Resistance training increases muscle size (i.e., causes hypertrophy) and muscle strength, particularly in untrained individuals. Hypertrophy is widely believed to be one of the mechanisms (i.e., a mediator) by which… Click to show full abstract
Resistance training increases muscle size (i.e., causes hypertrophy) and muscle strength, particularly in untrained individuals. Hypertrophy is widely believed to be one of the mechanisms (i.e., a mediator) by which resistance training increases strength. However, some researchers have questioned whether training-induced hypertrophy causes increases in strength. One approach to resolving this issue has been to use correlations between training-induced changes in muscle size and strength to infer the effect of hypertrophy on strength. This is problematic because correlations between changes in muscle size and strength may be confounded (i.e., correlation is not causation). Another approach has involved randomizing participants to different volumes of exercise to create different levels of hypertrophy and then comparing the strength increases associated with different levels of hypertrophy. This approach is also problematic because the unit of randomization is exercise volume rather than hypertrophy, and the potential for confounding remains. Thus, a new approach is needed to determine the extent to which hypertrophy increases muscle strength. Here, we introduce resistance training researchers to causal mediation analysis and recommend that it be used to resolve the current debate. Causal mediation analysis potentially provides an unconfounded estimate of the effect of a mediating variable (hypertrophy) on an outcome (strength). This analysis is supplemented by causal maps that help conceptualize research questions and identify potential confounders. In addition to resolving the debate on hypertrophy, causal mediation analysis can be used to answer a host of other questions about mechanisms in the health sciences.
               
Click one of the above tabs to view related content.