Construct validity (CFA) eksempelklausuler

Construct validity (CFA). A total of 500 responses were used for the Confirmatory Factor Analysis (CFA), which was conducted on all 43 MJS items using the 7 predetermined subscales, with the use of different reference values provided by the program, IBM SPSS AMOS, to conclude on how the data fitted the model. Hereby, the model fit, the factor loadings in terms of item loading on their subscales, and the correlations of the subscales are further discussed. The construct validity of the MJS instrument was evaluated using CFA on data from 500 re- spondents. The GFI and AGFI values were 0.766 and 0.736, respectively. Both values were found to fall below the commonly accepted threshold of 0.9 for GFI and 0.95 for AGFI, sug- gesting that the model did not achieve a perfect fit with the data (29). The NFI value of 0.805 and the TLI value of 0.838, did not reach the ideal threshold of ≥0.95, indicating an acceptable level of validity of the satisfaction scale with room for improvement (28). The RMSEA value of 0.073 fell within the accepted value of a good fit below 0.08. However, the RMR value of 0.087, exceeded the preferred maximum of 0.08, highlighting an area for potential improve- ment. The Comparative Fit Index (CFI) of 0.982, which was very close to 1, indicated an ex- cellent model fit. A CFI value above 0.95 was considered good, further supporting the ade- quacy of the model (29). In summary, the results of the CFA suggested that the 7-factor model of the MJS items was acceptable. CFA for the translated MJS questionnaire yielded factor loadings ranging from 0.44 to 0.91. Notably, most items demonstrated strong associations with their respective subscales, with the majority of loadings exceeding 0.7 (26). For an item to demonstrate a good correlation with its factor, in this study, a loading of 0.65 or higher was preferred. Items that fell below this thresh- old suggested weaker associations with their respective subscales, potentially due to issues in item formulation or misalignment with the construct they were intended to measure. Out of 43 items, only 10 did not load well with 5 of their respective factors, with most of these poorly loading items concentrated in the subscale related to Prospects. The lower loadings in the Pro- spect-subscale reflected variability in perceptions of promotion opportunities, job security, and prospects. Moreover, this the sample was quite heterogeneous, as it consisted of both hospital- and community pharmacists and PTs. This diversity might have ...