Title:
|
The psychometric properties of the Irish Management Standards Indicator Tool and its associations with the WHO-Five Well-being Index.
|
The primary purpose of this thesis was to psychometrically evaluate the Irish version (English language) of the
UK Health and Safety Executive's Management Standards Indicator Tool (ROI-MSITL a self-report survey
instrument that assesses perceptions of stressors in the workplace. To that end, the thesis undertook a
thorough examination of the ROI-MSIT's scales and items to appraise its criterion-related validity, construct
validity, factor structure, reliability, and measurement invariance across gender. Associations between the
ROI-MSIT and the WHO-Five Well-being Index (WHO-5) were also investigated in order to explore the
salutogenic potential of the Management Standards' risk model, the secondary purpose of this thesis. The aim
of Study I was to provide a preliminary investigation of the psychometric properties of the ROI-M5IT. In
particular, the study examined the factor structure, internal consistency, and convergent validity (with the
WHO-5) of the ROI-MSIT in order to determine its suitability for use in Ireland. Exploratory factor analysis
revealed that the factor structure of the ROI-MSIT is practically identical to that of the Italian MSIT, consisting
of six factors; the Demands, Control, Peer Support, Relationships, and Role factors were equiva lent to the
original UK factors. As with the Italian version, the principal factor was a merger of the Manager Support and
Change domains. Reliability analysis of the subscales revealed Cronbach's alpha scores ranging from .80 to .91.
Finally, the ROI-MSIT's (six) subscales and WHO-5 were found to be positively correlated (r = .36-44, P <.001),
while logistic regression analyses showed that increased scores on the individual items and subscales were
associated with a decreased likelihood of being classified in the poor well-being category of the WHO-5 (score
<13). Study II used confirmatory factor analysis (CFA) to assess competing measurement models of the ROIMSIT,
including the seven-factor structure of the UK Indicator Tool. The construct reliability and internal
consistency of the ROI-MSIT were also assessed, as were its convergent and discriminant validity. The seven-factor
measurement model of the UK-MSIT provided the best fit to the data compared with a six-factor and
one-factor measurement model. A second-order measurement model also proved to be of adequate fit. The
construct reliability (CR) and internal consistency were excellent, as each factor achieved a CR and alpha score
of .79 or more. The convergent and discriminant validity results were, on the other hand, disappointing, as
several subscales failed to meet the minimum values necessary for demonstrating validity. The primary aim of
Study III was to test the measurement invariance (MI) of the ROI-MSIT across gender. The seven-factor
measurement model and items of the ROI-MSIT were tested for MI; multiple-group confirmatory factor
analysis (MG-CFA) was used to test the model, while tests of differential item functioning (DIF) were applied to
each item through ordinal logistic regression. The results of these analyses indicated that the ROI-MSIT is fully
invariant across gender. The primary aim of Study IV was to investigate associations between scores on the
ROI-MSIT and the likelihood of being categorised in the 'good wel l-being' category of the WHO-5 (total score
>12). Correlation analysis revealed positive, moderate associations between the (uni-dimensional) ROI-MSIT
and the WHO-5 (r = .53, P <.001). In the logistic regression analyses, those in the top ROI-MSIT scoring
category were 18.44 times (95% CI: 14.46-23.50) more likely to be classified in the good well-being category of
the WHO-5 than those in the bottom scoring group. Similarly, ordinal logistic regression revealed that those in
the top category were 21.18 times (95% CI: 17.74-25.28) more likely to be in the good well-being category and
flourishing category (WHO-5 score >19). Strong results were also found for the ROI -MSIT's subscales.
Although additional research is needed to determine the full salutogenic potential of the Management
Standards' risk model, the initial findings of Study IV suggest that following the approach may help employees
experience positive mental states rather than just the absence of psychological harm. Small sh ifts in mean
scores on the Management Standards Indicator Tool could potentially move the working population towards
mental flourishing. The general conclusions of the studies are that a) the ROI-MSIT bears a very close
resemb lance to the UK version, b) the ROI-MSIT is a valid and reliable questionnaire, and c) the Management
Standards initiative may help employees (and working populations) experience greater 'positive mental
health.' Further research is needed to validate the findings of the thesis.
|