Are 30% of African Gamers Addicted? Exploring the Data

Very important update: On June 4th 2020, the research paper this article is based on was retracted by the publisher due to internal inconsistencies and “mathematically impossible” statistical analyses. While I am grateful that the publication is now removed, I am deeply concerned that a paper with so many errors and impossible statistics passed the peer review process. I am additionally concerned that this will lead to more bad faith towards gaming researchers. I know of at least two researchers ([1] [2]) who submitted publications criticising this paper, and an incredibly thorough dismantling of the dodgy statistics in this research paper can be read here. I can assure you academics were not silent about this paper, and these integrity-driven figures deserve your support.

Two months ago, a research paper was published stating that 30% of gamers sampled across Africa were addicted to video games. At this point, Gaming Disorder and Internet Gaming Disorder are familiar topics on my website. I have covered the topic of gaming addiction since 2018, so I wanted to explore this frankly unsettling claim further.

The paper can be read for free online, but I appreciate that academic research papers are not the easiest reads in the world. I will be explaining the findings of this paper in language that is easy to understand, while also offering my own perspective as a researcher and as someone who has read video game addiction research for years.

Before I begin, I need to make two things crystal clear:

1. Scoring highly on a video game addiction questionnaire doesn’t mean that you have Gaming Disorder. Gaming Disorder can only be diagnosed by a mental health practitioner, and they will be reluctant to diagnose you unless you experience impairment to daily life (e.g. lost a job and a relationship due to compulsive gaming).
2. Gaming Disorder is anticipated to represent a very small proportion of the gaming population (World Health Organisation, 2018). It’s not uncommon for this figure to be cited at around 1% (Gentile et al., 2017).

Baring these two things in mind, let’s begin. As usual, there will be a summary at the bottom if you do not wish to read everything.

Contents

1. Abstract
2. Why the Study was Conducted
3. How the Study was Conducted
4. Study Results
5. Discussion
6. Summary
7. References

Abstract

I’ve never specifically talked about an abstract before in an article, but this abstract is a special case. An abstract is a summary of the research and its findings written by the authors of the paper. Researchers tend to be pretty busy people, so it’s common for them just to read a study’s abstract to keep up to date with research in the field.

There are two things in this abstract that I feel I either need to explain or signpost for later.

Firstly, the abstract contains lines such as:

Results showed our sample of gamers (24 ± 2.8 yrs; 88.64% Male), 30% were addicted…

 

– Sosso et al., 2020

As mentioned in my crystal clear statements, scoring highly on a questionnaire does not mean that you are addicted to video games – you need to be diagnosed by a mental health practitioner. The video game addiction questionnaire used in this study, the Game Addiction Scale Short Form (Lemmens et al., 2009), places gamers into one of four categories: addicted gamers, problematic gamers, engaged gamers, and non-problematic gamers. Addicted gamers are gamers who score highly on the questionnaire, but we can’t actually say they’re addicted because they haven’t been diagnosed. I’ll be exploring how a person is classified as an addicted gamer soon.

The second point is where things get a little curious. To put this into basic language, the researchers state in the abstract that approximately 87% of insomnia, 83% of daytime sleepiness, and 82% of anxiety is due to participants playing video games. Remember this, it will be important later.

Why the Study was Conducted

The introduction/justification for the study is fairly strong. The researchers acknowledge previous research showing that rates of video game addiction can differ by population and country, and Africa has typically been unrepresented in this research. They argue that constantly focusing on video games, particularly on a bright screen, can influence mental health and sleep behaviour. The latter is under-researched in Africa, so they wanted to contribute to the research base.

Apologies for constantly saying ‘remember this’, but there are a few key details from the introduction that are important for later:

• The researchers acknowledge that gaming addiction is most likely to be experienced by adolescents and young adults.
• They also acknowledge that those suffering from video game addiction lead an impaired life as they ‘[dis]engage [from] the offline world’.
• They cite previous research indicating that smartphone gaming ‘does not appear to be problematic’. Gaming addiction is more likely to be facilitated by computers and consoles, particularly within genres such as MMOs.

How the Study was Conducted

Participants were recruited by distributing email questionnaires to nine universities in Africa, leading to a sample of 10,566 participants (89% male). They were asked to fill in questionnaires regarding insomnia, daytime sleepiness, anxiety and depression, video game addiction, how often they played video games, and demographic information such as employment and marital status.

So how exactly were people classified as addicted to video games? The gaming addiction questionnaire used, the Game Addiction Scale Short Form, contains seven questions that are associated with the diagnostic criteria for Gaming Disorder. However, the current study only uses four of these questions.

As someone familiar with the history and debate surrounding Gaming Disorder, I can understand this. These four questions are thought to represent the most sinister and dysfunctional aspects of a video game addiction, while questions such as asking if a person thinks about video games during work aren’t all that sinister. The study doesn’t show what these questions are, so I had to source them from another paper (Khazaal et al., 2016). These questions are:

1. Have others unsuccessfully tried to reduce your time spent on games? (Representing relapse)
2. Have you felt upset when you were unable to play? (Representing withdrawal)
3. Have you had arguments with others (e.g. family, friends) over your time spent on games? (Representing conflict)
4. Have you neglected important activities (e.g. school, work, sports) to play games? (Representing problems)

So we know which questions are involved in classifying people as addicted to video games. Now let’s look at the scoring.

According to the study, you need to score three or more on these four questions to be called an addicted gamer. They don’t expand on three or more, so I had to dig up the original research paper for this questionnaire.

There are five possible answers to these questions: never (1), rarely (2), sometimes (3), often (4), and very often (5). The creators of this questionnaire experimented with two different classification types: classifying people as addicted if they score three or more, and if they scored four or more.

They found that using three or more rather than four or more led to a 671% increase in people being classified as addicted gamers. They also cite previous research finding that comparing these two approaches resulted in a 2067% increase in addicted gamers (from 1.8% to 39%; Charlton & Danforth, 2007).

The creators of the questionnaire go on to discourage the use of the three or more approach:

[This] format is likely to lead to over-estimation of the frequency of addicted gamers…For the vast majority of adolescent players, their game addiction scores merely reflect enthusiasm for videogames or a relatively harmless displacement from other activities.

 

– Lemmens et al., 2009

Participants were classified as addicted gamers if they said sometimes to the above questions rather than often or very often, a practice that is discouraged by the original creators of the questionnaire. So what was found regarding these gamers?

Study Results

Using the questionnaire above, 30% of gamers were classified as addicted, 30% were classified as problematic (an ‘at-risk’ group), 8% were engaged, and 32% were non-problematic. To quickly explain the engaged group, these are people who are actively engaged with video games (e.g. think about them a lot), but didn’t answer sometimes or more to the four questions above. In Africa, the most popular method of gaming was smartphone gaming.

Through a type of data analysis known as an ANOVA, they found that when comparing addicted or problematic gamers to the non-addicted groups, addicted/problematic gamers experienced more insomnia, daytime sleepiness, and more anxiety than non-addicted groups, but not depression.

They then use a type of data analysis known as a regression analysis to try to explain why participants experience insomnia, daytime sleepiness, anxiety and depression. Using gaming addiction, sex, age, education, income, marital status and employment status, the researchers were able to explain approximately 89% of insomnia, 84% of daytime sleepiness, 83% of anxiety, and 76% of depression.

Remember those statistics from the abstract? It’s time to bring them back. To quote them exactly from the abstract:

Gaming significantly contributed to 86.9% of the variance in insomnia, 82.7% of the variance in daytime sleepiness and 82.3% of the variance in anxiety.

 

– Sosso et al., 2020

Let’s take insomnia as an example. They say that gaming explains 86.9% of insomnia, yet the overall model explained approximately 89%. This would mean that while gaming explains 86.9% of insomnia, sex, age, education, income, marital status, and employment status combined only explain around 2.1%

This doesn’t seem right to me.

Let’s take a look at the standardised beta values (β) of the model. We look at the standardised beta value as an indicator of how much a variable contributes to explaining a behaviour. The standardised beta values for gaming addiction types and insomnia are 0.23, 0.72, and 1.8. The next largest standardised beta value is the 24-30 age group at 0.752.

So 0.23, 0.72, and 1.8 together explain 86.9% of insomnia, yet 0.752 barely explains around 2%?

You may be interested to know that it is very, very easy in a regression analysis to see how much a variable contributes to a model in percentage form. For example, I could have a model to explain ice cream sales in a day that includes variables such as season and location (e.g. a park). I could see that by adding daily temperature, this explains an additional 30% of ice cream sales.

Despite how easy it is to get this information, it’s not disclosed in this paper.

How do I know how easy it is? I reported this information myself in a regression analysis I did for my website. This is their job and this is a peer-reviewed publication, yet they couldn’t provide this basic information for the sake of transparency.

Discussion

This is where I switch from a fatigued researcher to a prosecutor as I’m about to poke some holes in their logic.

According to the researchers, their findings suggest that attachments to the offline world such as having a romantic partner and having a job reduce the likelihood of being addicted to video games. Except their findings don’t seem to suggest this.

Compared to 3.52% of non-problematic gamers, 67% of addicted gamers both study at university and have a job. In fact, 83.5% of participants who both study and work were classified as addicted gamers. The researchers argue that having a busy offline life reduces the likelihood of gaming addiction, yet 83.5% of those with busy offline lives were labelled addicted gamers. This doesn’t add up.

Now is a good time to return to the points made at the beginning of the study to compare them to the study’s findings.

1. Gaming addiction is most likely to be identified in adolescents and young adults. In this study, the most common age group for addicted gamers (24-30) is older than the average age in previous research.
2. Those addicted to video games disengage with the offline world in favour of playing video games. In this study, 83.5% of those with busy offline lives were labelled as addicted gamers.
3. Those addicted to video games favour console and PC gaming, with smartphone gaming being labelled as ‘non-problematic’. Smartphone gaming was universally the most popular method of gaming in the current study.

All of these things just don’t add up in the context of years of research. My educated guess is that by classifying participants as addicted using criteria that the original authors recommended against, they have over-classified participants as addicted gamers. This limits the validity of these findings, so I would take them with a pinch of salt.

Something that I couldn’t help but notice is the fact that this is the second-most typo-filled research paper I’ve read in my academic career. There are words missing from sentences, incorrect statistics, formatting mistakes on tables, and missing references. This paper includes sixteen authors and would have been read by at least two peer reviewers before being accepted for publication. With so many mistakes going under the radar with so many pairs of eyes on this manuscript, part of me isn’t surprised that curiously omitted statistics and holes in logic weren’t pointed out.

I’m going to get personal here, which is something I very rarely do in my articles. I understand that working in the social sciences comes with a degree of original sin. I’ve received my fair share of hatred for being some sort of dishonest snake woman playing Pretend Science, and this hatred gets multiplied any time the topic of gaming addiction is mentioned. I try my utmost to be as honest, transparent, and respectful of data science as I possibly can when living my life as a researcher.

Research papers like this not only make it hard for me to have faith that others are doing the same, but also contribute to the original sin that hard-working, caring and conscientious researchers fight to shake off.

Summary

• This study was conducted to address the lack of video game addiction research conducted in Africa.
• The study recruited 10,566 participants (89% male) from nine universities in Africa. Participants were asked to fill in questionnaires on insomnia, daytime sleepiness, anxiety and depression, video game addiction, how often they played video games, and demographic information such as employment and marital status.
• Participants were classified as addicted gamers if they answered sometimes or higher on four video game addiction questions. The original creators of the video game addiction questionnaire warn against this sometimes or higher approach, finding that it can inflate cases of video game addiction by between 671-2067%.
• Using this classification method, 30% of African gamers were identified as addicted gamers, and another 30% were identified as at-risk of video game addiction (problematic gamers). Smartphone gaming was the most popular type of gaming in Africa.
• The researchers state that approximately 87% of insomnia, 83% of daytime sleepiness, and 82% of anxiety can be attributed to people playing video games. However, the statistics behind this don’t seem to match up with other statistics reported in the study. The statistics that would eliminate this doubt are omitted from the study, despite how easy they are to gather and report.
• The researchers argue that busy offline lives can protect people against gaming addiction, but this is contradicted by their own finding that 84% of those with busy lives (i.e. those who both study and work) were classified as addicted gamers.
• The profile of addicted gamer identified in this study (busy adults using smartphones) doesn’t match up with years of video game addiction research (isolated young people who separate themselves from the offline world via games such as MMOs). This may be due to the over-classification of participants as addicted gamers by using guidelines that were discouraged by the original creators of the questionnaire.

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: Albert S Calderon, Kyle T, redKheld, Dimelo ‘Derp’ Waterson, Hagbard Celine, Aprou, Austin Enright, SK120, NotGac, Shaemus, Joey Rodriguez, Marcus Lo Re-Sant, DarrenIndeed, Thomas Meszaros, Ciara Elizabeth, Dr. Jhin, Mulgar, Tobias Svensson, and Matt Demers. Thank you!

 

References

Charlton J. P., & Danforth I. D. W. (2007). Distinguishing addiction and high engagement in the context of online game playing. Computers in Human Behavior, 23, 1531–1548.

Gentile, D. A., Bailey, K., Bavelier, D., Brockmyer, J. F., Cash, H., Coyne, S. M., … & Markle, T. (2017). Internet gaming disorder in children and adolescents. Pediatrics, 140(Supplement 2), S81-S85.

Khazaal, Y., Chatton, A., Rothen, S., Achab, S., Thorens, G., Zullino, D., & Gmel, G. (2016). Psychometric properties of the 7-item game addiction scale among French and German speaking adults. BMC Psychiatry, 16(1), 132.

Lemmens, J. S., Valkenburg, P. M., & Peter, J. (2009). Development and validation of a game addiction scale for adolescents. Media Psychology, 12(1), 77-95.

World Health Organisation. (2018). WHO: Revision of ICD-11 (mental health) – questions and answers (Q&A). YouTube. https://www.youtube.com/watch?v=cb6hsq-IHfA