Abstract
A field study of supported employment for adults with mental illness (N=174) provided an experimental test of cognitive dissonance theory. We predicted that most work-interested individuals randomly assigned to a non-preferred program would reject services and lower their work aspirations. However, individuals who chose to pursue employment through a non-preferred program were expected to resolve this dissonance through favorable service evaluations and strong efforts to succeed at work. Significant work interest-by-service preference interactions supported these predictions. Over two years, participants interested in employment who obtained work through a non-preferred program stayed employed a median of 362 days versus 108 days for those assigned to a preferred program, and participants who obtained work through a non-preferred program had higher service satisfaction.