Job values and job characteristics are widely assumed to interact with each other, in that jobholders' preferences are thought to moderate associations of job content with well-being. However an examination of previous research revealed considerable between-feature inconsistency in findings about moderation, and a new contingency variable was introduced to account for that inconsistency. This construct, labelled "affective strength", was defined and investigated through the spread of a feature's desirability in a studied sample. A threesample examination of feature-by-value interactions across a broad set of job features confirmed that moderation by job values is often weak and that patterns vary between features. As predicted, associations between job characteristics and well-being were found to be significantly more influenced by worker preference when those characteristics were of lower affective strength -having greater variance in desirability. Models of job design need to incorporate worker preferences but also the varying influence of those preferences -in effect through a second-order interaction.In reviewing the "person-situation controversy", Pervin (1989) pointed to a general agreement that "most psychologists now see themselves as interactionists" (p. 350) -accepting that both the person and the environment affect people's experience and behaviour. However, as he argued, "interaction" has several meanings and we need to learn more about different forms of operation. For example, does the combined operation of personal and situational factors always involve a statistical interaction? Or are associations with personal and situational variables sometimes independent of each other? In the latter case, what variables might underlie the difference between patterns? Applying Pervin's profession-wide account to the specific area of organizational research, work and organizational psychologists typically see themselves as interactionists. This is exemplified in the widespread acceptance of conclusions drawn by Hackman and Lawler (1971) and Hackman and Oldham (1980) about the notion of "growth-need strength". Within their model of job characteristics and psychological outcomes, these authors proposed a moderating influence from workers' preference for a compound of intrinsic job featureshow much they would like to have that compound set of features in a job. They found that the correlation of intrinsic characteristics with job-related well-being tended to be greater for workers whose preference for the set of features (their "growth-need strength" or "GNS")