If I were feeling less lazy, I'd do a rank correlation with pay to test this hypothesis.
[WS:] Easier said than done. You would need to control for the amount of R's exposure to a particular occupational group, which is not easy to conceptualize or measure.
In a nutshell, there is a difference in perception based on actual experience and that based on what sociologists call status generalization (which has little to do with experience). According to stats generalization theories, high status people are also perceived as more skilled and honest than low status people in the absence of any other evidence. Therefore, absent personal experience with representatives of different occupational groups, there is a high correlation between the perceived status and the perceived skills or integrity almost *by definition.*
However, measuring experience/ exposure to different occupational groups is easier said than done because so many variables are involved: last occurrence, duration, intensity, reasons for interaction (e.g. a court-ordered professional treatment vs. voluntary professional treatment), R's education, socio-economic status to name just a few.
What is more, the relations between experience and perceived status of an occupational group is typically recursive (i.e. previous instances affect later instances) and in many possible ways i.e. sometimes with a negative reinforcement (e.g. negative or positive experience being reinterpreted to fit a preconceived image i.e. it is an exception that confirms the rule') , sometimes with positive reinforcements. Recursive modeling introduces another level of complication, but it needs to be accounted for if we are to claim any relationship.
Wojtek