The purpose of this study was to confirm the causal structure model of muscle, motor and living functions utilizing structural equation modeling (SEM) . As subjects, 103 community-dwelling older men and women, aged 65.7±6.9years of age, participated in the study to measure muscle cross-sectional area, maximum voluntary contractions, muscle power, 4 physical performance tests, and 16 questionnaires regarding ability of activities of daily living. The causal structure model of muscle, motor and living functions was hypothesized to be a hierarchical causal structure. The causal structure model of muscle function was hypothesized to be a hierarchical causal structure consisting of 3 sub-domains of muscle mass, muscle strength, and muscle power. Data analysis procedures were as follows : a) testing of construct validity of muscle function variables using confirmatory factor analysis (CFA) in SEM ; b) testing of causal structure using SEM ; c) testing of factor invariance using multi-group analysis for gender. The highest goodness of fit indices was obtained in the causal structure model of muscle, motor and living functions (NFI= .928, CFI= .978, RMSEA =.061) . The causal coefficient of muscle function to motor function was .98 (p<.05), followed by.34 for motor function to living function. From the results of multi-group analysis, the measurement invariance model indicated the highest goodness of fit indices (TLI=.968, CFI .977) . It was concluded that the hierarchical causal relation was among muscle, motor and living functions, and in which muscle function was consisted of 3 sub-domains.

The purpose of this study was to confirm the reliability and validity of a physical fitness questionnaire (PFQ) with self-rating for elderly people applying structural equation modeling (SEM) . As subjects, 105 community-dwelling older men and women aged 67.1±6.1 years participated in the study to measure 13 PFQ items and 13 performance tests. The data analysis procedures were as follows : a) testing reliability of PFQ ; b) testing of construct validity of PFQ using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) ; c) testing of criterion-related validity of PFQ to the performance tests using SEM ; d) testing of correlations of the PFQ to walking ability using SEM. Cronbaeh's alpha coefficient for consistency reliability of the PFQ was .83. Four common factors of muscle strength-power, endurance, coordination, and flexibility were extracted in EFA. The high and enough goodness of fit indices were obtained in the confirmatory factor structure model, and in each sub-domain of criterion-related validity to performance tests and correlation to walking ability models. The criterion-related validity coefficient of muscle strength and power was .77, followed by .66 for endurance, .59 for coordination and .82 for flexibility. The correlation coefficient of muscle strength and power to walking ability was -.51, followed by -.58 for coordination, - .43 for endurance and - . 28 for flexibility. These results indicated that the PFQ consisting of 13 items and 4 sub-domains satisfied reliability and construct validity although criterion related validity to performance tests was insufficient. It was concluded that the PFQ is of useful for physical fitness checking of elderly people.

The purpose of this study was to confirm the causal effect model of strength on walking ability development as a result of exercise participation among elderly people in a community, utilizing the latent curve model (LCM) in structural equation modeling (SEM) . Twenty-six male and 57 female subjects, 83 in total, aged 67.8±5.7, 63.9±7.1 and 65.1±6.9 in a pooled sample participated in the exercise program which lasted for two years. Grip strength and sit-ups used in the Japan Fitness Test were measured for muscular strength, 10-m hurdle walk and 6-min walk for walking ability, and the fitness test score for physical ability. The data analysis procedures were as follows : a) analysis of test-retest reliability and construct validity of measurement items, b) analysis of causal structure model of aging, muscular strength and walking ability, c) analysis of variance for repeated measurement of walking performance by sex, age and year, d) analysis of LCM for walking performance development. The highest goodness-of-fit indices of SEM were obtained in the LCM of 10-m hurdle walk performance development (GFI=0.989, AGFI=0.920, CFI=0.998, RMSEA=0.038) . The path coefficient of sit-ups at pre-test effect on the intercept of 10-m hurdle walk performance development was significant (p<0.05) . The path coefficients of age to intercept and slope of 10-m hurdle walk development were also significant (p<0.05) . It was concluded that walking ability development through participation in exercise age and strength level was more effective for maintaining walking ability in older age.