User study data - Summaries with personalized persuasive suggestions to mitigate confirmation bias during interaction with online debates

Description

This data was collected to test the effect of debate summaries and personalized persuasive suggestions to engage with them on participants argument recall after engaging with the debate. It contains interaction data and questionnaire results of 212 participants who interacted with one out of four versions of an online debate page.

Variables

(names/column headers, description, coding)

display_con: debate display condition, coding: 1: without summary, 2: with summary and neutral suggestion, 3: with summary and personalized persuasive suggestion, 4: with summary and random persuasive suggestion

correct_comp: proportion of correctly recalled arguments (10 arguments)

AO_correct_comp: proportion of correctly recalled attitude-opposing arguments (5 arguments)

AC_correct_comp: proportion of correctly recalled attitude-confirming arguments (5 arguments), coding

assigned_topic: debate topic participant was assigned to

clicked_contribute: indicates whether participant made a contribution to the debate, binary

att_strength: strength of prior attitude, coding: 3: strong, 2: moderate

time_debate: time spent on the debate page in seconds

clicked_showmore: indicates whether participant clicked on the show more button to reveal two additional items of the summary, binary

att_change: change of prior to post attitude, coding: negative values indicate a weakaning, positive a strengthening of the initial attitude (attitude was measured on a seven-point Likert scale)

stps_highest: highest scoring persuasion category (persuasion profile)