And... This was the third and final part of my dissertation, which allowed me to successfully defend my PhD last week! 🎉

The raw data, formatted data, codebook, formatting code, analysis code, materials, and documentations are openly accessible on OSF. There’s also a very thorough set of extra analyses that are reported as “supplementary materials” that address matters of validity, robustness, and generalizability.

We found the same pattern in both studies (with real articles in study 1, with standardized articles in study 2).

The more participants were habitual VVG players, the more they “distanced” their belief from the claim “VVGs increase aggression”, when exposed to the claim.

And with heavy players (on the right side of the x-axis on the previous post), the finding was quite interesting. Before reading, they believed in a null effect of VVGs.
Then after reading “VVGs increase aggression”, they started believing in a negative effect, i.e., that “VVGs decrease aggression”.

However, the effect diminished the more participants were VVG players. For moderately habitual players: before reading, they believed in a null effect, and after reading “VVGs increased aggression”, they did not update their beliefs and continued to believe in a null effect.

(middle of the x-axis)

After reading, they updated their belief toward what they read in both conditions:
Those who read that “VVGs increase aggression”, believed in a stronger effect.
Those who read that “VVGs have no effect”, believed in a null effect.

(still all the way to the left of the x-axis)

Main result:
Before reading, non-gamers and non-habitual VVG players believed in a small positive effect of VVGs on aggressiveness.

(all the way to the left of the x-axis)

For each person, a slope would link VVG exposure to Δaggressiveness.
The slope is the belief.
Each participant gets a slope.

Positive slopes = belief that VVGs increase aggression.
Flat slopes = belief that VVGs have no effect on aggression.
Negative slopes = belief that VVGs decrease aggression.

Belief measure:
we used 6 vignette scenarios that involve “a typical video game player” (with randomized levels of pre-play aggressiveness, duration of play, violence level during gaming sessions).

Participants had to estimate the post-play aggressiveness.

In Study 1, we used real study abstracts (published meta-analyses and published experiments).
In Study 2, we used standardized fictional abstracts (same fictional authors, same methods, varying conclusions), and we also used a press-style lay article without technical jargon.

We conducted two online experiments (each N = 788). We measured participant belief before and after reading research summaries that say either “VVGs increase aggression” or “no effect”. And we measure long-term VVG exposure.

We asked: “Do people update their beliefs towards psychological evidence—or sometimes away from it—when the evidence threatens them?” We focused on the case of VVG players (the threatening claim for them would be “VVGs increase aggression”).