| 英文摘要 |
To reduce research cost and to increase participation, as well as to make it convenient for initial counseling guidance and clinical interview, this study developed a shortened version of Big Five personality scale on basis of Goldberg’s IPIP-50. The study presents among all its construct validity, criterion-related validity, convergent validity, internal consistent reliability, and also checks its across age-level measurement invariance for middle-aged and older people. There are two sample data surveyed from middle-aged and older people in Taiwan. The first survey had a valid sample with 455 participants, and one year later, a second survey was conducted and 283 non-duplication participants were obtained. The two samples included 426 below 60 year-old middle-aged people and 312 over 60 year-old people, respectively. The items of the shortened version were analyzed and chosen by measurement model of structural equation modeling. Its construct validity was checked by confirmatory factor analysis with model-fit indices and factor loadings. Its criterion-related validity was established by correlations with the affect balance scale and the self-esteem scale. Its convergent validity was obtained by correlations with IPIP-50 and with another Big Five personality scale, TDA-40. The results show that a 15-item shortened version of Big Five personality, named IPIP-15, was developed, and its each personality included 3 items. The IPIP-15 has good model-fits with factor loadings ranging from .61 to .83. Its internal consistent reliability coefficients α ranged from .67 to .83, and its criterion-related validity and convergent validity are also as expected. Besides, IPIP-15 has an across age-level measurement invariance between middle-aged and older people. In general, IPIP-15 has acceptable reliability and validity. |
