The selection of wrong answers for multiple choice tests appears not to be blind guessing, as the current theory assumes, but represents systematic choice based upon the ways in which learners extract meaning and solve problems. It appears the departures from "expected" answers arise when the learner interprets the question in a manner different from that intended, uses a solution strategy differently than intended (perhaps incorrectly), or works from different information (perhaps misinformation) to solve problems. It would appear that the learner is unaware of these differences from expectation, and gives the best answer that he or she is able. The limiting factor confounding educators would seem to be inclinations to expound expectations rather than exploring those of the learners. Both teachers and learners may be unaware that their meaning extraction and thinking processes are not equivalent. To overcome this problem an alternative teaching strategy and testing procedure which is derived from wrong answer information is proposed, and a description is given of a teacher inservice course based on these procedures. (JD)