exercise, the internet, sexual activity; Bargeron & Holmes, 2016)? Simultaneously, how do researchers make clear distinctions between addicts and healthy individuals (Kardefelt-Winther et al.,

2016)?

No matter what degree of distinction is made between an addict and a healthy consumer of video games, media reports on mental disorders tend to act as both information and misinformation for the public (Slopen et al., 2007). Without contextualizing events regarding IGD and online/offline gambling, we risk assuming a hypodermic needle approach to the precursors – or antecedents – of these disorders (i.e. one that implies that video games, by themselves, are enough to warrant a diagnosis).

Previous findings on representations of mental disorders in the media also show that corrective articles (e.g. stories that educate and clarify the reader about the mental disorder in question) act as a buffer to misconceptions (Thornton & Wahl, 1996; “Impact of a Newspaper Article on Attitudes Toward Mental Illness”).

Likewise, it is important to first understand where terminology for IGD originates in the DSM-V. Broader research on addiction to the internet and analyses of anecdotal cases from Young (1998) prompted the construction of an eight-item questionnaire based on pathological gambling criteria from the DSM-IV (APA, 1994). Of these items, three common internet addiction variables and their operational definitions used to characterize internet addiction include preoccupation (thoughts of internet use whenever the user is

35 unable to access the web), compulsion (to use the internet despite present tasks or responsibilities) and withdrawal (drawback effects on the user from extended periods of no internet usage; Young, 1998). Today, IGD is characterized as a nine-item list wherein the patient in question must adhere to five or more items to be considered problematic:

(1) “Preoccupation with Internet games. (The individual thinks about previous

gaming activity or anticipates playing the next game; Internet gaming becomes the

dominant activity in daily life).

(2) Withdrawal symptoms when Internet gaming is taken away. (These symptoms are

typically described as irritability, anxiety, or sadness, but there no physical signs of

pharmacological withdrawal.)

(3) Tolerance – the need to spend increasing amounts of time engaged in Internet

games.

(4) Unsuccessful attempts to control the participation in Internet games.

(5) Loss of interest in other hobbies and entertainment as a result of, and with the

exception of, Internet games.

(6) Continued excessive use of Internet games despite knowledge of psychosocial

problems.

(7) Has deceived family members, therapists, or others regarding the amount of

Internet gaming.

(8) Use of Internet games to escape or relieve a negative mood (e.g. feelings of

helplessness, guilt, anxiety).

36 (9) Has jeopardized or lost a significant relationship, job, or educational or career

opportunity because of participation in Internet games” (APA, 2013).

The DSM-V additionally notes that this disorder is completely distinct from

Internet gambling, a condition elaborated within the “gambling” section of the manual.

Thus, no criteria listed here correlates IGD with online gambling, but as mentioned, media representations of video games tend to highlight their risks and negative effects more frequently (Copenhaver et al., 2017). This study seeks to contextualize this terminology as reported by MSM outlets and gaming media.

Certain studies undertake a synergetic approach toward understanding symptoms of IGD with gamers themselves. Carras et al. (2017) suggest that gamers and the DSM-V both strongly agree on three criteria that indicate IGD: (1) neglect of responsibility, (2) a dissociative attitude toward reality and (3) withdrawal effects, such as stress, when not playing (Carras et al., 2017). While this qualitative approach builds a healthy rapport between gamers and psychology researchers, Carras et al. (2017) acknowledge the lack in withdrawal specificity and the generalization of gaming culture. Additionally, King et al.

(2016) conducted a qualitative study to understand the precise withdrawal effects and coping habits of heavy gamers that abstained from gaming. Using a thematic approach with an IGD-lite survey, respondents (20 male, four female) provided both positive and negative aspects of substituting games for another activity (King et al., 2016). Likewise, both internal and external cues to play games were proportional to individual levels of gamers’ cravings (King et al., 2016). Three consistent criteria for withdrawal effects were

(1) sadness, (2) irritability and (3) anxiety, though boredom and a lack of mental

37 stimulation were additionally cited items. Thus, King et al. (2016) provide extra criteria to strengthen the empirical IGD list.

Also, are players that identify with their characters/avatars more susceptible to

IGD? Sioni et al. (2017) conducted interviews over forum message boards (n = 394, 50 percent female) with scales rated with alphas of 0.90 to answer this question. They demonstrated associations between social phobia, avatar identification and IGD symptoms, leading to the speculation that player-avatar identification could be a valid indicator for IGD. However, these conclusions do not explain how individuals identify with their avatars nor how self-esteem relates to vicarious player-avatar identification.

Additionally, the sincerity of responses and frequent hostility met on these forums mar the full implications of their findings.

As this paper is concerned, there is cause for speculation of a relationship between online gaming and gambling. However, research would consider this claim to be a severe leap in logic, because video games, by themselves, have no associative properties with a rise in gambling potential (Macey & Hamari, 2017). Motivated by the lucrative games- as-a-service model (free-to-play games), Macey & Hamari (2017) fit a least partial squares model to 613 individuals eSports consumption and gambling potential. As their above-stated conclusion indicated, video game addiction was inversely related to online, offline and problematic gambling behavior. Thus, an emphasis has been put on the convergence of media with digital culture when considering the structural relationship between games and online gambling.

38 The inquiry into gaming’s correlation with gambling also concerns the role of social media. Jacques et al. (2015) categorized 100 Facebook app games into three categories: (1) standard gambling simulations, (2) non-standard gambling simulations and

(3) gambling references. While 13 games of the initial sample were omitted as casino simulators, 54 percent of the remaining sample significantly held gambling content

(Jacques et al., 2015). Their findings highlight the need for semiotics research of

Facebook games containing references to gambling and how to indicate individuals at- risk. More importantly, how frequently – if, at all – do virtual experiences transfer neatly into their equivalencies in the real world?

King et al. (2015) observed how 1,287 adolescent Australians performed simulated gambling tasks and correlated their observations with demographic information. While their study agreed with most findings here (i.e. unrelated gambling behavior is a predictor, not a symptom, of pathology), it is unclear which causal relationship is assumed here (King et al., 2013). Simulated instances of gambling are also structurally similar to those in the real world, but there lacks study on what exposure to these simulations do to participants over time. King & Delfabbro (2015) examined academic papers on this topic between 2000 and 2015 and found significant gaps in knowledge between researchers’ findings about social media/simulated gambling and actual pathological gambling.

In terms of gender differences in online gambling addiction studies, Edgren et al.

(2015) observed 3,555 individuals’ (46.2 percent female) responses to a Finnish gambling survey. Their findings indicated that men were more likely to show signs of

39 gambling behavior online than their female counterparts (Edgren et al., 2015). Unlike previous findings cited however, gamblers were more likely to be gamers in this context, providing depth to IAD’s relationship with gambling and IGD combined.

On the other hand, Adiele & Olatokun (2013) are two of the strongest proponents who contest the idea that the internet cannot be defined so crudely as a substance, but rather as a medium of information. Adiele & Olatokun (2013) attribute Katz et al.’s

(2016) uses and gratification theory to this definition, because for internet addictions to be legitimized, two criteria would need to be fulfilled: (1) individuals would need motivation for using the internet, and (2) addictive behaviors online would need to be unrelated to those occurring offline. By studying 1,022 adolescent respondents to internet addiction questionnaires, Adiele & Olatokun (2013) found an IAD prevalence rate of 3.3 percent. Additionally, adolescent use of the internet was primarily motivated by reasons external to the internet itself (i.e. motivations from the real world, such as offline behavioral addictions). Hence, pathological internet users are more closely defined as

“addicts on the internet” rather than “addicted to the internet,” because the offline behavioral addictions reported were motivational factors instead of causal ones (Adiele &

Olatokun, 2013).

Thus, developing theories on diagnosing internet addiction symptoms have lead researchers to investigate other pathological issues within its proximity. In other words, while behavioral addictions must be characteristically distinct from one another, IAD on its own cannot warrant a pathological issue (Adiele & Olatokun, 2013; Gunuc, 2017).

Often times, IAD symptoms overlap with excessive gaming and pathological gambling,

40 creating motivations for researchers to accurately characterize the antecedents to each issue (Toker & Baturay, 2016). For example, Gunuc (2017) observed a strong correlation between IGD and IAD and wanted to know if tolerance of symptoms affected each disorder. Gunuc’s (2017) study observed 200 Turkish students aged 15-18 and their responses to addiction questionnaires. Subsequently, any symptoms of addictive behavior alone did not warrant a prognosis, but the prevention of tolerance to IGD and IAD symptoms worsened withdrawal effects in participants. Hence, IGD must be considered in tandem primarily with IAD. Similarly, Canale et al. (2015) fielded 14,778 adolescent students from the European School Survey Project on Alcohol & Other Drugs to characterize the linkage between IAD and gambling pathology. They determined that problematic gamblers among adolescents were much more likely to use online devices for gambling than their at-risk counterparts. However, any possible association of online gambling with IGD in Canale et al.’s (2015) sample is marred by a larger emphasis on context and family structure of the demographics studied.

Burgeron & Holmes (2016) also sought to hone in on a conceptualization and assessment of IGD. As mentioned, IGD would fit into the category of behavioral addictions, yet how does it differentiate from other behavioral addictions? Burgeron &

Holmes (2016) examined the clinical complications that accompany DSM-V criteria for

IGD, notably its relation to social phobia and low perceptions in well-being. By testing

257 self-identified gamers, Bargeron & Holmes (2016) found moderate to strong significance in all DSM-V criteria used, indicating an association between IGD and social psychological impairments. However, while Bargeron & Holmes’ (2016) findings

41 suggest pathological gamers exhibit poor emotional regulation and coping skills, their findings could not distinguish between the pathological gamer and the healthy gaming enthusiast with high levels of engagement.

Subsequently, a recurring theme in IGD studies since Bargeron & Holmes would be the fine distinction between the pathological gamer and the average gaming enthusiast.

Seok & Decosta (2012) used both a mono- and polythetic classification system on 1,332

South Korean students to assess what determines pathological gaming as opposed to high levels of engagement with video games. The group found rates between 1.7 and 25.5 percent for addiction perception with an overall 2.7 percent addiction rate, leading the researchers to see that IGD rates in South Korea were lower than expected (Seok &

Decosta, 2012). Their findings provide implications for cultural differences in video game play and perceptions of IGD. Likewise, classifying the transition from avid gamer to pathological addict requires an additional thorough examination of IGD criteria.

Lopez-Fernandez et al. (2014) observed two pools of European adolescents (1132 participants from Spain, 1224 from the UK) with respect to a problematic video game scale. A 7.7 and 14.6 prevalence percentage rate of pathological gaming was found for each sample respectively with the difference in rates attributed to cultural differences in withdrawal effects (i.e. respondents from Spain reported “playing through” IGD symptoms, while UK respondents reported feelings of “regret” when abstaining from video gaming). Because of their findings, Lopez-Fernandez et al. (2014) determined an abridged 5-item verification for pathological gaming according to these cultural differences.

42 Even though researchers have acknowledged the salience of pathology in gaming, a clear definition of “pathological gamer” by itself has remained elusive (Buono et al.,

2015). Namely, how can clinics assess the role/function video games play in “video game addiction?” Buono et al. (2015) sampled six psychiatrists and 467 peer-reviewed articles for their motivations to properly define IGD terminology. Their outcome was a reasonably factorable 21 to 24-item questionnaire designed to assess video game function in IGD cases (Buono et al., 2015). Thus, closer examinations of IGD literature has produced refined categories for functional assessments of video games with respect to behavioral addiction. At the same time, however, the variety in IGD questionnaire criteria leaves much to be desired for a precise characterization of IGD by itself.

Perhaps the current media coverage of loot boxes within video games may spark a new approach to how players interact with games. Particularly, there is scarce research on the economy within games (e.g. trading, selling and buying virtual items for in-game/host platform currency), particularly on the digital game store platform known as Steam. It is important to note that this virtual currency (i.e. the “Steam wallet”) cannot be transferred to real-world currency, but it can be used to purchase additional games, downloadable content and purchases that would alternatively require credit card input. Thus, certain games on Steam exhibit profit-making systems wherein the player can add to their

“Steam wallet” without having to provide a line of credit more than once. However, these systems create a new swarm of speculations: how do gamer exploit this system to

“maximize profit?” What hierarchies exist within this system (e.g. consumer level, corporate level)? How do these systems economically benefit Steam itself? Nonetheless,

43 this line of research is considered for future study but not entirely for the case at hand, because not all video games with “loot box” rewards contain the same systems as the

Steam platform for users to exploit monetary gains.

Supplementary research on IGD also helps characterize pathological personality patterns with respect to gaming addiction. Testing 1,004 adolescent males on personality inventory tests, Spekman et al. (2013) discovered a positive correlation between problematic gamers and both direct and indirect measures of addictive personality.

Although some gamers exhibit personality patterns reminiscent of addiction, game exposure was not correlated with any measure of addictive personality.

Video game effects frequently reveal positive cognitive ability in users (Forrest et al., 2015). This has led researchers to investigate said effects’ correlation with IGD. In

Forrest et al.’s (2015) Australia-based study, 657 individuals were assessed on 22 items pertaining to games’ cognition such as perfectionism, cognitive salience, regret and behavioral salience. While their results showed effects for cognitive salience and regret,

Forrest et al. (2015) discuss the limitations of not taking motivations for playing games into account. Subsequently, needs filled via video games can, at times, substitute for needs satisfied in real life (Allen & Anderson, 2018). Motivated by findings that cite 8 percent to 9 percent of Singapore youth as pathological gamers (Gentile, 2009; Gentile et al., 2011, as cited by Allen & Anderson, 2018), Allen & Anderson (2018) developed a study of MMO players based on self-determination theory, and subsequent hypotheses of need-density and real-world superiority. Using a 9-item DSM-V criterium for IGD, their findings revealed strong support for pathological gaming behavior whenever there was a

44 large discrepancy in need satisfaction between real life and video games (Allen &

Anderson, 2018), making the dichotomy a reliable predictor of IGD.

With the concurrent research regarding digital addiction and its intersection with online gambling, the author anticipates some congruencies between the MSM and GM reports on “loot boxes.” However, the limitations of a content analysis restrict such an inquiry to a second broad research question, rather than a directive hypothesis:

RQ2: What comparisons can be made between MSM and GM reports on “loot

boxes?”

Likewise, subsequent research hopes to closer examine the treatment of video games in MSM with a more qualitative approach. Nonetheless, an inevitable attempt will be made at depicting any significant discrepancies between MSM and GM reports. Given the implications of the regulation of games with loot boxes, the unavoidable concern of federal intervention with the games industry highlights the importance of some qualitative prescription for future research.

In conclusion of this section, the author hopes to convey the impact of growing literature to contextualize the various problems gaming provides. Although at first glance this section appears to denote the lack of consensus between researchers, the opposite point has already been made: the relationships between IGD, IAD and online/offline gambling are nonetheless substantial and detrimental, considering the comorbidity these disorders share with pre-existing issues (e.g. ADHD, OCD). Thus, scholars are not aloof to the overlap between these three disorders and what they imply.

45 Methods

The working hypotheses were tested by analyzing the content of 274 articles pertaining to loot boxes in video games. This sample size was determined using a 95±5 percent level of confidence within a 1,200-article population. A total of 17 duplicate articles were omitted from the initial testing. The articles for this study have been collected using a Boolean search from the LexisNexis Academic online database.

Specifically, no more than two search terms were combined at a time, and the ensuing search always contained either “loot boxes” or “gambling” as a part of their criteria (e.g.

“loot boxes” AND “video games;” OR “gambling” AND “video games;” NOT “video games” AND “addiction” because this lattermost criterion excludes a reference to gambling or loot boxes). For specifics on the search criteria, please refer to the concluding section of the codebook below. This is to ensure that the sample provided will contain content pertinent to this study about “loot boxes” explicitly and gambling references implicitly within video games. As an additional measure, only articles after the release of Blizzard’s Overwatch will be considered (after May 24, 2016) because the game itself recognizes its awards specifically and, more importantly, explicitly as “loot box” rewards. Since then, a cascade of references to “loot boxes” has been encoded in media to signify these randomized rewards.

A total of 40 composite weeks was constructed from the larger population of

1,200 articles. Two student coders examined the sample based on the attached codebook.

At the completion of their work, each coder was compensated with a $20 Steam gift card for their efforts. Intercoder reliability was tested using Cohen’s Kappa in addition to

46 running a Cronbach alpha for each variable defined below. For all variables used, an alpha of 0.72 was achieved. Linear regression was performed for each media channel to predict the average number of framing and critical variables for any given article.

The procedures for this study fall within objective practice because of the recent codification of the term “loot box.” Since Overwatch’s inception, the term “loot boxes” have strung together the growing presence of randomized rewards within video games regardless of whether such rewards are explicitly categorized as “loot boxes.” Hence, the term has become more easily applicable and accessible to the outsider looking into video game culture. Blizzard’s game has indirectly provided the MSM and gaming media a shared terminology in how the media describes systems of rewards within games today.

Therefore, the search process this study follows additionally abides by the rhetoric of systems found within games themselves to ensure that the sample studied pertains to the matter of discussion (i.e. “loot boxes”) and not to a peripheral issue alone (i.e. “video game addiction”).

47 Results

After measuring intercoder reliability, eight variables were generated to characterize 274 articles containing the term “loot box(es).” Among these eight variables, there are three specific subcategories (1) descriptive, as in Media Outlet, Word Count and

Story Type; (2) framing, as in Normative Judgment, Gambling and Addiction

References; and (3) critical, as in Personal Testimonies and Political Intervention.

Likewise, these subcategories are designed to conceptualize (1) the tangible characteristics of each article, (2) the media’s framing process of each story and (3) what interactions third parties make with the issue of “loot boxes.”

Additionally, a t-test was conducted to examine any potential differences between the MSM (nMSM = 199) and GM subsamples (nGM = 75) used within the study. Table 1 outlines the comparison of means between each subsample, including that of the total sample and intercoder reliability ratings given as Cohen’s Kappa, κ. Four scaling variables are also included within Table 1, as well as their respective Cronbach alphas, α.

These α-scores indicate the level of cohesion between the variables used to create such a scale. For instance, framing is comprised of Normative Judgment, Gambling and

Addiction References with an α = 0.66. Generally, an acceptable level of cohesion is α >

0.70, but the reduced sample size along with the number of variables limiting this study’s finding warrants flexibility. Thus, the following coefficient (C) scale shall determine the impact of various categories used within this study:

 C = 0.00 to 0.20, … 0.00 to -0.20; Very Weak

 … 0.20 to 0.40, … -0.20 to -0.40; Weak

48  … 0.40 to 0.60, … -0.40 to -0.60; Moderate

 … 0.60 to 0.80, … -0.60 to -0.80; Strong

 … 0.80 to 1.00, … -0.80 to -1.00; Very Strong (Wimmer & Dominick, 2013)

As such, any inferences regarding the results of this study will consult this level of measurement when assessing suggestions for future study.

Table 1

Descriptive Statistics of Full Sample and Subsamples.

MSM GM Mean, Cohen’s Variable Mean, 푥̅ Mean, t-test, t 푥̅퐺푀 Kappa, κ 푥̅푀푆푀

Media Outlet 0.27 n/a n/a 0.62 n/a

Word Count 829.16 726.44 867.88 n/a 1.82*

Story Type 1.24 1.31 1.22 0.58 -0.76**

Personal 0.19 0.29 0.15 0.62 -2.43** Testimonies

Political 0.32 0.27 0.34 0.74 1.14* Intervention

49 Table 1 continued

Normative 0.46 0.59 0.41 0.62 -2.61 Judgment

Gambling 0.37 0.31 0.39 0.58 1.34** References

Addiction 0.13 0.12 0.14 1.00 0.34 References

Frames1 1.72 1.65 1.75 n/a 0.70

Frames AND 2.95 2.90 2.97 n/a 0.26** Politics2

Frames AND 2.62 2.64 2.6 n/a -0.13 Testimonies3

Commentary4 1.47 1.57 1.43 n/a -0.70

Sample size, n = 274 (nGM = 75; nMSM = 199) Cronbach alpha’s for scaling: α1 = 0.66; α2 = 0.77; α3 = 0.65; α4 = 0.74. * Significant at the 0.05 level (two-tailed). ** Significant at the 0.01 level (two-tailed).

Moderate (κ = 0.40 to 0.60) to strong (κ = 0.60 to 0.80) levels of consistency between coders were observed for the eight finalized variables. A total of three beta tests were conducted using 10 control articles each. The omission of variables resulting in weak (0.20 to 0.40) to very weak (0.00 to 0.20) κ-scores is elaborated in the discussion section, but the discrepancies between scores deemed acceptable for this study must be noted. Ideally, all acceptable κ-scores would be of a strong to very strong (0.80 to 1.00)

50 level of consistency, so any inferences resulting from this study might be verified using a larger sample size.

Word Count is the only ratio-level variable with substantial standard deviation (σ

= 708.17) and variance (σ2 = 501507.64). This large discrepancy is due to the range

(6062 words) between the minimum (40 words) and maximum (6102 words) values of this variable. Story Type has four separate values assigned as “News Writing” (1), “Op-

Ed” (2), “Review” (3) and “Other” (0).” The lattermost category “Other” in this variable characterizes any number of story types that originate from press releases of a company, trade press information, announcements and/or sources unlisted by the codebook. The remaining four variables are on the nominal level, and they are codified by the numbers zero (0) and one (1).

A reliability test yielded a moderately strong range of Cronbach alphas for each scale included in Table 1 (α1 = 0.66, α2 = 0.77, α3 = 0.65, α4 = 0.74). This test was conducted to verify the similarities between individual framing variables and the two critical variables. For reference, as a score approaches 1.00, the closer the criteria is between each variable within the created scale. For reference, “Frames” contains the three framing variables “Normative Judgment,” “Gambling Reference” and “Addiction

Reference.” Although the “Political Intervention” and “Personal Testimonies” variables appear to positively impact the scaling of “Frames,” the criteria needed to code these two variables were non-exhaustive in comparison to the three “Framing” variables. Thus, these Boolean-formatted scales are represented as follows: (1) Frames with respect to

Political Intervention; (2) Frames with respect to Personal Testimonies and (3)

51 Commentary Density, a variable symbolizing the summation of “Frames,” “Political

Intervention” and “Personal Testimonies.”

Figure 1. Count of Story Type by Media Outlet.

As demonstrated additionally by Figure 1, the Gaming Media articles represent

27.37 percent of the total sample. Between these story types, News Writing (i.e. facts- based reporting with little to no personal opinion by the author) outweigh its three counterparts with 50.67 percent (GM) and 73.87 percent (MSM), respectively. While this discrepancy is further explored in the Results section, GM’s reduced focus on News

Writing style demonstrates how GM articles serve a different purpose than MSM articles.

52 Table 2

Cross-tabulation of Framing and Critical Variables by Media Outlet.

Media Outlet

Framing AND Critical MSM GM Total Variables

Yes 30 22 52 Test. (χ2 = 7.20) No 169 53 222 Total 199 75 274

Yes 67 20 87 Poli. (χ2 = 1.23) No 132 55 187 Total 199 75 274

Yes 82 44 126 Norm. (χ2 = No 117 31 148 6.69) Total 199 75 274

Yes 78 23 101 Gamb. (χ2 = No 121 52 173 1.70) Total 199 75 274

Yes 27 9 36 Addc. (χ2 = No 172 66 238 0.12) Total 199 75 274

T-testing also reveals several significant differences in means between the MSM and GM subsamples. At the 0.01-level, these differences occur between the variables

53 Story Type, Personal Testimonies, Gambling References and Framing combined with

Political Intervention; Word Count and Political Intervention are significant on the 0.05- level. Using Table 2, a crosstabulation of the framing and critical variables quantifies whether each variable is observed within each subsample. Only two of these variables,

Normative Judgment and Addiction References, are not significantly different between the MSM and GM subsamples.

Correlations for all articles (Table 2a) was subsequently divided into two subcategories as shown by Tables 2b and 2c to further visualize the differences between the GM (nGM = 75) and the MSM (nMSM = 199).

Table 3a

Correlations of full sample.

Medi Wor Stor Norm Gamb Addc Frame Cmmt All Test. Poli. a d y . . . s . 0.16 Media 1 -0.09 0.05 -0.07 0.16* -0.08 -0.02 -0.06 0.04 * 0.18 Word 1 0.03 0.07 0.09 0.03 0.02 0.01 0.07 * - - - Story 1 -0.05 0.19 -0.07 -0.11 -0.16 0.16* 0.17* * 0.17 Test. 1 0.32* 0.19* 0.20* 0.12 -- * Poli. 1 0.41* 0.70* 0.41* 0.52** -- Norm. 1 0.43* 0.29* -- -- Gamb. 1 0.49* -- -- Addc. 1 -- -- Frame 1 -- s

54 Table 3a continued

Cmmt 1 . n = 274 * Significant at the 0.01 level (two-tailed). -- Nullified due to grouping variable (i.e. Frames = Norm. + Gamb. + Addc.; Cmmt. = Frames + Poli. + Test.)

Among the number of significant correlations (p < 0.01) for all articles, only one is a moderately strong, positive correlation between “Gambling References” and

“Political Intervention” (ρ = 0.70). Likewise, positive associations indicate that as one variable’s value increases, the other paired value increases alongside it. Significant correlations suggest that this association—whether positive or negative—does not occur by chance. For additional reference, the impact of a correlation is assessed according to the following scale:

 ρ = 0.00 to 0.20, … 0.00 to -0.20; Very Weak

 … 0.20 to 0.40, … -0.20 to -0.40; Weak

 … 0.40 to 0.60, … -0.40 to -0.60; Moderate

 … 0.60 to 0.80, … -0.60 to -0.80; Strong

 … 0.80 to 1.00, … -0.80 to -1.00; Very Strong

There are a few positive moderate-level correlations are between “Frames” and

“Political Intervention” (ρ = 0.52), “Addiction References” and “Gambling References”

(ρ = 0.49), “Gambling References” and “Normative Judgments” (ρ = 0.43), “Addiction

References” and “Political Intervention” (ρ = 0.41) and “Normative Judgment” and

55 “Political Intervention” (ρ = 0.41). The relative distances between this cluster of moderate-level correlations is discussed further in the Results section.

The remaining positive correlations are all weak to very weak associations:

“Normative Judgment” and “Personal Testimonies” (ρ = 0.32), “Addiction References” and “Normative Judgment” (ρ = 0.29), “Addiction References” and “Personal

Testimonies” (ρ = 0.20), “Gambling References” and “Personal Testimonies” (ρ = 0.19),

“Story Type” and “Word Count” (ρ = 0.18), “Personal Testimonies” and “Political

Intervention” (ρ = 0.17), “Personal Testimonies” and “Media Outlet” (ρ = 0.16) and

“Normative Judgment” and “Media Outlet” (ρ = 0.16; H1a not supported). There were no significant correlations between Media Outlet and Gambling or Addiction References

(H1b and H1c not supported).

Likewise, all negative correlations are very weak associations, wherein as one variable’s value increases, its counter decreases, and vice versa: “Gambling References” and “Story Type” (ρ = -0.16), “Commentary Density” and “Story Type” (ρ = -0.17; H2b confirmed in part) and “Political Intervention” and “Story Type” (ρ = -0.19). There was no significant correlation between Commentary and Media Outlet (H2a not supported).

Although a very weak correlation, the relationship between Story Type and Commentary

Density indicates that as any given article approaches a news writing format, the total amount of framing and critical variables decreases. However, H2b cannot fully be confirmed because of the nominal leveling that defines the Story Type variable (i.e. beginning with “0” and ending in “3,” the categories go from Other, News Writing, Op-

Ed and Review). Although an argument could suggest that an “increase” in Story Type

56 indicates more opinion-based reports, this line of reasoning undermines the factual components of Reviews while simultaneously ignoring the vagueness of Other stories.

Despite this shortcoming, an equal line of reasoning highlights that as commentary decreases, the “Story Type” increases within the sample, thus suggesting—however weakly—that a small relationship exists between the level of commentary and the article formatting. Future studies might provide a more succinct procedure. Nonetheless, the omission of Other writing formats yields a weak, negative correlation (ρ = -0.23) significant at the 0.01 level. Therefore, H2b is only confirmed in part according to the assumptions of this study.

Table 3b

Correlations of Mainstream Media articles.

MSM Word Story Test. Poli. Norm. Gamb. Addc. Frames Cmmt. Word 1 0.14* 0.02 0.09 0.12 0.04 0.04 0.04 0.09 - - - Story 1 -0.05 -0.08 -0.10 -0.19 0.20** 0.22** 0.19** Test. 1 0.09 0.28** 0.15* 0.20** 0.13 -- Poli. 1 0.46** 0.78** 0.37** 0.50* -- Norm. 1 0.46** 0.29** -- -- Gamb. 1 0.46** -- -- Addc. 1 -- -- Frames 1 -- Cmmt. 1 nMSM = 199 * Significant at the 0.05 level (two-tailed). ** Significant at the 0.01 level (two-tailed). -- Nullified due to grouping variable (i.e. Frames = Norm. + Gamb. + Addc.; Cmmt. = Frames + Poli. + Test.) 57 Table 3c

Correlations of Gaming Media articles.

GM Word Story Test. Poli. Norm. Gamb. Addc. Frames Cmmt. Word 1 0.37* 0.14 -0.07 0.09 -0.04 -0.06 -0.14 0.02 Story 1 -0.09 -0.17 -0.07 -0.03 -0.12 -0.10 -0.13 Test. 1 0.41* 0.36* 0.33* 0.21 0.17 -- Poli. 1 0.32* 0.45* 0.52* 0.57* -- Norm. 1 0.44* 0.31* -- -- Gamb. 1 0.56* -- -- Addc. 1 -- -- Frames 1 -- Cmmt. 1 nGM = 75 * Significant at the 0.01 level (two-tailed). -- Nullified due to grouping variable (i.e. Frames = Norm. + Gamb. + Addc.; Cmmt. = Frames + Poli. + Test.)

A similar cluster of correlations occur in both Tables 2b and 2c. However, a second level of significance occurs for the MSM sample in Table 2b (p < 0.05). Aside from minor differences between the two subsamples, the MSM presents a strong, positive correlation between Gambling References and Political Intervention (ρ = 0.78).

Surprisingly, H2b is only supported in part by MSM media outlets (ρMSM = -0.19), suggesting that the lack of Reviews—or the outweighing News Writing reports—within the MSM is what validates this hypothesis. Likewise, omitting Other story formats results in a weak, negative correlation (ρ = -0.23) significant at the 0.01 level. Thus, it can be

58 concluded that H2b is confirmed in part by the MSM subsample of articles.

Additionally, the significance between Word Count and Story Type (ρMSM = 0.14;

ρGM = 0.37) can be omitted from these results. In conjunction with the argumentation provided regarding the Other subcategory of Story Type, there is no qualitative measure to suggest that the entirety of a single article pertains to loot boxes. In other words, certain GM articles may also contain subheads regarding the loot box controversy, addressing the matter only in part. However, the ρGM correlation with respect to Word

Count and Story Type does highlight a significant difference of the two variables between the GM and MSM outlets (Table 1). As indicated by Figure 1, the comparative ratios of News Writing to Review types of articles between the two outlets shows this significance because of MSM’s focus on News Writing. Although GM has been shown to possess News Writing ability, the Word Count-Story Type correlation within its subsample infers, potentially, that Reviews and Op-Eds receive slightly more attention than News Writing.

59

Figure 2. Distribution of commentary density among sample size.

As mentioned, the Commentary Density is the summation of Frames, Political

Intervention and Personal Testimonies variables. This variable was created to characterize the extent a single article mentioned a frame or critical variable at least once when discussing “loot boxes.” For all articles, Figure 1 is characterized by a skew to the right, as 39.4 percent (108 articles) had no mention of any framing or critical variables.

Subsequently, 19.7 percent (54 articles) maintained very weak commentary, followed by

15.3 percent (42 articles) were moderate commentaries.

Likewise, one of the main focuses of this study is to predict the average number of frames and critical variables per article with respect to each media outlet. Tables 4a-d explore this using a linear regression of Commentary and Framing for GM and MSM

60 individually. Since each dependent variable is a scale comprised of framing and critical variables, the coefficients used in each table represent the independent variables that can be used for linear regression.

Table 4a

Commentary Regression for MSM subsample.

Commentary (DV) Regression for MSM subsample R 0.23 R Square 0.05 F 5.26 Sig. 0.01 Coefficients (IVs) Word Count Story Type Unstandardized Beta 0.00 -0.45 Std. Error 0.00 0.15 t 1.71 -2.97 Sig. 0.09 0.00

For instance, Commentary can only be predicted using Word Count and Story

Type because the scale contains all framing and critical variables. Table 4a shows that only Story Type can be used to predict the level of Commentary within the MSM outlet

(p < 0.01), supporting H2b in part. However, the same does not hold for the GM subsample as shown by Table 4c (p = 0.21). Thus, a larger sample of GM articles is needed to fully conclude an association between Commentary and Story Type between

Media Outlets.

61 Table 4b

Frames Regression for MSM subsample.

Frames (DV) Regression for MSM subsample R 0.54 R Square 0.29 F 10.36 Sig. 0.00 Personal Political Coefficients (IVs) Word Count Story Type Testimonies Intervention Unstandardized Beta 1.30E-5 -0.10 0.31 0.78 Std. Error 0.00 0.11 0.15 0.13 t 0.18 2.04 2.04 5.79 Sig. 0.86 0.35 0.04 0.00

Similar results occur among Personal Testimonies between the MSM (p = 0.04) and GM (p = 0.58); while the former is significant at the 0.05-level, the latter cannot be used as a predictor of Frames for any given GM article, despite having a significant difference in means (Table 1). Thus, a greater number of GM articles is needed to support

H1a through H1c. Despite this, Political Intervention is a reliable predictor of Frames between the two media outlets, while Personal Testimonies can only be used to predict

MSM frames.

Most notably, linear regression predicting both Frames (p < 0.01) and

Commentary (p = 0.01) for MSM articles are both significant are similar p-levels.

Although these findings do not support significant differences according to the proposed hypotheses, there is a reliable difference between Frames predicted by Political

Intervention among the two media outlets not hypothesized by this study.

62

Table 4c

Commentary Regression for GM subsample.

Commentary (DV) Regression for GM subsample R 0.15 R Square 0.02 F 0.82 Sig. 0.44 Coefficients (IVs) Word Count Story Type Table 4c: continued Unstandardized Beta 0.00 -0.27 Std. Error 0.00 0.21 t 0.64 -1.27 Sig. 0.52 0.21

Table 4d

Frames Regression for GM subsample.

Frames (DV) Regression for GM subsample R 0.58 R Square 0.33 F 5.14 Sig. 0.00 Personal Political Coefficients (IVs) Word Count Story Type Testimonies Intervention Unstandardized Beta -9.25E-5 0.07 -0.13 0.99 Std. Error 0.00 0.13 0.23 0.24 t -0.42 0.50 -0.56 4.17 Sig. 0.68 0.62 0.58 0.00

Hence, Political Intervention can also be used to predict Frames within the GM subsample of articles. Likewise, it is the only observed variable that makes linear

63 regression for GM Frames a reliable measure (p < 0.01), unlike GM Commentary predictions (p = 0.44). This observation reinforces the need for more GM articles to elaborate on the significant differences between Media Outlet means. Only then can the first set of hypotheses be further investigated without scrutiny from a greater amount of

MSM articles.

64 Discussion

Despite the lack of confirmed hypotheses regarding differences in the ways GM and MSM report loot boxes in video games, Political Intervention has shown to be a reliable predictor of Frames between both Media Outlet categories. This is foreshadowed additionally within Table 1, wherein the scaling variable Frames is not significantly different between the MSM and GM until Political Intervention is added to the scale (p <

0.01). Arguably, this finding suggests a congruence between the types of reporting GM and MSM conduct about loot boxes. In other words, for framing to occur in larger amount, it is possible that some form of policy or regulation of loot boxes is present (i.e. discussion of gambling among youths might invoke associative frames such as addictive personalities or normative judgments). Thus, not only MSM agents are aware of the amount of framing that occurs with respect to loot box reporting.

The Gaming Media, largely thought of to be an enthusiast press, is nonetheless a journalistic practice to a certain degree (Foxman & Nieborg, 2016). Although this study cannot confirm that the MSM contributes more/less frames to the loot box debate than the

GM, it can highlight that there are significant differences between the two outlets in terms of Gambling References, Political Intervention, Personal Testimonies and Story

Type. Thus, gaming media outlets are nonetheless conscious contributors on the topic of loot boxes, and their inputs may align with those of the MSM. As mentioned in the literature review, MSM reports on video gaming tend to imply more negative connotations about gaming media and its effects on children (McKernan, 2013;

Copenhaver et al., 2017). This study implies that GM may share these sentiments—

65 perhaps to a lesser extent—to convey a sense of journalistic responsibility; if the gaming media caters to the gaming community, it would logically be worthwhile to inform their audience of the debatable effects of loot boxes if they are deemed “predatory” (Dingman,

2017). Thus, some degree of conscious reporting is true for even an ambivalent media as gaming.

RQ1a: What themes generally define the “loot box” controversy?

After running three intercoder reliability tests, a string of variables defined by

“Uses & Gratifications” required omission because of substantially low kappa scores.

Upon inspection of the codebook used throughout the testing, these variables were determined by the coders to contain overlapping definitions. Additionally, beta testing was conducted using a double-blind methodology to constrict bias at the cost of accuracy

(i.e. no control beta articles with fixed inputs in mind were issued to the coders with fixed inputs). Thus, future research may consider an approach using thoroughly tested control articles instead of a larger attempt to reduce bias.

Since a gaming genre did not precisely limit this study, a single use or gratification could not be uniformly applied to all games containing loot boxes; nor would the subject of the article regard a use or gratification with respect to loot boxes.

This predicament occurred when considering a game review that spoke at length about gaming mechanics but made no mention of “loot boxes.” Indeed, an empirical observation of gaming’s uses and gratifications would be severely lacking if no actual game was observed first-hand. Therefore, the preceding argumentation in tandem with

66 the poor Cronbach alphas resulted in the omission of this tertiary-level analysis (i.e. how a single media frames another media’s uses and gratifications).

Nonetheless, the data suggest—albeit, weakly—that regardless of media outlet, there are congruent themes that generalize the “loot box” discussion as predicted by H1 and H2. The web of relationships between the framing and critical variables show that controversy regarding loot boxes tends to fragment at least two individual variables together rather than a full commentary cluster. On average, MSM gave 1.43 frames per article with a standard deviation of 1.54 frames, while GM gave 1.57 frames with a standard deviation of 1.57 frames. Thus, it is unlikely that “loot boxes” reports are sensational according to either medium, yet the relationships between the individual framing and critical variables demonstrate how the discussion was amplified.

RQ1b: How are they related?

Table 2 maintains a cluster of positive, moderate-level correlations between

Political Intervention, Personal Testimonies, Gambling References, Addiction References and Normative Judgments. One positive, moderately strong correlation between Political

Intervention and Gambling References (ρ = 0.70) confirms a logically sound association; this can also be said of Political Intervention’s correlation with Addiction References (ρ =

0.41). Despite moderate strength, associations within this level and above confirm what might logically follow from the variables’ definitions and their implications: normative judgments increase as references made to gambling and/or addiction (i.e. vice) increase; if the summation of these variables (e.g. “Frames”) increases, mentions of policy increase.

67 Subsequently, this web of relationships appears to act as the basis for the remainder of findings; regardless of media outlet, the amount of commentary provided by a single article tended to increase, even by a small amount, as the story type progressed toward more fact-based types of reporting. However, further inspection of this finding revealed that the MSM subsample of articles were the primary support for this hypothesis because the GM subsample was insignificant for H2b. Thus, there is a clear difference between GM and MSM media outlets judging by the discrepancy of News Writing and

Review story types within each outlet. As mentioned, future studies might seek to develop ordinal criteria for a fact-to-opinion scale.

RQ2: What comparisons can be made between MSM and GM reports on “loot boxes?”

Although Personal Testimonies and Normative Judgments both tended to increase as Media Outlets shifted toward Gaming Media, a further qualitative examination is needed to characterize the precise ways in which “loot boxes” are being described, combatted or scrutinized. While it is surprising to find that Gaming Media—albeit, marginally—appears to frame “loot boxes” more frequently than its mainstream counterpart, this might be the result of an additional variable: a (general) gaming community. Indeed, communication scholars must not underestimate the impact a consumer culture has on its own media. Therefore, future studies may wish to incorporate some form of cooperative research akin to Carras et al. (2017).

As also mentioned, several substantial limitations constrict the inferences that can be discussed within the parameters of this study. Most prominently, the subsample of GM

68 articles restricts any further investigation of the statistically significant differences between GM and MSM outlets. Political Intervention is the only reliable variable that can used to predict Frames for each media outlet category, even though there are observable differences in means of Gambling References, Personal Testimonies and Story Types.

The simplest solution to this conundrum would be to re-examine the data using a larger subsample of GM articles.

Subsequently, a potential byproduct of this increase could reconcile the severe lack of intercoder reliability during the first two steps of testing. The κ-scores for previous variables were not listed in the preceding data because they were deemed either weak or very weak. As a result, it was not possible to observe any variables pertaining to the uses and gratifications of video games that contain loot boxes. One possible solution to this would be to increase the amount of control articles used for reliability testing.

However, this study’s design brings forth the question of whether frames or uses and gratifications of gaming media—both video games themselves and their media equivalent in streaming and infotainment—should be used to evaluate the loot box controversy.

Therefore, I recommend the latter for a qualitative analysis to better conceptualize what impact loot boxes have on gamers themselves. As mentioned, future studies regarding the media frames of video games may consider codifying the “legitimacy” of video games in

MSM as well as further contextualizing loot boxes (i.e. what effects do loot boxes have in a specific game, rather than a generalized effect?).

It is clear from this study that media frames pertaining to normative judgment, gambling, addiction, testimonies and regulation can be used to convey loot boxes in a

69 certain light. However, there is no qualitative measure that can determine what effect they have on their audience according to this study alone. Thus, I must reiterate the importance of some qualitative measure to examine the impact loot boxes have on not only public perception but the gaming community itself.

Additionally, this data shows that even gaming media professionals are not entirely complacent with the speculative effects and perceptions of loot boxes; rather, it is more reasonable to believe that because the loot boxes of Star Wars Battlefront II directly affected gameplay in such a dubious manner, it was an irrevocable responsibility for reviewers to denote the predatory nature of such a system. Although this study made little mention of the gaming community beyond gaming media outlets, the spread of controversy would have been bolstered by opinion leaders on streaming and video- sharing platforms such as Twitch and YouTube. Thus, gaming media personnel might best use this data to examine how their audiences have alternatively engaged with the topic of loot boxes. Similarly, future researchers could question how various networks of the gaming community influence reports on gaming themselves, if there is a connection at all between them.

Political Intervention: An Unforeseen Conclusion

Given Political Intervention’s prevalence among the results, there is no doubt that political agency plays a strong role in the framing of video games with loot boxes in the media. Less matter-of-factly, however, are the implications that follow this predictor of frames and overall commentary; the qualitative measure to which we can rationalize these implications to follow must be supported by future case study. Nonetheless, the

70 political pressure to regulate video games with loot boxes is substantive enough to warrant concern from the ESRB and the ESA.

Should political intervention by some means of expression prove to be a valid measure of frames predicted, what exactly would stop this agency from controlling the conversation about video game loot boxes? Although this study attempts to simultaneously depict the nuanced and general attitudes toward loot boxes, positions of power might make greater use of their influence to gloss over the former. One possible outcome of this, as suggested through Rosenberg’s (2017) take on loot boxes, would be independent developers receiving the axe financially because smaller studios tend to rely on more service-based forms of income, such as microtransactions and loot boxes.

It would also be easier to assume the more predatory nature of loot boxes as the ailment—or norm—of greed within the games industry rather than just a symptom. As

Sinclair (2018) mentions in his detailed inquisition toward loot box self-regulation:

Star Wars: Battlefront II was pilloried for introducing a loot box scheme where the box contents would give players competitive advantages, yet games like FIFA Ultimate Team and Hearthstone mostly get a pass from their communities despite tying purchases and performance together more directly. (6)

While it may be unclear to those unfamiliar with games how these scenarios co- exist, such speculation is an apt insight toward cognitive dissonance in the loot box conversation; perhaps it is naïve to assume that since players of FIFA Ultimate Team and

Hearthstone are mostly blasé about their own microtransactions that they pay no mind to other more lucratively predatory systems in different games. It is entirely possible that cognitive elements are added and/or substituted to compensate for the narrative of loot boxes within video games today, assuming loot boxes were never considered to be

71 gambling in some form (Rosenberg, 2017). Thus, arguments in favor of nuanced views on microtransactions could combat a more generalized narrative of loot boxes as gambling mechanisms to assist self-regulation attempts. Nonetheless, Political

Intervention’s impact on this study was not entirely unexpected given the citations of themes throughout games reporting (Copenhaver et al., 2017; McKernan, 2013). As mentioned, the impact it makes offers an insight toward how the ESRB and the ESA can act to combat outside pressures toward federal regulation of games containing loot boxes.

For starters, several case studies could be employed to examine the ways in which games’ players interact with loot boxes in both an amicable consumer rapport and those that are less than friendly. Such qualitative examinations can provide uses and gratifications of positive loot box microtransaction while simultaneous cataloging the ways in which such interactions become predatory and/or exploitative. In any case, it must be re-emphasized that political agency can overtake the conversation on loot boxes unless the ESRB, ESA and the gaming community can prove—as in has in the past—that self-regulation of games is possible.

72 Conclusions

While only one of the working hypotheses was supported by this study’s findings, the aforementioned limitations in conjunction with significant t-tests implies that there is more to be done with the media framing of loot boxes—and video games in general. Although future researchers might consider categorizing samples by story type to examine which type could best predict frames, the impact political intervention had on this study has shifted the focus toward possible legal ramifications of loot boxes, especially according to those in power unfamiliar with video games. However, like any other entertainment media, our nation’s leaders and political actors are under no obligation to investigate the nuances of this loot box matter. To best conceptualize the conflicting space the loot box controversy holds in this study, and—perhaps—under the larger umbrella of media, a quote from Ian Bogost (2015) on the state of affairs with games’ popularity today will suffice:

The downside … of having games books and games degrees and games festivals and games retail channels and games communities—is thinking that their influence and their impact extend farther than they really do … (L)ike it or not, games are still a niche tricked by the echo chamber of internal success into thinking that they are approaching the mainstream. (203)

In a self-defeating sense, Bogost’s (2015) statement is a stern reminder of our nascent understanding of games from a media research perspective, as well as how we treat games as media themselves without a critical viewpoint; why take seriously the first amendment protections offered to games if their marketable purpose is akin closer to household appliances? Should games have less to offer in terms of expressive value, there is arguably little holding back legislation that regulates the content within games, especially that of loot boxes. Yet, history has shown the opposite to be true; games are as 73 serious as any other media, and Bogost’s (2015) words act as checks and balances so that gaming understands its cultural boundaries.

By no means does Bogost’s (2015) account represent the entire views of this study on games representation within the media. This is much more to be said on how an interactive media functions in our society and what facilities it offers to its players—not to mention the prevalence of “serious games,” otherwise known as games that focus on education through interactive media. On the contrary, Bogost’s (2015) work was inspiration enough to know not to take games too seriously altogether; growing up with games myself, it was easy for me to fall in and out of the antagonist view of pitting the mainstream media against games. Yet, upon observing the congruency of frames between mainstream and gaming media, it had begun to occur to me that a more mutual relationship could be possible to break the stereotypically hostile relationship between games and the layperson.

Coincidentally, this study shows that not only is the media outlet irrelevant in terms of framing loot boxes, but political intervention takes seriously the speculative repercussions of loot boxes such that it is the greatest predictor of frames and commentary on the matter. Given the current understanding of games media and how quickly political agency is capable of framing video games as imminent danger, it is imperative that the ESRB and ESA themselves exercise their authority on loot boxes before it is too late. While future research is nonetheless necessary to further constitute a meaningful investigation, this research shows that possibly taking games with loot boxes too seriously could mar reasonable action to self-regulate more predatory practices while

74 pardoning more amicable ones. If allowed to prove itself once again with self-regulation, the ESRB could retain its authority over games and the efforts that its creators make to sustain the industry.

75 References

Adiele, I., & Olatokun, W. (2013). Prevalence and determinants of Internet addiction

among adolescents. Computers in Human Behavior, 31, 100-110.

Alexandra, H. (2017). Loot Boxes Are Designed To Exploit Us. Kotaku. Retrieved

from: https://kotaku.com/loot-boxes-are-designed-to-exploit-us-1819457592.

Allen, J. J., & Anderson, C. A. (2018). Satisfaction and frustration of basic

psychological needs in the real world and in video games predict internet gaming

disorder scores and well-being. Computers in Human Behavior, 84, 220-229.

American Psychiatric Association. (2013). Internet Gaming Disorder. American

Psychiatric Association.

American Psychiatric Association. (1994). Gambling. American Psychiatric

Association.

Bailey, K., & West, R. (2016). Did I do that? The association between action video

gaming experience and feedback processing in a gambling task. Computers in

Human Behavior, 69, 226-234.

Bargeron, A. H., & Hormes, J. M. (2016). Psychosocial correlates of internet gaming

disorder: Psychopathology, life satisfaction, and impulsivity. Computers in

Human Behavior, 68, 388-394.

Bogost, I. (2015). How to talk about Videogames (Vol. 47). U of Minnesota Press.

Braun, B., Stopfer, J. M., Müller, K. W., Beutel, M. E., & Egloff, B. (2015).

Personality and video gaming: Comparing regular gamers, non-gamers, and

76 gaming addicts and differentiating between game genres. Computers in Human

Behavior, 55, 406-412.

Brewer, R. L. (2018). Writer’s Market 2018. Writer’s Digest Books, 97.

Brown v. Entm't Merchs. Ass'n, 564 U.S. 786, 131 S. Ct. 2729, 180 L. Ed. 2d 708,

2011 U.S. LEXIS 4802, 79 U.S.L.W. 4658, 22 Fla. L. Weekly Fed. S 1259 (U.S.

June 27, 2011).

Buelow, M. T., Okdie, B. M., & Cooper, A. B. (2015). The influence of video games

on executive functions in college students. Computers in Human Behavior, 45,

228-234.

Buono, F. D., Upton, T. D., Griffiths, M. D., Sprong, M. E., & Bordieri, J. (2015).

Demonstrating the validity of the Video Game Functional Assessment-Revised

(VGFA-R). Computers in Human Behavior, 54, 501-510.

Canale, N., Griffiths, M. D., Vieno, A., & Siciliano, V. (2015). Impact of Internet

gambling on problem gambling among adolescents in Italy: Findings from a

large-scale nationally representative survey. Computers in Human Behavior, 57,

99-106.

Carras, M. C., Porter, A. M., Van Rooji, A. J, King, D., Lange, A., Carras, M., &

Labrique, A. (2017). Gamers’ insights into the phenomenology of normal gaming

and game “addiction”: A mixed methods study. Computers in Human Behavior,

79, 238-246.

Comedy III Productions, Inc. v. Gary Saderup, Inc., 25 Cal. 4th 387, 21 P.3d 797, 106

Cal. Rptr. 2d 126, 2001 Cal. LEXIS 2609, 58 U.S.P.Q.2D (BNA) 1823, 2001 Cal.

77 Daily Op. Service 3380, 2001 Daily Journal DAR 4163, 29 Media L. Rep. 1897

(Cal. April 30, 2001).

Copenhaver, A., Mitrofan, O., & Ferguson, C. J. (2017). For video games, bad news

is good news: news reporting of violent video game studies. Cyberpsychology,

Behavior, and Social Networking, 20(12), 735-739.

Davis v. Elec. Arts, Inc., 775 F.3d 1172, 2015 U.S. App. LEXIS 154, 113

U.S.P.Q.2D (BNA) 1341, 43 Media L. Rep. 1073, 2015 WL 66510 (9th Cir. Cal.

January 6, 2015).

Destiny 2. (2018). Engrams. Destiny Game Wiki. Retrieved from:

https://d2.destinygamewiki.com/wiki/Engrams.

Dingman, H. (2017). How loot boxes are turning full-priced PC games in pay-to-win

games of chance. PCWorld, 30-34.

E.S.S. Entm't 2000, Inc. v. Rock Star Videos, Inc., 547 F.3d 1095, 2008 U.S. App.

LEXIS 23294, 88 U.S.P.Q.2D (BNA) 1691 (9th Cir. Cal. November 5, 2008).

Edgren, R., Castrén, S., Alho, H., & Salonen, A. H. (2017). Gender comparison of

online and land-based gamblers from a nationally representative sample: Does

gambling online pose elevated risk? Computers in Human Behavior, 72, 46-56.

Entm't Software Ass'n v. Blagojevich, 469 F.3d 641, 2006 U.S. App. LEXIS 29356,

35 Media L. Rep. 1008 (7th Cir. Ill. November 27, 2006).

Entm't Software Ass'n v. Chi. Transit Auth., 696 F. Supp. 2d 934, 2010 U.S. Dist.

LEXIS 1156, 38 Media L. Rep. 1257 (N.D. Ill. January 7, 2010).

78 Entm't Software Ass'n v. Swanson, 519 F.3d 768, 2008 U.S. App. LEXIS 5634 (8th

Cir. Minn. March 17, 2008).

Ferguson, C. J. (2013). TAMIU Student, Professor Research on Media Violence

Published. Texas A&M International University. Retrieved from:

https://www.tamiu.edu/newsinfo/newsarticles/2013-MediaViolence011713.shtml.

Ferguson, C. J. (2013). Violent video games and the Supreme Court: Lessons for the

scientific community in the wake of Brown v. Entertainment Merchants

Association. American Psychologist, 68(2), 57.

Forrest, C. J., King, D. L., & Delfabbro, P. H. (2015). The measurement of

maladaptive cognitions underlying problematic video-game playing among adults.

Computers in Human Behavior, 55, 399-405.

Foxman, M., & Nieborg, D. B. (2016). Between a rock and a hard place: Games

coverage and its network of ambivalences. Journal of Games Criticism, 3-1.

Funk, D. C., Pizzo A. D., & Baker, B. J. (2017). eSport management: Embracing

eSport education and research opportunities. Sport Management Review, 21, 7-13.

Gach, E. (2018). The FTC Says It Will Investigate Loot Boxes. Kotaku. Retrieved

from: https://kotaku.com/the-ftc-says-it-will-investigate-loot-boxes-

1830714932?utm_medium=sharefromsite&utm_source=kotaku_copy&utm_camp

aign=top.

Gardner, E. (2012). No Doubt, Activision Settle Lawsuit Over Avatars in ‘Band

Hero.’ The Hollywood Reporter. Retrieved from:

79 https://www.hollywoodreporter.com/thr-esq/no-doubt-activision-lawsuit-band-

hero-376217.

Garrelts, N. (Ed.). (2006). The meaning and culture of Grand theft auto: critical

essays. McFarland.

Good, O. S. (2018). Star Wars Battlefront 2’s rollback on loot crates clears up what

the game really is. Polygon. Retrieved from:

https://www.polygon.com/2018/3/16/17130402/star-wars-battlefront-2-loot-crates

progression-changes-opinion.

Grayson, N. (2018). CSGO Skin Traders Dumping Their Inventories After Valve

Announces New Rules. Kotaku. Retrieved from:

https://www.google.com/amp/s/steamed/kotaku.com/csgo-skin-traders-dumping-

their-inventories-after-valve-1824220495/amp.

Gunuc, S. (2015). Relationships and associations between video game and Internet

addictions: Is tolerance a symptom seen in all conditions. Computers in Human

Behavior, 49, 517-525.

Hallmann, K., & Giel, T. (2017). eSports—Competitive sports or recreational

activity? Sports Management Review, 21, 14-20.

Hart v. Elec. Arts, Inc., 717 F.3d 141, 2013 U.S. App. LEXIS 10171, 107 U.S.P.Q.2D

(BNA) 1001, 41 Media L. Rep. 1985, 2013 WL 2161317 (3d Cir. N.J. May 21,

2013).

Heere, B. (2017). Embracing the sportification of society: Defining e-sports through a

polymorphic view on sport. Sport Management Review, 21, 21-24.

80 Helton v. Good, 208 F. Supp. 2d 597, 2002 U.S. Dist. LEXIS 12436 (W.D.N.C. July

5, 2002).

Helton v. Hunt, 330 F.3d 242, 2003 U.S. App. LEXIS 10489 (4th Cir. N.C. May 27,

2003).

Hu, M., Zhang, M., & Wang, Y. (2017). Why do audiences choose to keep watching

on live video streaming platforms? An explanation of dual identification

framework. Computers in Human Behavior, 75, 594-606.

IGN. (2018). IGN Entertainment. IGN. Retrieved from: https://corp.ign.com/.

In re NCAA Student-Athlete Name & Likeness Litig., 2010 U.S. Dist. LEXIS

139724, 2010-2 Trade Cas. (CCH) P77,273, 2010 WL 5644656 (N.D. Cal.

December 17, 2010).

Jacques, C., Fortin-Guichard, D., Bergeron, P., Boudreault, C., Lévesque, D., &

Giroux, I. (2015). Gambling content in Facebook games: A common

phenomenon? Computers in Human Behavior, 57, 48-53.

Jenkins, H. (2004). Reality Bytes: Eight Myths About Video Games Debunked. PBS.

Retrieved from:

http://www.pbs.org/kcts/videogamerevolution/impact/myths.html.

Jerauld, N. (2018). Gambling with Gamers: The Global Controversy Surrounding the

Growth of Virtual Loot Boxes and Their Ties to Gambling. Syracuse Law Review.

Retrieved from: http://lawreview.syr.edu/gambling-with-gamers-the-global-

controversy-surrounding-the-growth-of-virtual-loot-boxes-and-their-ties-to-

gambling/.

81 Kardefelt‐Winther, D., Heeren, A., Schimmenti, A., van Rooij, A., Maurage, P.,

Carras, M., Edman, J., Blaszczynski, A., Khazaal, Y. & Billieux, J. (2017). How

can we conceptualize behavioural addiction without pathologizing common

behaviours? Addiction, 112(10), 1709-1715.

Keller v. Elec. Arts, Inc., 2010 U.S. Dist. LEXIS 10719, 94 U.S.P.Q.2D (BNA) 1130,

38 Media L. Rep. 1353 (N.D. Cal. February 8, 2010).

Keller v. NCAA, 2015 U.S. Dist. LEXIS 166546, 2015-2 Trade Cas. (CCH) P79,409,

2015 WL 8916392 (N.D. Cal. December 10, 2015).

Kim, T. (2017). State legislators call EA’s game a ‘Star Wars-themed online casino’

preying on kids, vow action. CNBC. Retrieved from

https://www.cnbc.com/2017/11/22/state-legislators-call-eas-game-a-star-wars-

themed-online-casino-preying-on-kids-vow-action.html.

King, D., Delfabbro, P. H., Kaptsis, D., & Zwaans, T. (2013). Adolescent simulated

gambling via digital and social media: An emerging problem. Computers in

Human Behavior, 31, 305-313.

King, D., Kaptsis, D., Delfabbro, P. H., & Gradisar, M. (2016). Craving for internet

games? Withdrawal symptoms from an 84-h abstinence from Massively

Multiplayer Online gaming. Computers in Human Behavior, 62, 488-494.

King, D., & Delfabbro, P. H. (2015). Early exposure to digital simulated gambling: A

review and conceptual model. Computers in Human Behavior, 55, 198-206.

Kümpel, A. S., & Haas, A. (2016). Framing Gaming: The Effects of Media Frames

on Perceptions of Game(r)s. Games and Culture, 11(7-8), 720-744.

82 Laconi, S., Tricard, N., & Chabrol, H. (2015). Differences between specific and

generalized problematic Internet uses according to gender, age, time spent online

and psychopathological symptoms. Computers in Human Behavior, 48, 236-244.

Lang, A. (2013). Discipline in crisis? The shifting paradigm of mass communication

research. Communication Theory, 23(1), 10-24.

Lawrence, N. (2017). The Troubling Psychology of Pay-to-Loot Systems. IGN.

Retrieved from: http://www.ign.com/articles/2017/04/24/the-troubling-

psychology-of-pay-to-loot-systems.

Lopez-Fernandez, O., Honrubia-Serrano, M. L., Baguley, T., & Griffiths, M. D.

(2014). Pathological video game playing in Spanish and British adolescents:

Towards the exploration of Internet Gaming Disorder symptomatology.

Computers in Human Behavior, 41, 304-312.

Macey, J., & Hamari, J. (2017). Investigating relationships between video gaming,

spectating esports, and gambling. Computers in Human Behavior, 80, 344-353.

McKernan, B. (2013). The morality of play: Video game coverage in The New York

Times from 1980 to 2010. Games and Culture, 8(5), 307-329.

No Doubt v. Activision Publishing, Inc., 192 Cal. App. 4th 1018, 122 Cal. Rptr. 3d

397, 2011 Cal. App. LEXIS 174, 98 U.S.P.Q.2D (BNA) 1728 (Cal. App. 2d Dist.

February 15, 2011).

O'Bannon v. NCAA, 2016 U.S. Dist. LEXIS 44131, 2016-1 Trade Reg. Rep. (CCH)

P79,646, 2016 WL 1255454 (N.D. Cal. March 31, 2016).

83 Perloff, R. M. (2013). Progress, paradigms, and a discipline engaged: A response to

Lang and reflections on media effects research. Communication Theory, 23(4),

317-333.

Recktenwald, D. (2017). Toward a transcription and analysis of live streaming on

Twitch. Journal of Pragmatics, 115, 68-81.

Rosenberg, A. (2017). Gaming’s loot box controversy just took a dangerous new

turn. Mashable. Retrieved from: https://mashable.com/2017/11/22/ea-star-wars-

battlefront-loot-box-gambling/#3vSmCWhDLiqW.

Scutti, S. (2018). Do video games lead to violence? CNN. Retrieved from:

https://www.cnn.com/2016/07/25/health/video-games-and-violence/index.html.

Seo, Y. (2015). Professionalized consumption and identity transformations in the

field of eSports. Journal of Business Research, 69, 264-272.

Seok, S., & DaCosta, B. (2012). The world’s most intense online gaming culture:

Addiction and high-engagement prevalence rates among South Korean

adolescents and young adults. Computers in Human Behavior, 28, 2143-2151.

Shaw, A. (2017). Encoding and decoding affordances: Stuart Hall and interactive

media technologies. Media, Culture & Society, 39(4), 592-602.

Shaw, E. F. (1979). Agenda-setting and mass communication theory. Gazette (Leiden,

Netherlands), 25(2), 96-105.

Sinclair, B. (2018). Self-regulating loot boxes is harder than it sounds.

GamesIndustry.Biz. Retrieved from:

84 https://www.gamesindustry.biz/articles/2018-05-03-self-regulating-loot-boxes-is-

harder-than-it-sounds.

Sioni, S. R., Burleson, M. H., & Bekerian, D. A. (2017). Internet gaming disorder:

Social phobia and identifying with your virtual self. Computers in Human

Behavior, 71, 11-15.

Sjöblom, M., Törhönen, M., Hamari, J., & Macey, J. (2017). Content structure is

king: An empirical study on gratifications, game genres and content type on

Twitch. Computers in Human Behavior, 73, 161-177.

Sjöblom, M., & Hamari, J. (2017). Why do people watch others play video games?

An empirical study on the motivations of Twitch users. Computers in Human

Behavior, 75, 985-996.

Slopen, N. B., Watson, A. C., Gracia, G., & Corrigan, P. W. (2007). Age analysis of

newspaper coverage of mental illness. Journal of health communication, 12(1), 3-

15.

Spekman, M. L. C., Konijn, E. A., Roelofsma, P. H. M. P., & Griffiths, M. D. (2013).

Gaming addiction, definition and measurement: A large-scale empirical study.

Computers in Human Behavior, 29, 2150-2155.

State v. Blackmon, 304 S.C. 270, 403 S.E.2d 660, 1991 S.C. LEXIS 89 (S.C. April

22, 1991).

Toker, S., & Baturay, M. H. (2015). Antecedents and consequences of game

addiction. Computers in Human Behavior, 55, 668-679.

85 United States v. Lynn, 1999 U.S. App. LEXIS 3026 (6th Cir. Tenn. February 19,

1999).

Video Software Dealers Ass'n v. Maleng, 325 F. Supp. 2d 1180, 2004 U.S. Dist.

LEXIS 13533, 33 Media L. Rep. 1017 (W.D. Wash. July 15, 2004).

Watters, C. A., Keefer, K. V., Kloosterman, P. H., Summerfeldt, L. J., & Parker, J. D.

A. (2013). Examining the structure of the Internet Addiction Test in adolescents:

A bifactor approach. Computers in Human Behavior, 29, 2294-2302.

Willett, B. (2018). If Game Makers Don’t Address Loot Crates, Lawmakers Will. The

Federalist. Retrieved from: http://thefederalist.com/2018/03/01/game-makers-

dont-address-loot-crates-government-will/.

Wimmer, R. D., & Dominick, J. R. (2013). Mass media research. Cengage learning.

Yin-Poole, W. (2018). Now Belgium declares loot boxes gambling and therefore

illegal. Eurogamer. Retrieved from: https://www.eurogamer.net/articles/2018-04-

25-now-belgium-declares-loot-boxes-gambling-and-therefore-illegal.

86 Appendix: Loot Box Codebook

1. Media outlet: variable describing any source that channels information 0 - Mainstream media (MSM): any media outlet that distributes news to a mass audience; to differentiate this variable from gaming news outlets, MSM sources are additionally defined as outlets whose primary content focuses on politics, entertainment and information unrelated to video games. Below are examples of sources coded as MSM for the purposes of this study (non-exhaustive): i. NBC, USA TODAY, MSNBC, ABC, FOX, WSJ, New York Post, CBS, CNN, The New York Times, NPR, BBC, PBS, The Washington Post, Forbes, Business Insider ii. INCLUDE LOCAL NEWS PROVIDERS (e.g. WTOV9, WTRF) iii. INCLUDE ALL OTHER MEDIA OUTLETS (e.g. industry trade press, newswires & press releases, aggregate news sources) 1 – Gamers’ media (GM): any media outlet whose primary content focuses on gaming news and commentary. Contrary to MSM, GN sources may or may not feature sections of gaming such as “culture” or “tech” (i.e. as opposed to “gaming” as a section itself). Below are examples of GN sources in this study (non-exhaustive): iv. Polygon, Kotaku, IGN, Eurogamer, Gamezebo, Game Informer, Destructoid, PC Gamer, GamesRadar+, Metacritic, Nintendo Life, GameSpot, VG247, Joystiq, GoNintendo, Shacknews, Touch Arcade 2. Length of story: word count 3. Story type: the writing format, or stylistic structure, of a report. These categories include: 1 - News writing (fact-based reporting with little personal input/commentary) 2 - Op-Ed (i.e. look for use of “I,” “me,” “my”) 3 - Review 0 - Other 4. Personal testimonies: quotes that detail personal anecdotes about tropes such as “using a (parent’s) credit card to make large purchases,” reports of individuals “spending more than $100 (USD) on microtransactions;” code 1 ONLY IF a testimony is in a quote. DO NOT code 1 if the anecdote is paraphrased; a personal testimony MUST BE A DIRECT QUOTE. Otherwise, code 0. 87 5. Political intervention: code 1 if there is any reference to pending or successful legislation that regulates “loot boxes,” disclose odds of winning, etc.; otherwise, code 0 if not applicable. 6. Valence frame (Framing series; FR): code 1 if there is an explicit use of “good” or “bad” OR AN INFERENCE OF GOOD AND BAD (e.g. loot boxes -> [underage] gambling <-> illegal -> bad; therefore, loot boxes -> bad) when referring to “loot boxes;” otherwise, code 0. 7. Gambling reference frame (FR); explicit references to real-world gambling that can be combined with virtual terminology; code 1 if keywords are found; otherwise, code 0: a. characterizing “games as virtual online casinos,” “loot boxes as virtual slot machines;” “the sounds and visuals of receiving a reward are like winning at the slots;” other references to the “flashiness” of rewards and their likeness to winning at the casino; “virtual gambling of skins/items;” “problematic gambling;” “stealing money/credit cards” to obtain more 8. Video game addiction reference frame (FR); explicit references to internet gaming disorder (IGD); code 1 if keywords are found; otherwise, code 0: a. Keywords: “World Health Organization acknowledgement of video game addiction;” “academic failure due to game addiction;” “playing despite knowledge of problem;” “pathological gaming,” “social isolation,” “excessive play,” “withdrawal” effects, and “stress” due to excessive play; explicit use/colloquial use of “addict/addicted/addiction;” “reference to the term “Internet Gaming Disorder/IGD”

88 ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

Thesis and Dissertation Services ! !