Do Games Make Young People Less Engaged?

Paul Musgrave points to a pair of interesting papers by Pavel Bačovsky, a CU alum (represent!) and currently a scholar at Bates College, into what effects playing video games may have on political engagement in young people. I wish he had consulted his peers in communication research. The first, “Gaming alone: Videogaming and sociopolitical attitudes,” published at New Media and Society, is worth unpacking in full. From the abstract:

Using the Swedish Political Socialization Panel dataset and partial-pool time series methodology, I investigate the relationship between playing videogames and adolescents’ political and social attitudes over time. I find that those gamers who spend more time engaging in their favorite pastime become less interested in sociopolitical issues and less prosocial than non-gamers from year to year. My findings tell a cautionary tale about the adverse effects of extensive gaming on the development of democratic attitudes among adolescents.

The adverse effects of extensive gaming is a thoroughly researched topic, and it’s suspect that Bačovsky uses 10-20 year old studies. Unfortunately, in doing so, he amplifies some debunked research in his theoretical development that has been largely discounted in game studies research. This is not necessarily his fault, because games causing adverse psychology is a conventional wisdom. But the peer reviewers should have caught it and asked him to update his framework. Let’s examine the claims one by one.

Heightened exposure to violent digital media is linked to increased levels of aggressive thoughts and behavior (Anderson and Dill, 2000), in both the short and long term (Anderson et al., 2003). Playing violent videogames may also lead to decreased empathy, prosocial behavior, and civic engagement (Anderson, 2014; Anderson et al., 2010)

Anderson’s research suffers from the same problem that plagues academic psychology more broadly, but especially in games research: he didn’t do rigorous research. This, again, isn’t a personal dig at Anderson, though I am aware of his entire career being oriented around problematizing gaming. But, the field more broadly faces this problem: the vast majority of psychology research into game effects did not bother to measure the effect of competition on metal state and arousal until 2019. I’m not joking: nearly 30 years of psych research alleging a connection between gaming and aggression never asked if competition could lead to aggressive feelings. That might have an effect to consider when investigating social attitudes, especially when only measured right after play! Dowsett and Jackson finally examined competition in 2019 and came to a conclusion that should prompt a cautious uptake of the previous body of research:

While numerous studies have assessed the impact of violence within video games, only a few have assessed the impact of competition. This study found that competition increased aggressive affect but not aggressive behaviour, and when participants lost in the competitive condition their aggressive affect increased even further. In contrast, violence within the video game had no effect on aggressive affect or behaviour. In addition, a combination of violence and competition did not increase aggression further than competition alone. This study provides further evidence that competition in video games plays a significant role in determining aggression and should be included in future studies.

Studying video games without accounting for competition is like studying sports without accounting for competition: it is an arbitrary fencing of a central concept with a pretty clear predisposition toward one type of conclusion. So, if you’re basing one of your theoretical foundations on 20-year old studies that suggest games “cause” aggression, you really need to have a rethink. This could lead to a much more interesting question: how do competitive environments affect socio-political attiudes? This is probably answerable using Bačovsky’s other paper on hobbies that Paul links to, and perhaps even in the dataset Bačovsky used for his current study, but it’s very different from asking whether games are the causal agent here. Bad questions, bad results. Speaking of:

Furthermore, a wealth of research suggests that even online interactions with other players may fail to address the many adverse effects of gaming on youth political behavior. Kraut et al. (1998) and Young and Rogers (1998) connect excessive Internet use to shrinking of social circles, which leads to increased rates of depression and alienation. In his study on the role of the Internet in political participation among adolescents from Scandinavia, Canada, and the United States, Milner (2010) cautioned against the optimism toward digital media as a stimulant of political engagement, adding that “the Internet has not yet lived up to its potential” as an instrument of youth political involvement (Milner, 2010: 73). Psychologists also warn that prolonged online gaming may erode existing interpersonal relationships and lead to social isolation (Williams, 2006a). Using a sample of Belgian teenagers, Quintelier and Vissers (2008) found a negative correlation between online gaming and political participation.

Again, this is a rich topic and it’s really suspect to me that these studies are all so old. For starters, internet use, and the internet’s role in society, is fundamentally different in the 2020s than it was in the 1990s. More importantly: there are lots of recent studies about this! As one example, in 2010 Jennifer Brundidge found that online political discussion increased exposure to heterogenous views, which is very much the opposite of the 1998-era research Bačovsky cites. More pointedly, as uses and gratifications theory has shown, people use the internet differently from how they use video games. There are strong parallels with how people consume sports media and video gaming media, for example, with data showing they derive similar gratifications if at different intensities. People use different social media platforms because they derive different gratifications from different uses on them. It is not appropriate to use broad (and decades old!) internet research and to apply that to video games.

Also? Data show that the correlation between internet usage and depression, is, like with gaming, misattributed to the internet itself when it seems to be more closely tied to physical inactivity than any specific medium. In other words, it is the sitting down and not moving for long periods of time that seems to account for the mental health outcomes of surveys and experiments, not some magical attirbute of the game or the internet itself. These studies are straightforward to get published and align with the biases of many (especially older) academics who find everything online and gaming problematic, but they just aren’t very good as studies. Sorry.

Back to Bačovsky:

Clinical psychologists have listed several of online gaming’s adverse effects, including loneliness, aggressiveness, alienation, and anxiety, under the umbrella term of the “Internet gaming disorder” (Gentile et al., 2011, 2012; Müller et al., 2014). Both online and offline variations of “gaming disorder” have also been recognized in the 11th Revision of the International Classification of Diseases as equivalent to other addictive or substance abuse behaviors (World Health Organization, 2018).

Sorry, again, but gaming disorder is, let’s be gentle, a contested idea (the invocation of a WHO designation is its own problem). Multiple psychologists have argued against it existing in the first place, or, if they concede it might exist, note that it is based on flimsy evidence and logical fallacies.

Again: if Bačovsky had engaged with recent studies on this, I don’t think he would have a solid theoretical foundation to then proceed with the rest of his study. I want to reiterate that I don’t intend this as a personal criticism, but rather as identifying a systemic problem with relying on psychology research that is reified through a deeply flawed peer review process. Psychology, especially behavioral and media research, has a massive replication crisis for good reason: lots of poorly designed studies have been accepted as conventional wisdom when they really shouldn’t be.

I am personally vested in this question, however! It is the topic of my own research: how is the military using video games to strategically communicate, and does it work? Applying a media effects paradigm to video games is difficult to do well: as Hopp noted in 2018, the initial observed effects in gaming experiments do not last, which calls into question the utility of surveys and experiments for arguing that gaming affects attitudes. In my specific field, decades of worry about the political indoctrinating effects of military themed games has not resulted in more positive societal attitudes toward the military–rather, the opposite has occurred (see Phillip Hammond’s opening essay in War Games: Memory, Militarism, and the Subject of Play for more on this). In specific gaming titles, there seems to be no real causal relationship between gaming and attitudes. In 2013 Marcus Schulzke, also a political scientist, examined a decade of the America’s Army game to see whether it was conveying a realistic picture of military service, but also whether or not people were persuaded to enlist because of it (the key purpose of the game). He decided the answer to both questions was “no”: most people who played it did so because it was free to play and offered a degree of realism they did not get from other games. Rather than infinite bullets, item drops, and health bars, the game instead focused on tactical realism and one-shot kills, which is a closer match to reality but not the lived experience of service. Injuries are not lifelong maimings, there is no blood, and the drudgery of military life (“hurry up and wait” is the running joke) is ignored. Similarly, while the Army itself claimed in 2003-2004 that there was a major uptick in enlistment tied to the game launch, this also coincided with a renewed interest in military service after the 9/11 terror attacks and the 2003 invasion of Iraq, and the Army never differentiated its enlistment numbers in enough detail to say for certain. It seemed to be overbaked.

There are other ontological issues with the methods in that paper, like how “gamer” is defined in a general social survey, the generalizability of studying Swedish teens in a small city, and the many statistical hacks designed to pull insights from a non-specific data set. It isn’t productive to examine all of them in detail right now, but an exampple might help demonstrate my concern. Bačovsky says that he “predict[s] the influence of gaming on adolescents’ interest in politics and social issues,” but does not define any sort of baseline for this definition. The Swedish survey data seems to only ask a general question about gaming based on a Likert scale. From this, Bačovsky averages the answers to conclude that more than 50% of the adolescents polled say they play a game for 30 minutes or more per day. This is a highly skewed finding: most studies of youth gaming have far lower thresholds (1 hour per week is standard) which suggests either the Swedish survey population is unique or it is measuring something totally different. What Bačovsky is using is not a standard or accepted measure of “gaming,” but rather is measuring heavy game use, and not considering the type of game being played: a distinction with real implications for drawing broad conclusions about society! Gaming and identity have a complex relationship that just isn’t captured here.

More narrowly: game play type matters at least as much as game play quantity (women playing Animal Crossing for 30+ hours per week have a different set of issue networks than men playing Call of Duty for 30+ hours per week), and that is a confounding nuance that just could not be captured in a general survey instrument. Which is why I’m always deeply skeptical of their use for answering causative questions about beliefs. In fact, this is a genre of question no one is really sure how to answer. Game studies folks suspect it could be structural. In the 2000s, Ian Bogost developed his idea of procedural rhetoric to argue that the mechanics of games, and the act of playing those mechanics, constituted an argument about the worldview of the game designers. While he has backed off many of those initial ideas (game studies is a new field and constantly updating its priors–another reason why the reliance on decade-old studies is so suspect), he was right about the content of the game mattering as much as the time being played. Mary Flanagan observed in 2006 that games communicate values, even if the designer did not intend to: they say one thing is valuable over another, they instruct players to prefer one way of solving puzzles over another, they tell players who is important and who can be discarded. To repeat: that is a nuance about how gaming affects people and beliefs that cannot be captured in a stats hack on survey data.

There is emerging research that extraludic communities–that is, communities about gaming, but not in games themselves–play a far stronger role in shaping attitudes than the game itself. This is one of the foci of my work: moving beyond the game itself to look at what social role gaming as a concept plays in attitude formation and transmission. Gaming communities can organize around misogyny, or they can organize around the joy of crafting. They can be queer, or toxicly masculine. It’s complicated!

But, as Musgrave notes, this is an interesting topic that deserves investigating: does playing games influence youth political engagement? Sadly, that isn’t answered in this paper, which engages in issues-fencing I dispute (what’s a “nonpolitical hobby?”) and limited study design. And he’s right that this is a question to be taken seriously. But, as a part of a very old genre of research into youth gaming since the early 1980s, it keeps asking the same question again and again without asnwering it rigorously. I just wish there was a stronger emphasis in the field, and in its journals, on good theory and good study design.