Data derived from more than one type of performance and that, when summarised statistically, misrepresent the performances from which they were derived (as when an avoidance schedule produces both a moderate response rate and bursts of high-rate responding after shock, so that the average rate does not represent either contribution to overall rate).