Meta-analytic methods were used to examine ADHD-related risk-taking in children and adults, and to compare the magnitude of risk taking across behavioral, self-report, and virtual reality metrics. Potential moderators of effect size heterogeneity were also examined via a three-level multi-level approach and a hybrid meta-analytic/systematic review approach. Aggregated effect sizes obtained from 56 behavioral-task studies (82 effect sizes; ADHD_N = 2577; TD_N = 2606), 51 self-report studies (130 effect sizes; ADHD_N = 18,641; TD_N = 113,163), and 8 virtual reality studies (16 effect sizes; ADHD_N = 382; TD_N = 436) suggest that children and adults with ADHD exhibit moderately more risk-taking compared to children and adults without the disorder. Notably, the aggregated effect size obtained from virtual reality simulations (Hedges', g = 0.43) was 30–40% larger than effect sizes obtained from self-report and behavioral task metrics (Hedges' g = 0.31 and 0.27), respectively. Suboptimal Decision Making was the only significant moderator identified via multi-level modeling; however, comparison of subgroup effect sizes revealed potential moderating effects of ADHD presentation and trial-by-trial feedback on behavioral tasks. Collectively, findings suggest that ADHD is reliably associated with small to moderate magnitude greater risk-taking behavior and virtual reality simulations appear be the most sensitive currently available metric.

元分析方法用于检查儿童和成人与 ADHD 相关的冒险行为，并比较行为、自我报告和虚拟现实指标的冒险程度。还通过三级多级方法和混合元分析/系统审查方法检查了影响大小异质性的潜在调节因素。从 56 项行为任务研究（82 项效应量；ADHD _N  = 2577；TD _N  = 2606）、51 项自我报告研究（130 项效应量；ADHD _N  = 18,641；TD _N  = 113,163）和 8 项虚拟研究中获得的汇总效应量现实研究（16 个效应量；ADHD _N  = 382；TD _N = 436) 表明与没有这种障碍的儿童和成人相比，患有 ADHD 的儿童和成人表现出适度的冒险行为。值得注意的是，从虚拟现实模拟 (Hedges', g  = 0.43)获得的总效应量比从自我报告和行为任务指标 (Hedges' g  = 0.31 和 0.27)获得的效应量大 30–40% 。次优决策是通过多级建模确定的唯一重要调节器；然而，亚组效应大小的比较揭示了 ADHD 表现和逐次试验反馈对行为任务的潜在调节作用。总的来说，研究结果表明 ADHD 与小到中等程度的更大冒险行为可靠相关，虚拟现实模拟似乎是目前最敏感的可用指标。