## **Project Overview** ## **This page contains materials associated with:** Meidenbauer, K. L., Choe, K. W., Bakkour, A., Inzlicht, M., Meidenbauer, M. L., & Berman, M. G. (in press). Characterizing the role of impulsivity in costly, reactive aggression using a novel paradigm. *Behavior Research Methods*. Preprint DOI: https://doi.org/10.31234/osf.io/kw3by ## **Project Components** ## ## Data ## ***Data from Computer Opponent Version - Confirmatory Study:*** - **"RC-RAGE_3pt_confirmatory_data.csv"** is a csv with data from large confirmatory study (N = 354) reported in paper and that is the input for all analysis scripts of the Computer Opponent Version ***Data from Human Opponent Version*** - **"RC-RAGE_Human_Opponent_3pt_scored_anonymized_data.csv"** is a csv with data from the full sample of participants who completed the Human Opponent Version of the RC-RAGE (N = 366). This csv is used for all Human Opponent Version analyses reported in paper. The label "Suspicion_YES" specifies participants who suspected deception. Analyses on these data are done on either the full sample (N = 366) or those who did not detect deception ("Non-suspcious sample"; N = 281). ***Pilot Data - Computer Opponent*** - **/ Preliminary_Data_0-pt Version:** *"RC_RAGE_prelimResults.csv"* is a csv with pilot data from a previous version of the task in which retaliation was always costly (participants got 0 points back for retaliating) - this task led to relatively low rates of retaliation and was changed to the current (3-point version). - **/ Preliminary Data 3-pt Version:** *"RC-RAGE_3pt_prelimData.csv"* is the pilot data for the current version of the task. This data was used to generate the pre-registration for the confirmatory study. ## Analysis Code ## All analysis scripts are in the form of Rmarkdown files. Output from each of these scripts is in the Results folder as a .html file of the same name. ***Analysis Scripts*** ***/ Computer Opponent Version*** - **"Self-Report_RetalationRate_Correlations.Rmd"** includes correlation analysis and z-tests comparing correlation coefficients (results from sections "Confirmatory Correlation Results" and "Comparing the Relationships between Impulsivity/Self-Control and Costly Aggression" in manuscript) - **"LogMixedEffects_bySelfReport_Part1_BPAQ.Rmd"** includes the logistic mixed-effects models predicting retaliation rate by costliness of retaliation and BPAQ measures of trait aggression. **"LogMixedEffects_bySelfReport_Part2_BIS_BSC.Rmd"** contains the same analysis pipeline for the BIS impulsivity measures and brief self-control. **"LogMixedEffects_bySelfReport_Part3_Affect.Rmd"** has the same analysis for hostile and irritable affect. These analyses are reflected in the "Predicting Costliness of Retaliation by Self-report Measure" section of the manuscript and Table 3: Logistic Mixed Effects Regression Tables with SignificantInteractions. - **"LogMixedEffects_bySelfReport_AllSig_wSocDesirability.Rmd"** has the code for all of the logistic mixed effects regressions with an added term taking into account social desirability score. These results are referenced in "Predicting Costliness of Retaliation by Self-report Measure" subsection "Effects of Social Desirability" ***/ Code for Figures*** - **"CorrelationMatrixPlot.R"** creates the correlogram for the Computer Opponent Version in the supplementary materials (also CorrPlot.png in the Results section of this project) - **"PredPlots_withGroupedSelf-Report_Data.R"** generates the predicted probability plots in Figure 4 (also SR_PredPlots_wData.zip in Results section of this project). These plots show the fixed effects of the logistic regressions predicting retaliation rate by costliness of retaliation and self-report measure. In addition to the predicted probabilities, actual data from the sample (divided into 5 groups based on quintiles of self-report measures) are overlain on the plots to show the overlap between model predictions and real data. ***/ Human Opponent Version*** - ***/ Full Sample***: Analyses done on the full sample of participants (N = 366) in the Human Opponent Version study. - **"RC-RAGE_Human_Opponent_analysis_FullSample.Rmd"** includes code for descriptive statistics, table of retaliation rates, correlation analyses and z-tests for comparing correlation coefficients, generates the correlogram for the full sample (see Results/Human Opponent Version/Corrplot_alldata.png), as well as predicted plots + quintile data for the logistic regressions. - **"HumanOpp_LogMixedEffects_bySelfReport_FullSample.Rmd"** includes the logistic mixed-effects models predicting retaliation rate by costliness of retaliation and BPAQ subscales, BIS-Motor, and Hostile affect. - **HumanOpp_LogMixedEffects_bySelfReport_wSD_FullSample.Rmd** includes the logistic mixed-effects models predicting retaliation rate by costliness of retaliation and BPAQ subscales, BIS-Motor, and Hostile affect with social desirability as a covariate. - ***/ Non-Suspicious Participants***: Analyses done on those participants (N = 281) who did not suspect deception in the Human Opponent Version study. - **"RC-RAGE_Human_Opponent_analysis_NonSusSubset.Rmd"** includes code for descriptive statistics, correlation analyses and z-tests for comparing correlation coefficients, generates the correlogram for the non-suspicious subset (see Results/Human Opponent Version/Corrplot_notsus_subset.png), as well as predicted plots + quintile data for the logistic regressions - **"HumanOpp_LogMixedEffects_bySelfReport_Nonsus_Subset.Rmd"** includes the logistic mixed-effects models predicting retaliation rate by costliness of retaliation andPAQ subscales, BIS-Motor, and Hostile affect - **"HumanOpp_LogMixedEffects_bySelfReport_wSD_Nonsus_Subset.Rmd"** includes the logistic mixed-effects models predicting retaliation rate by costliness of retaliation and BPAQ subscales, BIS-Motor, and Hostile affect with social desirability as a covariate ## Results ## Statistical results are html files and output of Rmarkdown file by same name. ***/ Computer Opponent Version*** - *Corrplot.png* contains the correlogram created by CorrelationMatrixPlot.R - *SR_PredPlots_wData.zip* is a zipped folder of all the individual plots (for each of BPAQ subscales, BIS subscales, BSC, and affect) shown in manuscript Figure 5 and in the supplementary materials. ***/ Human Opponent Version*** - ***/ Full Sample***: - *Corrplot_alldata.png* contains the correlogram created by "RC-RAGE_Human_Opponent_analysis_FullSample.Rmd" - ***/ Non-Suspicious Participants***: - *Corrplot_notsus_subset.png* contains the correlogram created by "RC-RAGE_Human_Opponent_analysis_NonSusSubset.Rmd" ***/ Test-retest data and results*** - Inlcudes a merged dataframe of the 188 participants who completed both sessions for test-retest reliability analysis and Rmarkdown output of the test-retest analysis. ## Preregistration ## Preregistration for the confirmatory study is linked to this project also available at this link: https://osf.io/czn43 ## Task Code ## The RC-RAGE task is **coded in jsPsych** and can be accessed at: https://github.com/SCENeLabWSU/RC-RAGE A **working demo** of the task is available at: https://kywch.github.io/RC-RAGE_jsPsych/rc-rage-demo.html