Stuart Schechter and Cristian Bravo-Lillo
15 July 2014
We surveyed 3,539 workers on Amazon's Mechanical Turk to gauge their response to five scenarios describing scientific experiments---including one scenario based on Facebook's emotional contagion experiment. Respondents who reported being already aware of Facebook's experiment responded very differently to the scenario based on it than those who reported being unaware, so we focused on 2,102 respondents who reported being unaware. We asked these respondents whether they would want someone they cared about to be included as a participant, interpreting an answer of `no' as indicating concern for participants. A greater fraction of respondents were concerned about the two of the four scenarios inspired by university-approved experiments than expressed concern for Facebook's experiment. We also asked whether the experiment should be allowed to proceed, interpreting a `no' answer as disapproval of the experiment. A similar or greater fraction of respondents disapproved of the two more controversial scenarios based on university-approved studies as disapproved of the Facebook-experiment scenario. We found a statistically significant reduction (for $\alpha=0.05$) in disapproval and concern for participants in a group of respondents shown a hypothetical variant of Facebook's experiment in which the manipulation performed by researchers was to insert extra positive posts into users' news feeds---instead of removing positive or negative posts based on treatment group.