Date created: | Last Updated:

Category: Uncategorized

Description: Preregistration – Chat Study Sample Size We will recruit 2,000 individuals, aiming for at least 300 usable pairs, from the MTurk website. Participant Criteria Inclusion Criteria: • 18 years of age or older Exclusion Criteria: • None Measures Primary Dependent Variables: The extent to which the participants 1) believe they changed their partners’ opinions and 2) believe that their partners changed their own opinion. Secondary Dependent Variables: How highly participants rate the warmth and competence of their chat partner. Procedure Participants will be matched to have two-person, 15-minute long conversations with one another of a pre-selected topic. Partners will be matched based on having opposing views on the topic (e.g., one will support, the other oppose). In every dyad, one partner will be randomly assigned to receive instructions to speak in a politically correct or politically incorrect manner, and the other person will not receive instructions about how to speak during the conversation. After the conversation, participants will answer manipulation check questions. Then each participant will be asked to rate how much they believe they changed their partner’s opinion, how much their partner changed their own opinion, and how warm and competent their partner seemed. Finally, they will be surveyed on their demographic characteristics. Prediction Our primary prediction is that the uninstructed chat partners of the instructed politically correct (vs. politically incorrect) speakers will believe that they changed the opinions of their partners more. We do not think that politically correct speakers will actually report that their opinions changed more than the politically incorrect speakers. We also predict that politically incorrect partners will be rated as less warm than politically correct partners. Analysis Plan Our analysis will be a 2 (Instructed vs. Uninstructed) x 2 (Instructed Language: politically correct vs. politically incorrect) ANOVA on our primary and secondary dependent measures. We will run a follow-up analysis examining position as well: 2 (Instructed vs. Uninstructed) x 2 (Position: Support vs. Oppose) x 2 (Instructed Language: politically correct vs. politically incorrect) ANOVA. To control for pair-level variance, we will run two other analyses: 1. Regression analysis in which each experimental condition is included as a dummy variable (0 or 1) and each pair is also included as a dummy variable (e.g., Pair 1 labeled 0 or 1 and so on). 2. Linear mixed model analysis where the pair is included as the “group ID”. We will next split the file among those who were instructed and those who were uninstructed and separately test the effect of the language instructions on our primary measures within each of these two groups, controlling for position and pair number.