Traditionally, decision making under risk has been studied using scenarios with fixed levels of outcomes and probabilities (e.g., a 50% chance of winning a particular amount of money). In real-life, however, people can often actively influence the probability of a risky outcome through investing effort (e.g., working overtime to increase the likelihood of a raise). It is currently unclear what factors influence people’s effort investment to manage risk in their favor. We report two experiments in which participants were presented with monetary gambles and had the opportunity to engage in an essay-writing task to increase (decrease) the probability of a gain (loss) outcome of the gamble. Experiment 1 (N=102) shows that participants invested more effort (i.e., wrote more words) the larger the outcome and the lower the initial probability of the outcome. In addition, they tended to invest less effort to decrease the probability of a loss than to increase the probability of a gain—suggesting that effort investment is driven by risk attitudes. Based on a computational model, we find that people’s effort decisions are characterized by diminishing sensitivity to outcomes and probabilities and that they displayed lower probability sensitivity when managing losses (vs. gains). Experiment 2 (N=100) showed that although participants do not exert the effort required to maximize their chances of a favorable outcome, their effort investment increases when the pay rate was higher. Our results identify behavioral regularities in how people trade-off internal resources against risk and point to a theoretical framework to capture these regularities.