A longitudinal test of the association between motivation to undertake a precautionary health action and subsequent behaviour was conducted on women's uptake of the cervical screening test. A sample of never-screened women (N = 166) completed measures derived from protection-motivation theory (PMT; Rogers, 1983). One year later, screening uptake was reliably determined from medical records. While regression analyses demonstrated that PMT variables predicted both motivation to undergo cervical screening and screening uptake, there was, nonetheless, a good deal of inconsistency between protection motivation and screening behaviour. Fifty-seven per cent of those who indicated they were willing to undergo the test did not do so within a one-year period. Discriminant analysis was therefore used to test the ability of cognitions specified by PMT to distinguish four patterns of association between motivation and behaviour: inclined actors, inclined abstainers, disinclined actors and disinclined abstainers. While PMT variables could successfully classify the four groups, it was not possible to derive a reliable discriminant function which distinguished between inclined actors and inclined abstainers. The results suggest that PMT provides a useful account of choice motivation but does not address the psychological processes by which intention is translated into action. Recent calls for the development of a social psychology of volition are discussed in the light of these findings.