Did COVID-19 have any impact on why people play video games? Beyond the shared, global anxiety of watching the pandemic unfold, many countries put in place similar shelter-in-place orders in roughly the same timeframe, disrupting work, leisure, and social interactions. Certainly, being kept indoors and having fewer sources of entertainment has boosted spending and engagement with video games, but we wondered if this global event had any impact (even temporarily) on what people want out of video games?

One plausible shift is that gamers may want more social interaction in gaming as face-to-face interactions are reduced. And might the barrage of worrying COVID-19 news bites on TV and social media lead gamers to prefer more serene and relaxing gameplay? Or does the tedium of daily indoor routines lead gamers to want more thrilling, visceral gaming experiences?

Because we’ve been running the Gamer Motivation Profile continuously over the past few years, we have a unique data set to compare responses to the same set of questions before and after COVID-19.

The Gamer Motivation Profile

The Gamer Motivation Profile is a 5-minute survey that allows gamers to get a personalized report of their gaming motivations, and see how they compare with other gamers. Over 500,000 gamers worldwide have taken this survey. The 12 motivations that are measured in our model were identified via statistical analysis of how gaming motivations cluster together. You can get a more detailed description of our gamer sample here.

See how you compare with other gamers. Take a 5-minute survey and get your Gamer Motivation Profile

Our Pre/Post-COVID Analysis

In our analysis, we first selected a pre-COVID period for the baseline—the 15 months from 1/1/2019 through 3/31/2020. We then selected a post-COVID period for comparison—the 2.5 months from 4/1/2020 (when most countries had issued shelter-in-place orders) through 6/15/2020 (when we began this analysis).

We controlled for the impact of gender, age, casual/core/hardcore, and limited the analysis to US, Canada, Australia, and the UK.

Because the gender and age of our respondents shifts a little in different time periods, we also wanted to make sure any detected differences weren’t simply due to sampling shifts (since we know that there are some large motivational differences by gender and age). To this end, we controlled for the impact of gender, age, and casual/core/hardcore gamers. For the same reason, we limited the countries for analysis to the US, Canada, Australia, and the UK. This resulted in a sample size of 45,277 for the pre-COVID period, and 10,810 for the post-COVID period.

Statistically, it means to remove the effect of a potentially confounding (i.e., related) variable to better see the effect of the variable we’re actually interested in.

As a concrete example, we know that younger gamers are more competitive. So if our post-COVID sample just happens to have a slightly higher % of younger gamers, then a detected difference in Competition between the two time periods may simply be due to age rather than COVID. When we control for age, it means the statistical model removes the expected increase in Competition from the larger number of younger players. If we still detect an increase in Competition, then we know it’s not just because of the younger players and we’re more confident that it may be driven by COVID.

Completion and Community Changed The Least

The appeal of Completion (collecting stars/trophies, completing quests/tasks, collecting collectibles) and Community (chatting, teamwork, social interaction) changed the least between the two time periods. In fact, the differences were not statistically significant even with the large sample size we had. This suggests that social distancing hasn’t changed the degree of social interaction gamers are seeking out in video games.

Challenge and Excitement Changed The Most

The appeal of Challenge (high difficulty, challenging bosses/missions, practice & mastery) and Excitement (fast-paced action, thrilling, adrenaline rush) changed the most and both had downward trends. This suggests gamers as a whole are seeking out more relaxing, less challenging gaming experiences (even after age, gender, and casual/core/hardcore gamer types are controlled for).

Gamers are seeking out more relaxing, less challenging gaming experiences … but these are tiny changes.

While statistically significant, both were tiny changes—roughly a difference of 3.5 percentile points. In comparison the gender difference in Excitement spans 22.5 percentile points. And the effect of COVID explains only 4% of the variance that gender does, and 2.5% of the variance that age does. The pattern is very similar for Challenge. So these are detectable, but tiny, changes.

The p-value framework is geared towards the smaller sample sizes typically encountered in lab studies (< 500, and typically < 100). With increasing statistical power in large sample sizes (> 1,000), even tiny changes become statistically significant. While there are methods to adjust p-values to account for large sample size (and the p-values here except for Challenge and Excitement were significant even after adjusting for sample size and multiple comparisons–< .0002), we find it’s often more useful to look at the effect sizes in comparison with other benchmarks in the same domain.

In large sample sizes, statistical significance becomes largely meaningless. The meaningful question is: Is this difference something we should care about?

Some Thoughts and Caveats

There are several aspects of this analysis that bear pointing out:

COVID isn’t the only notable world event that happened in the past few months. We also had a market crash and BLM protests worldwide.

  • Correlation isn’t Causation. Because we can’t experimentally introduce COVID to one population and have another one as control, we’re limited to correlational analysis and we can’t definitively attribute any findings (or non-findings) to COVID. After all, COVID isn’t the only notable world event that happened in the past few months; we also had a market crash and BLM protests worldwide.
  • Short Timeline. Even as a significant global event, a 2.5-month time period may still be too short to sustain any meaningful impact. It may take longer than that for people’s gaming habits and preferences to change, if only due to inertia. And perhaps it’s also short in the sense that some states in the US only maintained shelter-in-place orders fairly briefly, which would also minimize any potential impacts.
  • Game Release Schedule. Major video games are typically released in the fall, to coincide with the holiday season, to build up marketing from gaming conferences in the summer (e.g., E3), and when more people are staying indoors to keep warm. Assuming that it’s easier to shift to different gaming motivations when trying out new games (rather than games you already own), then the first 3 months of the year offer far fewer new releases to try.
  • Internal Reversals. Because we focused on the top-level differences, there is the potential that there are internal reversals that are being concealed—e.g., older women became much more competitive but younger women became much less competitive, and this cancels each other out at the top level. However, when we looked at the underlying two-way interactions, we didn’t see any substantive effects along these lines.

Did COVID Change Your Gaming Pattern/Preferences?

What about you? Did you have more time to spend on gaming during the past few months? And did you find yourself preferring games or game mechanics different from the ones you usually enjoy? Tell us in the comments below.

Subscribe to our newsletter to stay up to date with our data-driven insights into video gamers and board gamers. When new findings are released, you’ll be the first to know.