User participation in information systems development is considered the key to system success in organizations. The empirical evidence, however, does not support this. A review of the literature suggests that one critical weakness in empirical investigations is inadequacy of operational measures of participation to gauge user influence on system design. Furthermore, there is also a growing consensus that the contradictory evidence may be due to a contingent, rather than a direct, relationship between participation and system success. This conception asserts that the outcome of user participation may depend on various contextual variables. One variable in particular--users' system-related functional expertise--is believed to moderate the outcome of participation. This paper derives the contingent effect of user expertise and reports the results of a controlled laboratory experiment and a field survey conducted to test it. The data suggest that users who perceive themselves as functional experts, relative to others, are unlikely to accept a system unless they exerted a substantive influence on its design. On the other hand, users who perceive themselves as functional nonexperts, relative to others, are likely to accept a system regardless of the extent of their influence on its design. This finding suggests user expertise as a useful criterion for selecting participants to serve on design teams and for determining the appropriate extent of a participant user's influence on system design.