Final Post – Results

The following findings were derived from the experimental design outlined in a previous post, entitled “A dataset of my very own,” in which participants answered survey questions and uploaded screenshots of posts from their own Facebook newsfeeds.

Of the 30 participants who completed the survey/screenshot activity, 14 were assigned to the political treatment group and 16 to the sports treatment group. All were timed from the moment they advanced to the page with the screenshot activity to when they clicked on the next button at the end of the page. Participants in the political condition took 561 seconds on average to complete the Facebook activity, while participants in the sports condition took an average of 436 seconds (356 seconds if one outlier is excluded). While the sample size is too small for statistical significance, this suggests that sports posts are more common in the Facebook newsfeeds than political posts.

The partisanship distribution in the sample was skewed heavily to the left. 20 of the 30 participants reported identifying more closely with the Democratic party, and the remaining 10 were split evenly between independents and Republican-leaning individuals (only 1 identified as a strong Republican). This skew, combined with the small sample size, made it difficult to identify relationships between partisanship and other key variables, such as time.

As for the 39 screenshots uploaded by the political treatment group, 29 had partisan cues. Of these posts, 12 were consistent with the participant’s partisan identity, 12 ran counter to it, and 5 were uploaded by independents. This result suggests high exposure to heterogeneous opinions on Facebook, but it should be noted again that this is drawn from a very small sample size.

Finally, the main hypothesis, that the political treatment would increase polarization, was not supported. The polarization variable was operationalized several ways based on responses to the survey questions following the activity, but all yielded similar results between the political and sports treatment groups.