Who are Adult Gamers? A Data Analysis

Update 27th July 2018: The response to this has been difficult to read from some corners of the internet. I would like to reiterate that the introduction below establishes that I am curious about people who choose to identify as a gamer despite decades of negative discourse about how they are violent, childish time-wasters. I am not trying to pretend that every participant has beaten Dark Souls at Soul Level 1 or participated in an eSports tournament, and this is clearly acknowledged in the Criticisms section. Data analysis allows us as researchers to empirically explore our curiosities, and I was curious to explore and humanise the people who are happy to say they are a gamer despite the negative stigma this might carry. I did not take bribes to skew results, I did not tamper with the data, and I am not trying to ‘ruin gaming’ or further an agenda. Thank you for understanding and please have a nice day.

The video game industry is an economic powerhouse, earning over $100 billion in revenue in 2017 and employing thousands of workers (Statista, 2018a; McDonald, 2017; Zackariasson & Wilson, 2012). This hard work results in one of the world’s most popular hobbies, with global research placing the number of people who play video games at over two billion (Statista, 2018b).

Despite the popularity of gaming and its contribution to the economy, gaming as a hobby has received numerous criticisms that may not be levied at hobbies such as reading and watching television. Three of the most prevalent criticisms include:

1. Gaming, more so than other hobbies, is “a colossal waste of time” (Whittel, 2008). Parents may refuse to let their children play video games as they would prefer them to read or play outside.
2. Gaming promotes violent behaviour and rewards the suffering of others (McCartney, 2008).

[It is] wrong for anyone, child or adult, to spend long hours electronically rehearsing the prolonged agony and detailed humiliation of other human beings for their own amusement.

 

– McCartney, 2008

3. Gaming is a child’s hobby and adults shouldn’t play video games. This is debated in editorials on whether grown men should be allowed to play video games (Simons, 2013).

Something that may also be unique to gaming is how it can contribute to a person’s identity. While people who play a lot of chess may not identify as ‘chessers’, the term ‘gamer’ is used by someone to identify that their passion and love for video games contributes to their sense of self. There is no hard rule for what qualifies someone to be a gamer; if someone loves gaming and identifies as a gamer, they are free to do so.

The gamer identity is something that has always intrigued me. With criticisms that gamers may throw away their time, are violent and may be childish, I set out to explore who gamers actually are. Are they truly underachieving time wasters? Using post-graduate level data analysis techniques, I will be looking at whether the gamer identity can encompass different groups of people or whether they are just one big similar group of people. I will explore who these people are demographically and what their beliefs are regarding gaming.

To be clear, I will be looking at three main questions:

1. Who plays video games?
2. Who chooses to identify as a gamer?
3. Are there different subgroups of gamers or are they a homogenous population?

As usual, there will be a summary at the bottom if you do not wish to read everything. Findings will also be presented in nice diagrams if you wish to browse at your leisure. Please enjoy!

Contents

1. 1. The Data
   2. Who Plays Video Games?
   3. Who Are Adult ‘Gamers’?
   4. Gamer Subgroups: A Latent Class Analysis
   5. Discussion
   6. Criticisms
   7. Summary

The Data

The dataset that I will be analysing is provided by the Pew Research Center. Between June-July 2015, data was collected from 2,001 Americans regarding gaming, jobs and broadband via a telephone questionnaire. 1,012 females and 989 males answered the questionnaire, with an age range of 18-99 and a mean age of 52. For my analyses, I will be focusing on sociodemographic variables (gender, income, marital status etc.) and gaming variables.

All participants were asked nine questions pertaining to gaming. Participants were first asked if they ever play video games ‘on a computer, TV, game console, or portable device like a cell phone’. Following this question, participants were provided with six statements that they could respond to with either ‘True for most games’, ‘True for some, but not others’, ‘not true for most games’ or could refuse to answer/say they were unsure. These statements are:

• Video games help develop good problem solving and strategic thinking skills.
• Video games are a waste of time.
• Video games portray women poorly.
• Video games promote teamwork and communication.
• Video games portray minority groups poorly.
• Video games are a better form of entertainment than watching TV.

Participants were then asked if they agreed or disagreed with the following statements: ‘most people who play video games are men’, and ‘people who play violent video games are more likely to be violent themselves’. Finally, participants were asked if they considered themselves to be a gamer. A gamer was defined in the questionnaire as ‘a fan of gaming or a frequent game player’.

The dataset is available to download if you sign up to the website here.

Who Plays Video Games?

The first analysis conducted was an analysis of who reportedly played video games. All participants were asked the question ‘Do you ever play video games on a computer, TV, game console, or portable device like a cell phone?’. 925 (46.2%) participants said yes. I set out to explore what sociodemographic factors were related to people answering ‘Yes’ to this question and what their gaming beliefs are.

This was achieved by conducting Pearson’s correlations between this variable, sociodemographic variables and other gaming variables. Nothing too complicated for now, that’ll come later. In Pearson’s correlation, a large r value (with the maximum being 1) denotes a strong relationship between two variables. We want the p-value to be below 0.05 to show that there is a significant relationship between the two variables. In short, we want a big r and small p.

I will present the findings in a diagram so I don’t lose you all in a sea of statistics. So, what variables are related to people playing video games?

To summarise the diagram, people who play video games are more likely to: be younger, be a student, have kids under 18, have a job and live in households earning up to $70,000 per year. They are less likely to have ever been married or consider themselves to be Republicans. They don’t believe: games are a waste of time, games portray women or minority groups poorly, gamers tend to be male, or games make you more violent. They believe: games help problem solving and strategic thinking, games promote teamwork and communication, and games are a better form of entertainment than TV. Gender, highest education and type of job (e.g. manual labour, office work) were not related, with 11 more women reportedly playing video games than men.

Now that we have looked at who plays video games, it’s time to look at the gamers.

Who Are Adult ‘Gamers’?

The second analysis investigated who chose to identify as a gamer. To reiterate, the definition of a ‘gamer’ in this questionnaire is ‘a fan of gaming or a frequent game player’. Out of 925 people who play video games, 165 (17.8%) chose to identify as a gamer. Similar to the first analysis, I will be looking at sociodemographic factors associated with being a gamer and their gaming beliefs.

Time for a diagram and a summary below.

To summarise the diagram, gamers are more likely to: be male, be a student, be younger and vote for the Democratic political party. Gamers are less likely to have been married. Gamers are more likely to believe: games develop good problem solving and strategic thinking, games promote teamwork and communication, and games are a better form of entertainment than TV. Gamers are less likely to believe games are a waste of time or make people more violent. Interestingly, gamers have no clear pattern of beliefs for whether games portray minority groups and women poorly or whether gamers tend to be male. Income and raising a child under 18 became non-significant for gamers, meaning there is no clear relationship between these variables and being a gamer.

It is interesting to note that there are indeed differences between people who play games and those who consider themselves to be gamers; some non-significant variables become significant and vice versa. However, simply conducting Pearson’s correlations does not tell us whether different types of people relate to the gamer identity.

In order to achieve this, I will be moving from Pearson’s correlations to something called Latent Class Analysis.

Gamer Subgroups: A Latent Class Analysis

Latent Class Analysis (LCA) is a data analysis technique that I have professed my love for in previous research breakdowns. LCA involves creating data-driven categories for people based on how similar they are to one another (Haagenars & McCutcheon, 2002).

An example I have given to describe LCA in the past is asking people whether they play six games: Street Fighter, Halo, Crash Bandicoot, Tekken, Call of Duty and Spyro. LCA may generate three groups from this: fighting gamers, FPS gamers and platforming gamers. However, one part of LCA is that it does not specifically tell you how many groups it has found; the analyst must decide for themselves how many groups to retain based on specific criteria. Please feel free to ignore anything written in purple as it contains stats talk related to this criteria.

To be brief, the criteria for choosing how many groups to retain includes: low AIC, BIC and SSABIC values (Akaike, 1987; Schwartz, 1987; Sclove, 1987), a non-significant LRT (Lo, Mendell, & Rubin, 2001), and an entropy value close to one (Ramaswamy et al., 1993).

With this criteria in mind, I conducted several LCAs using only those who identified as a gamer. In the analyses, I included all sociodemographic variables found to be significantly related to being a gamer and all gaming belief questions. I decided to include all of these questions as I was interested in how they differ across groups. After conducting multiple LCAs, it was time to consult the selection criteria to decide how many gamer groups to retain.

 

The decision was made to retain four groups of gamers as it has the largest entropy value, a non-significant LRT and the differences in AIC and SSABIC are only slight between 4 and 5 groups; BIC actually begins to increase for five groups.

After deciding to retain four groups of gamers, a data file was generated which organised each gamer according to their respective group. This classification was added to the main file and each gamer group was looked at in isolation. I will briefly summarise each group in the diagram below and then discuss each group individually. Please note that each group’s response classification is too large to include in the article, so I will be uploading this for each group on an external site. You do not have to view it as I will be summarising each group under their subheading.

Group 1

Group 1 could be thought of as ‘Older Gamers’ (n = 48) due to their age range of 42-77; their response classification can be viewed here. The vast majority are married, but do not tend to have children under 18. This group has the highest proportion of females by a slight margin (52.1% female) and tends to favour the Democratic party in US elections.

This group believes that some games help with strategic thinking, promote teamwork, are better than TV and may portray women poorly. They do not believe most games do a poor job in portraying minority groups. Their response to the question ‘games are a waste of time’ is tied between ‘true for some’ and ‘not true’. This group disagrees with the idea that most people who play games are men and that gamers are more likely to be violent themselves.

Group 2

Group 2 could be considered the ‘Younger Gamers’ (n = 79) as they have the youngest age range of the groups (18-40); their response classification can be viewed here. In contrast with the previous group, they are overwhelmingly male (89.9%) and have the highest population of students across all groups (43.6%). The vast majority of the group are unmarried, don’t have children under 18 and vote for the Democratic party in US elections.

This group holds massively positive beliefs about video games. They strongly believe that games help develop strategic thinking skills, aren’t a waste of time, promote teamwork and are more entertaining than TV. They don’t tend to believe that games do a poor job of portraying minority characters, but they have the highest proportion of people who believe that games in some capacity depict female characters poorly. They are more likely to believe that most people who play games are men, but massively disagree (97.5% disagree) with the idea that gamers are more likely to be violent themselves.

Group 3

Group 3 could be considered the ‘Family Gamers’ (n = 35) as they have the highest population of parents of under 18s and may still be involved in raising a family; their response classification can be viewed here. This group is aged between 22-44 and is slightly more likely to be male and married. They are equally likely to vote Republican and Democrat in US elections.

‘True for some games’ is the dominant response for this group. This response is given for questions regarding gaming helping problem solving, gaming helping teamwork, whether gaming is a waste of time, and whether it is better than watching TV. This group does not believe that games portray women or minority characters poorly. They agree more than any other group that most people who play games are men.

This group is the only group that tends to believe that there is a relationship between video gaming and violence.

Group 4

When I introduced what LCA was, I said that it is a method of grouping people together based on their similarities. When I conducted LCAs with 4+ groups, I noticed that the analysis started grouping three people together. I wondered if these people had some significantly different sociodemographics or gaming beliefs.

Well, it turns out that it’s best to think of this group as the ‘error group’. The LCA grouped three people together who refused to state how old they are (one person said they didn’t know) or what their political beliefs are. The big similarity here is that they didn’t tell the interviewer their age or political beliefs, that’s all. However, you can still view their response classification here if you wish.

When I saw that the fourth group was an error group, I thought about using three groups instead of four. When I looked at the output file for the analysis, it looked like a dozen people (ignoring the three errors) would be rearranged in the three group LCA. Since the selection criteria argues that people best belong in their group when using four groups, I decided to stick to four groups. Just please note that there’s nothing substantial in this group, it’s more of a containment group for errors.

Discussion

The preliminary analysis found that a wide variety of people enjoy video games, from young college students to parents to professionals. The analysis also found that it may be unfair to consider ‘gamers’ as a homogenous group as subgroups exist within the identity. ‘Older Gamers’, ‘Younger Gamers’ and ‘Family Gamers’ each have their own sociodemographics and gaming beliefs that make them both interesting and informative.

The ‘Older Gamers’ group is the only group to both contain more females and to not believe that most people who play games are male. I have considered two potential arguments for why this could be the case.

The first argument is that these women may have been exposed to the thriving arcade culture of the 1980s, with men and women alike rushing to play popular games such as Ms. Pacman (The History Chasers, 2015). The second argument may involve increased time to explore new hobbies and interests once children have left home. In the interview, participants are asked if they have ‘children under 18 living at home with them’. It is possible that their children have grown up and left home, leaving more free time to discover the fun of video games. Video games would be especially accessible to this population through mobile phones and tablets.

‘Younger Gamers’ also have their share of interesting findings. They are overwhelmingly male and have the most critical opinions for how female characters are portrayed in gaming. However, it is important not to mischaracterise this as ‘young women don’t play games’. We can see from the preliminary findings that slightly more women say they play video games than men, but young women seem to prefer not to call themselves a gamer. It is difficult to say exactly why, but two potential explanations include the stigma of negative stereotyping mentioned above, or the belief that they may not play enough video games to qualify as a ‘gamer’. Regarding the skepticism of female portrayals in video gaming, this appears to be a generational difference as similar patterns of skepticism have been found for female representation in movies (YouGov, 2018).

There may be a history lesson hidden inside the ‘Family Gamers’ group. This group would have been raised in or exposed to the time period of fervent debates surrounding the relationship between gaming and violence. This time period would see video game critics such as Jack Thompson argue that every school shooter is a gamer (Thompson, 2000) and that video games encourage children to enjoy killing (Benedetti, 2007). The idea that gaming is related to real-life violence seems all but extinct in younger generations, but is the majority opinion in this group.

It is interesting to note that while this group still considers themselves to be gamers, they seem to enjoy video games less than the other two groups; 0% of participants felt that most games are more entertaining than television. In the case of Jack Thompson-esque rhetoric about gaming damaging children’s brains, having young children yourself may make this belief more salient, while enjoying video games may minimise it (as seen in the ‘Older Gamer’ group).

Criticisms

All good research tends to pinpoint its flaws so that future research can improve upon it. In the interest of transparency, I would like to pinpoint some flaws below.

• This dataset involves an all-American sample, so American gamers may not necessarily represent gamers across the world.
• I wish I had more ‘gamers’ to work with than 165, but it is the nature of the data. If anything, it represents a neat starting point for looking at gamer subgroups.
• People may not agree with how ‘gamer’ is defined in this questionnaire. There is no real way to distinguish between a gamer who plays 30 minutes of Candy Crush on their phone every day versus someone who contributes a lot of time and money to the video game industry. However, I would justify this as looking at who exactly is happy to call themselves a ‘gamer’ rather than creating arbitrary gamer standards.
• Every percentage that I have reported is the valid percentage, meaning that I have not included the cases where people refused to answer a question. To my surprise, the questions that tended to receive the most refusal responses were questions regarding politics. This is interesting to me as people were happy to speak about their family, their income and share their ZIP code, but questions on politics seemed to cross some sort of line. Perhaps the politics variables would look different if so many people did not refuse to answer, but alas, I can only work with what I have.
• This data was collected in 2015 and predates the recent Battle Royale boom of the gaming industry. It would be interesting to see what this data looks like in 2018.

Summary

• Gaming, moreso than other hobbies, may be criticised for being a waste of time, something that promotes violent behaviour and something that only children should do.
• I set out to analyse who plays video games in their adulthood and who is happy to call themselves a ‘gamer’. I wished to know who these people were demographically and what their beliefs were regarding gaming.
• A wide range of factors are related to playing video games, including being a student, a parent, being employed and political ideology; there was no clear relationship with gender, highest level of education and type of job. People who play video games view the hobby positively, believing that it can be beneficial for cognition and teamwork. They do not believe gaming is related to violent behaviour.
• Several factors are related to identifying yourself as a gamer, such as being male, being younger and voting for the Democratic party. Gamers also view games positively, but do not have a clear pattern of beliefs for whether games portray female/minority characters poorly or whether most gamers are male.
• To break down this ‘gamer’ group further, I conducted a Latent Class Analysis (LCA) to identify subgroups of gamers based on their sociodemographics and gaming beliefs. The LCA produced four gamer subgroups, with the fourth group simply being a containment group for non-responses.
• The three valid groups can be thought of as ‘Older Gamers’, ‘Younger Gamers’ and ‘Family Gamers’. Older Gamers contain the most females, vote Democrat and don’t believe that most gamers are male or violent. ‘Younger Gamers’ are overwhelmingly male, overwhelmingly Democrat, highly believe in the benefits of gaming, and massively disagree with the idea that gaming is related to violence. ‘Family Gamers’ are the group most likely to be raising under-18s. The group contains slightly more males than females, is equally divided between Republican and Democrat, is dismissive of the idea that games portray women/minority characters poorly, and is the only group where believing gaming is related to violence is the majority opinion.
• This study shows that it may be unfair to dismiss people who play video games as childish, violent time-wasters; the hobby seems to attract people from many walks of life with different levels of responsibility. This is demonstrated well in the three gamer subgroups identified by the LCA: they are different people at different stages of their lives with different beliefs.

Thank you all very much for reading! This hard work would not be possible without the support of my wonderful Patrons. I would particularly like to thank my Platinum Patrons: Matt Demers, Albert S Calderon, Kyle T, Kyle Ryan, Andrew Shirvis, redKheld and DigitalPsyche. Thank you!

 

References

Akaike, H. (1987). Factor analysis and the AIC. Psychometrika, 52(3), 317-332.

Benedetti, W. (2007). Were video games to blame for massacre? Retrieved July 20th, 2018 from https://web.archive.org/web/20121107104123/http://www.msnbc.msn.com/id/18220228/.

Haagenars, J. A., & McCutcheon, A. L. (2002). Applied Latent Class Analysis. Cambridge: Cambridge University Press.

Lo, Y., Mendell, N. R., & Rubin, D. B. (2001). Testing the number of components in a normal mixture. Biometrika, 88(3), 767-778.

McCartney, J. (2008). Cited in Purchese, R. (2008). Games a “colossal waste of time”. Retrieved July 20th, 2018 from https://www.eurogamer.net/articles/games-a-colossal-waste-of-time.

McDonald, E. (2017). The Global Games Market Will Reach $108.9 Billion in 2017 With Mobile Taking 42%. Retrieved July 20th, 2018 from https://newzoo.com/insights/articles/the-global-games-market-will-reach-108-9-billion-in-2017-with-mobile-taking-42/.

Ramaswamy, V., Desarbo, W. S., Reibstein, D. J., & Robinson, W. T. (1993). An empirical pooling approach for estimating marketing mix elasticities with PIMS data. Marketing Science, 12(1), 103-124.

Schwartz, G. (1978). Estimating the dimensions of a model. The Annals of Statistics, 6(2), 461-464.

Sclove, S. L. (1987). Application of model-selection criteria to some problems in multivariate analysis. Psychometrika, 52(3), 333-343.

Simons, J. W. (2013). Should grown men (and Prince William) be allowed to play video games? Retrieved July 20th, July from https://www.telegraph.co.uk/men/thinking-man/10488777/Should-grown-men-and-Prince-William-be-allowed-to-play-video-games.html.

Statista (2018a). Value of the global video games market from 2011 to 2010 (in billion U.S. dollars). Retrieved July 20th, 2018 from https://www.statista.com/statistics/246888/value-of-the-global-video-game-market/.

Statista (2018b). Number of active video gamers worldwide from 2014 to 2021 (in millions). Retrieved July 20th, 2018 from https://www.statista.com/statistics/748044/number-video-gamers-world/.

The History Bandits. (2015). “Lady Arcaders” and Ms. Pac-Man’s Significance in Women’s Video Gaming. Retrieved July 20th, 2018 from https://thehistorybandits.com/2015/11/25/lady-arcaders-and-ms-pac-mans-significance-in-womens-video-gaming/.

Thompson, J. (2000). 20/20 Debate on Whether Violent Video Games Desensitize Children. World News Now, ABC.

Whittel, G. (2008). Cited in Purchese, R. (2008). Games a “colossal waste of time”. Retrieved July 20th, 2018 from https://www.eurogamer.net/articles/games-a-colossal-waste-of-time.

YouGov. (2018). Representation in movies. Retrieved July 20th, 2018 from https://d25d2506sfb94s.cloudfront.net/cumulus_uploads/document/mjpe6c0tgp/Representation%20in%20movies,%20February%2010%202018.pdf.

Zackariasson, P. & Wilson, T.L. (2012). The Video Game Industry: Formation, Present State, and Future. New York: Routledge.