In this study, we conducted a cross-sectional experiment of NIH peer reviewers’ evaluations of mock primary reviewers’ comments (overall impact statements (OISs)) in which the level and sources of risks were manipulated. Manipulated sources of risk were the PI or the approach and manipulated levels of risk were low or moderate, resulting in four OISs: low risk PI and low risk approach (control OIS); low risk PI and moderate risk approach (approach risk); moderate risk PI and low risk approach (PI risk); and moderate risk PI and moderate risk approach (PI-approach risk). We evaluated the association of proposal risks and reviewer characteristics, including risk tolerance, with reviewers’ criteria and overall scores. A list of all included Variables Demographics gender_recoded race_ethnic_non_white eng_1lang degree_PhD degree_MD years_since_degree Review Experience NIH_panels NSF_panels AHRQ_panels DoD_panels other_panels TotalPanels Scoring of Control Statement ctl_overall ctl_sig ctl_inn ctl_PI ctl_app ctl_env Scoring of manipulated statement comp_overall comp_sig comp_inn comp_PI comp_app comp_env Expertise Level res_similar Evaluative pre-disposition grt_predispo Risk Tolerance Proxy NEO_Openness Version of manipulated statement comp_PI_gender comp_app_risk comp_PI_risk Ctrl1st_Yes_1