Standard measures of self-efficacy (SE) may confound perceived capability and motivation because respondents interpret the word "can" as "will". Objective: To test whether priming for the meaning of the word "can" changes self-efficacy ratings. Design: In an experimental test, 134 university students responded to an on-line standardized measure of exercise SE and provided definitions of the words "can" and "will". One month later participants were randomized to complete (a) the same questionnaire (control), (b) the same questionnaire but with presentation of each participant's definition of "can" prior to the SE measure (definition priming), or (c) the same questionnaire but with SE items ("I can exercise…") placed side-by-side with behavioral intention items ("I will exercise…") (side-by-side priming). Results: SE increased relative to controls for side-by-side (b = 12.08, SE = 2.70, p<.01) but not definition priming (ns), with the former even stronger among participants (n = 91) who provided strict (i.e., literal) rather than liberal definitions of "can" (b = 15.38, SE = 3.21, p<.001). Conclusion: Priming of the meaning of the word "can" led to increases in self-efficacy ratings among those who hold a literal meaning of the word "can". This suggests that for many respondents standard assessments of SE may be confounded by motivation.
Keywords: Self-efficacy; behavior; construct validity; intention; motivation.