Parties, agendas, and roll rates

For the US House of Representatives, Cox and McCubbins discover tiny majority-party roll rates and offer them as evidence of majority-party agenda control. However, the observed roll rates are approximately what would result from chance alone or from chance constrained in several natural ways. Besides that, we show that rolls themselves are not evidence of any lapse in partisan agenda control and may even occur as the intended consequence of agenda setting by the majority party. Innovations include a solution to the combinatorial problem of counting all possible rolls, the associated computations, hypothetical examples of strategically advantageous self-induced rolls, and a review of likely real examples of the same.