Timing of puberty has both short- and long-term impacts on physical and mental health for adolescent girls; understanding links between early puberty and health outcomes has important implications for healthcare and policy. This research concerns existing evidence for associations between early puberty in girls and risky behaviors. Such behaviors include risky sexual behaviors (including earlier sexual experiences, unprotected sex and teenage pregnancy) and risky health behaviors (such as alcohol and substance misuse). Although there have been numerous observational reports of links between early puberty and risky behaviors, there have also been inconsistencies. For example, some studies report associations to be persistent into adulthood, whereas others suggest that effects are only present in early, but not late, adolescence. Furthermore, observational studies have numerous limitations, including confounding, reverse causality and bias, which make findings difficult to interpret. Thus, it is currently unclear whether associations between early puberty and risky behaviors in girls may be causal. Although randomized controlled trials (RCTs) are considered the gold standard of epidemiological research methods, testing causal associations between early puberty and risky behaviors is not possible within an RCT design. Alternatively, Mendelian randomization allows the testing of causal relationships between exposure and outcome within a genetically informed design. We propose to test the causal relationship between early puberty in girls and risky behavior using a Mendelian randomization design.