The methodology for studies designed to investigate potential risk factors for sports injury is reviewed, using the case of hamstring strains as an example. Injuries result from a complex interaction of multiple risk factors and events. Therefore, a multivariate statistical approach should be used. In addition, the sample size of the study needs to be considered carefully. Sample size mainly depends on the expected effect of the risk factor on injury risk, and to detect moderate to strong associations 20–50 injury cases are needed, whereas small to moderate associations would need about 200 injured subjects. Studies published to date on the risk factors for hamstring strains have methodological limitations, and are too small to detect small to moderate associations.