Priming and Third Person Effects of 4chan.org

Taught by S. Shyam Sundar, my 400-level Media Effects course this spring was a survey of media effects theory that transitioned into analysis of online implications. The course culminated in a collaborative social sciences research project; our group attempted to measure the effects of race-related imagery on 4chan with a standardized implicit association test.

An excerpt of my contribution to the Results and Limitations portion of our report:

Evidence of the Third Person Effect matched the predictions of the social distance corollary nicely, and the established theories of related semantic cognitive networks were supported with our priming data and the results of the IAT.

Because the findings indicated a difference in race attitudes when exposed to racist content–though barely below the statistic significance threshold–users and hosts of such shock humor should note that even offensive content posted in fun or without serious context can have real-world effects on viewers.  Our tests for Third Person Effects show that though one doesn’t think they are affected by racist imagery, they believe others are.  Together, these arguments provide evidence that, if one wants to avoid cultivating racist mentalities, humorous racist content is not insignificant.

Conversely, because our results did not reach statistical significance, hosts and providers of shock content may still argue that there is no conclusive evidence that content has confidently measurable effects.  Those tempted to jump to the conclusion that such content always has negative net effects should perhaps be more conservative and acknowledge that racist content may indeed be demonstrated as largely harmless.

A number of theoretical limitations remain as caveats to our findings.  First and foremost, practical limits on the race percentages of our subjects and limits on our ability to vary race in testing materials kept our findings from being further generalized; our experiment used only African-American versus White American preference Implicit Association indicators, ignoring possible effects of other races, and we did not have enough non-white subjects to attempt to find significance for any other population.  Even if we had had enough black subjects to mirror our conclusions about one’s preference for one’s own race through the IAT, our results largely ignore the research question as it applies to people of other races.

There were also validity limitations in expanding our findings to other instances of the shock site or forum.  We tested only one message board on one site, the Random page on 4chan, and our results have not yet been replicated on other sites to determine if the effect is universal.  Even our stimuli were not identical  to the normal 4chan forums—vivid pornography advertisements, internal links to other 4chan pages that include illegal material, and some racist text comments were removed to keep our study within the limits of our institution’s Internal Review Board for research and may have had a lessening or otherwise unanticipated effect on subjects’ responses had they remained.

Because of technological limitations, our survey relied on subjects accurately self-reporting their IAT results, leaving us unable to detect uneven bias in self-reporting if it existed.  The IAT, while operationally effective and generally representative of the effects we were examining, is itself an incomplete measure of the relatively broad concept of racism.  Our experimental design also did not measure tests over any significant length of time, leaving open the possibility that increases in racist attributes decayed rather quickly after exposure to the stimulus.

Further, we had no measure of determining the existence or level of behavioral change in subjects between conditions.  Although exhibiting increased cognitive association with negative attributes when exposed to members of a different race is an important effect from a larger societal perspective in its own right, being able to determine the actual effects of racism is also significant.