Four decades of research into the influences of time pressure on risky decisions have produced widely contrasting findings: 38.5% of the effects indicate that time pressure increases risk preferences, whereas 61.5% show the opposite. A theoretical framework with four conceptual categories of moderators is proposed to explain these heterogeneous findings: nature of the time constraint, negative outcome salience, negative outcome severity, and vulnerability to the outcomes. This framework is tested through a meta-analysis of 213 effect sizes reported in 83 papers, representing 65,574 unique respondents. The four categories of moderators effectively resolve notable conflicts. For example, regarding the nature of the time constraint, an absolute versus relative constraint increases risk preferences, but an ambiguous versus objective constraint decreases risk preferences. In terms of negative outcome salience, risk preferences decrease if the risk is learned about from a description (vs. experience) or the outcome is framed as a loss (vs. gain). Negative outcome severity also exerts an effect, as discrete choices lower risk preferences compared with attitudinal risk. In addition to managerial and public policy implications based on simulations, a comprehensive research agenda that builds on the robust insights of this meta-analysis is offered.