Stress levels are said to be rising in many different occupations but one problem for cross-occupation comparison purposes is that different questionnaires have been used in different studies - often specially designed questionnaires for the occupation under study. The Occupational Stress Inventory-Revised (OSI-R) is one questionnaire that may help assess the same stress-related variables across different occupational groups. The OSI-R model is theory-based and assesses the effects on the individual of three 'factors' (occupational roles, psychological strain, and coping resources) across fourteen dimensions. This current study reports the findings of a re-analysis of the original Manual data of 983 mixed occupational respondents of the OSI-R, using confirmatory factor analyses of the inter-correlations given of the 14 dimensions. The findings show that the three-factor solution is not optimal (though two of the three original factors are accurately reproduced) and that a four-factor solution better fits the responses, but with more error than desirable in both solutions. This finding is consistent with an earlier confirmatory factor analysis of the responses of 141 teachers which suggested that a four-factor rather than three-factor solution was preferred. Implications for use of the OSI-R and for further research are drawn.