The College Student Self-report Mental State Scale: Item Selecting, Reliability, And Validity

Objective: The purpose of this research was to: analyze and select the items of the college student self-report mental state scale developed by the study group of"the theory and practice studies on college students'mental health education";develop the structural model of the scale;examine the reliability and validity of the scale to produce the reliability and validity proof to the scale's use.Method: 8152 college students of universities were taken from 12 provincial capitals and cities in Western China by stratified random sampling. There were 7362 useful samples (4173 males, 3189 females) after surveying. Item analysis and preliminary exploratory factor analysis were used to select the items. Then the big samples were split into two groups equally. Exploratory factor analysis was carried out in the one half of the samples to delete some items which were not fit to the demands of psychometrics and to develop the factor-structure of the scale. Confirmatory factor analysis was carried out in the other half of the samples to evaluate the factor-structure of the scale. The reliability analysis included Cronbach'sα, split-half reliability, and test-retest reliability. The validity analysis included content, construct, and criterion-related validity.Results: (1) Using item analysis and exploratory factor analysis, 71 items were included in the scale. (2)The further exploratory factor analysis identified that the scale had six-factor structure, explaining 61.136% of the total variance. Six factors was labeled as follow: Concept of career choosing, development and creative ability, self-concept, concept of love, interpersonal adaptation, learning adaptation. (3)Reliability analysis'results showed that the Cronbach'sαcoefficient, split-half and test-retest reliability coefficient ranged from 0.791 to 0.905, from 0.812 to 0.938, from 0.731 to 0.904 respectively for each factor, and were 0.938, 0.777, 0.897 respectively for the total scale. (4)The scale's internal consistency analysis showed that the correlation coefficients between items and total scale ranged from 0.314 to 0.643, the correlation coefficients between factors and total scale ranged from 0.589 to 0.720. All correlation coefficients were significant(P<0.01).Thus the scale had high internal consistency. (5)The scale's internal discrimination analysis showed that most of the correlation coefficients among six factors ranged from 0.146 to 0.576, and most of them were medium correlation coefficients. All correlation coefficients were significant(P<0.01). Thus the scale had high internal discrimination. (6)Confirmatory factor analysis showed that the theory structural model fit the observed variables well, and supported six-factor structure model of the scale. (7)We mainly studied the differences between male and female, only-child and non-only-child, Han and minority college students'mental state to analyze the criterion-related validity. The results showed that there was no difference between male college students and female college students in the mental health state, but the only-child college students'mental health state was significantly better than the non-only-child college students', and the Han college students'mental health state was also significantly better than the minority college students'. The results were same to the concerned studies conducted in the past, and showed that the scale's criterion-related validity achieved good to excellent levels.Conclusions: (1)The study finished selecting the items of the college student self-report mental state scale and developing the theory structure of the scale, and confirmatory factor analysis studies showed that the structure of the scale is comparatively reasonable. (2)The reliability and validity analysis showed that the scale's homogeneity, split-half, test-retest reliability, and content, construct, criterion-related validity achieved good to excellent levels of psychometrics.