Cycling has the potential to address a number of personal and societal challenges, not least with respect to health and the need for more sustainable modes of transport. However, the best way(s) to promote cycling is still unclear. In an effort to answer this question, we identified 39 interventions designed to promote cycling, with a total sample of 46,102 participants. Random effects meta-analysis estimated a small but statistically significant effect of interventions on cycling behaviour (g₊ = 0.14, 95% CI [0.05, 0.23]). To identify the most effective intervention strategies, we coded the behaviour change techniques used within each of the interventions. Interventions that prompted people to self-monitor their behaviour or added objects to the environment (e.g., provided shared bikes) were more effective than those that did not use these strategies. Interventions that restructured the physical environment (e.g., built new cycle paths) were less effective than the studies that did not do this. We also identified a number of factors that moderated the effect of the interventions on outcomes; specifically, interventions that targeted a specific group, used objective measures of cycling such as accelerometers, and that were tested using independent groups designs typically yielded stronger effects. The findings should help to guide interventions to promote cycling in the future.

骑自行车有可能解决许多个人和社会挑战，尤其是在健康和对更可持续的交通方式的需求方面。然而，促进骑自行车的最佳方式仍不清楚。为了回答这个问题，我们确定了 39 项旨在促进自行车运动的干预措施，样本总数为 46,102 名参与者。随机效应荟萃分析估计干预措施对骑行行为的影响很小但具有统计学意义（g ₊= 0.14, 95% CI [0.05, 0.23])。为了确定最有效的干预策略，我们对每个干预中使用的行为改变技术进行了编码。促使人们自我监控自己的行为或向环境中添加物品（例如，提供共享单车）的干预措施比不使用这些策略的干预措施更有效。重组物理环境的干预措施（例如，建造新的自行车道）不如没有这样做的研究有效。我们还确定了一些影响干预对结果影响的因素；具体而言，针对特定群体的干预措施，使用加速度计等客观的自行车测量方法，并使用独立的群体设计进行测试，通常会产生更强的效果。