Confirmatory factor analysis method was employed to verify the construct validity of the measurement model assumed to underlie the Athletic Cognitive Pattern Scale (ACPS, Liaw, 1993). The ACPS was designed to measure the psychological ski1ls of elite athletes in Taiwan. One hundreds and fifty-six college athletes were used in this structural equation model with LISREL 8.30. The results of the analysis were shown as follows: 1. The last model has been approached by Maximum Likelihood method (χ2 (574)=1043.07, p<.01, GFI=.72, AGFI=.68, RMSEA=.O7, TLI=.70) with ten error terms correlated within and across two different factors. Lambda coefficients ranged from .04 to .81. The correlations between the ACPS subscales ranged from .15 to .90, indicating low to high correlations as well as some multicollinearity problems. 2. The Cronbach α coefficient (an index of internal consistency) was moderately acceptable with value from .52 to .74. In conclusion, the measures of overall fit for the hypothesized five-factor structure of the ACPS were not acceptable. The ACPS didn't meet the construct validity standards required of a multidimesional instrument that was to be used for research.
Confirmatory factor analysis method was employed to verify the construct validity of the measurement model assumed to underlie the Athletic Cognitive Pattern Scale (ACPS, Liaw, 1993). The ACPS was designed to measure the psychological ski1ls of elite athletes in Taiwan. One hundreds and fifty-six college athletes were used in this structural equation model with LISREL 8.30. The results of the analysis were shown as follows: 1. The last model has been approached by Maximum Likelihood method (χ2 (574)=1043.07, p<.01, GFI=.72, AGFI=.68, RMSEA=.O7, TLI=.70) with ten error terms correlated within and across two different factors. Lambda coefficients ranged from .04 to .81. The correlations between the ACPS subscales ranged from .15 to .90, indicating low to high correlations as well as some multicollinearity problems. 2. The Cronbach α coefficient (an index of internal consistency) was moderately acceptable with value from .52 to .74. In conclusion, the measures of overall fit for the hypothesized five-factor structure of the ACPS were not acceptable. The ACPS didn't meet the construct validity standards required of a multidimesional instrument that was to be used for research.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。