Title

The Validity/Reliability of Occupational Performance Measurement: A Synthesis of Research Using the Validity Generalization Method

Presenter Information

Moses N. Ikiugu
Lynne M. Anderson

Start Time

24-10-2008 8:30 AM

End Time

24-10-2008 9:00 AM

Abstract

In this study, we used the Validity Generalization method (Hunter & Schmidt, 2004; Schmidt & Hunter, 1977) to synthesize findings from 18 studies investigating the validity/reliability of a variety of occupational performance assessments. Our objectives were to determine: 1) an overall estimate of the validity/reliability of occupational performance measurement scores; and 2) generalizability of the validity/reliability from research to clinical settings after correction for some attenuating statistical artifacts. To achieve the above two objectives, we computed weighted mean validity/reliability coefficients of studies validating a variety of occupational performance measurement instruments; determined the attenuation of variability of the validity/reliability coefficients by sampling and test criterion measurement reliability errors; and calculated the proportion of the variance of the validity/reliability coefficients accounted for by the attenuating factors. Our sample was comprised of 18 studies in which test-retest, inter-rater, alternate measure, and predictive reliability estimates of occupational performance scores were investigated. The instruments generating scores in the studies included: the Canadian Occupational Performance Measure (COPM); Occupational Performance History Interview (OPHI-II); Assessment of Motor and Process Skills (AMPS); Role Checklist (RC); Assessment of Living Skills and Resources (ALSAR); Australian Therapy Outcome Measures (AusTOMs); Functional Independence Measure (FIM); and Hessel Analogical Reasoning Test (HART) among others. Occupational Performance assessment scores based on self-report were found to have a higher corrected weighted mean validity/reliability coefficient than is typical for instruments in social science research. This was particularly significant in the context of the emphasis on client-centeredness in the current occupational therapy paradigm which encourages collaboration with clients in the assessment and intervention process. When observed variance was corrected for attenuation by sampling and test criterion reliability errors, less than 75% of the variance recommended by Hunter and Schmidt (2004) remained. Our findings indicated that assessment scores based on self-report instruments may be the most reliable/valid. It is not clear whether such validity/reliability can be assumed in clinical conditions that differ from research circumstances. Further meta-analysis is indicated to determine more conclusively such generalizability of validity/reliability of occupational performance assessments.

This document is currently not available here.

Share

COinS
 
Oct 24th, 8:30 AM Oct 24th, 9:00 AM

The Validity/Reliability of Occupational Performance Measurement: A Synthesis of Research Using the Validity Generalization Method

In this study, we used the Validity Generalization method (Hunter & Schmidt, 2004; Schmidt & Hunter, 1977) to synthesize findings from 18 studies investigating the validity/reliability of a variety of occupational performance assessments. Our objectives were to determine: 1) an overall estimate of the validity/reliability of occupational performance measurement scores; and 2) generalizability of the validity/reliability from research to clinical settings after correction for some attenuating statistical artifacts. To achieve the above two objectives, we computed weighted mean validity/reliability coefficients of studies validating a variety of occupational performance measurement instruments; determined the attenuation of variability of the validity/reliability coefficients by sampling and test criterion measurement reliability errors; and calculated the proportion of the variance of the validity/reliability coefficients accounted for by the attenuating factors. Our sample was comprised of 18 studies in which test-retest, inter-rater, alternate measure, and predictive reliability estimates of occupational performance scores were investigated. The instruments generating scores in the studies included: the Canadian Occupational Performance Measure (COPM); Occupational Performance History Interview (OPHI-II); Assessment of Motor and Process Skills (AMPS); Role Checklist (RC); Assessment of Living Skills and Resources (ALSAR); Australian Therapy Outcome Measures (AusTOMs); Functional Independence Measure (FIM); and Hessel Analogical Reasoning Test (HART) among others. Occupational Performance assessment scores based on self-report were found to have a higher corrected weighted mean validity/reliability coefficient than is typical for instruments in social science research. This was particularly significant in the context of the emphasis on client-centeredness in the current occupational therapy paradigm which encourages collaboration with clients in the assessment and intervention process. When observed variance was corrected for attenuation by sampling and test criterion reliability errors, less than 75% of the variance recommended by Hunter and Schmidt (2004) remained. Our findings indicated that assessment scores based on self-report instruments may be the most reliable/valid. It is not clear whether such validity/reliability can be assumed in clinical conditions that differ from research circumstances. Further meta-analysis is indicated to determine more conclusively such generalizability of validity/reliability of occupational performance assessments.