In legislative studies, a “roll rate” refers to the proportion of votes where a member or group of members is “rolled”—voting no on a measure that passes. Roll rates analysis is a potentially useful tool for evaluating arguments that the majority party in the House exercises influence by controlling what does and does not make it onto the agenda. Despite its promise, so far the use of roll rates to test these arguments has faced the problem of observational equivalence—that the predictions of partisan theories of agenda control are indistinguishable from those of arguments that have no place for parties. I address this problem by calculating party-less and partisan counterfactual roll rates data to pin down the predictions of these theories. This offers a powerful research design to evaluate observed roll rates from sample Congresses. The results provide intriguing evidence for partisan theories of the House agenda.