Previous research examining the outcomes of free votes concludes that voting behaviour is determined in large part by MPs’ personal preferences. However, most studies do not measure preferences directly and ignore other possible determinants of voting behaviour. This piece illustrates the need to address these shortcomings before one concludes that preferences explain the outcomes of free votes. I illustrate this by examining a series of divisions on the issue of House of Lords reform. Using direct measures of preferences and controlling for alternative explanations, the analysis suggests MPs’ preferences had little effect on voting behaviour on this issue.