How sensitive are handedness prevalence figures to differences in questionnaire classification procedures?

Questionnaire data from 572 subjects show that the proportion of right-handers to nonright-handers varies greatly (from 1:100 to 92:8) as a function of questionnaire length, graded or forced response to items, and cut-off criteria. For this reason, the right-handed/nonright-hander dichotomy is singularly unsuited as a basis for classification when handedness is to be related to some other neuropsychological factor as an independent variable. This is especially the case when very strict cut-off criteria are used to define right-handedness (e.g., only those individuals who prefer the right hand for all activities listed).