| Objective:Cognitive impairment of the elderly has become the key point in the 21 century with the fast steps of population aging. It is urgent to develop a tool and national norms for more completely assessing cognitive functions of the elderly. The Intelligence Scale for Chinese Adults (ISCA) was recently developed in our country. The ISCA contains eleven core subtests and an additional subtest, four of which were designed to measure crystallized intelligence (Vocabulary, Similarities, Information, and Comprehension), four developed to evaluate fluid intelligence (Picture Completion, Block Design, Matrix Reasoning, and Picture Arrangement), and three intended to assess memory and attention (Digit Symbol-Coding, Arithmetic, and Digit Span). An early evaluation of both the reliability and validity of this measure has been completed, and the national norms of ISCA for adults ranging from 16 to 64 years old have been developed. The purpose of this study was to develop the national norms of ISCA for the elderly over 65 years old and examine its psychometric characteristics to meet the urgent needs of cognitive aging study and practice.Methods:The ISCA standardization was completed on 435 aged subjects.2000 Census data was applied to the standardization sample in order to correctly represent individuals according to age, gender, educational level, geographic region, and occupation. ISCA provided a variety of scores such as age-corrected scaled scores, IQ scores, Index scores and percentile rank. The reliability analysis included split-half reliability, Cronbach's a, test-retest reliability, standard error of measurement, inter-rater reliability and generalizability analysis. The validity analysis consisted of construct validity and criterion-related validity.Results:1. Item analysis:the average difficulty of subtest was 0.34, and most of the items were acceptable, good or excellent discrimination. 2. Norms:The sums of subtest scaled scores for the ISCA IQ and Index scores were formed by summing each individual's actual age-corrected scaled scores on the relevant subtests. For each scale, the distribution of the sums of scaled scores was converted to a scale with a mean of 100 and a standard deviation of 15. This conversion was accomplished by preparing a cumulative frequency distribution of actual sums of scaled scores for each scale and index, smoothing and normalizing these distributions, and then calculating the appropriate IQ and Index score equivalent for each sum of scaled scores. These normative scores were reported in the user's manual in retail. The overall trend on the ISCA was that nearly all subtest raw scores decreased at successive ages, which were statistically significant (P<0.01). The full raw score of ISCA was statistically significant lower for individuals aged over 70 and even lower for those of aged 80 and older. The higher educational level was, the higher IQ scores and Index scores were. Gender difference was notable on the IQ scores and Index scores, and males outscored females. But, the effect size of educational factor on the full IQ score of ISCA was larger than that of gender, with 0.371 and 0.078, respectively.3. Reliability research:(1) The average spilt-half reliability for the IQ scales ranged from 0.96 to 0.97 and the factor indices had reliability coefficients ranging from 0.90 to 0.96.(2)Excepting the subtest Digit Symbol-Coding and Digit Span, the Cronbach's a coefficients of the other subtests ranged from 0.673 to 0.895, while the Cronbach's a coefficient of full scale was 0.893. (3) The average standard error of measurement of subtest ranged from 0.85 to 1.97, the IQ scales ranged from 2.37 to 3.09, and the factor indices ranged from 3.09 to 4.62. (4) The test-retest coefficients of the Verbal IQ, Performance IQ, Full scale IQ, Crystallized intelligence Index, Fluid intelligence Index, and Memory-attention Index were 0.95,0.95, 0.96,0.92,0.92, and 0.95, respectively. (5) The inter-rater reliabilities of the vocabulary, similarities and comprehension subtest were 0.98,0.95, and 0.95, respectively. (6) The generalizability analysis showed that the range of generalizability coefficient were from 0.67 to 0.90. In a word, the true score of the ISCA was above 0.89.4. Validity research:(1) The correlations between the sum of scaled scores of subtests and that of ISCA ranged from 0.64 to 0.83, there were median correlations among subtests. By the way of confirmatory factor analysis (CFA), three-factor model was fitted well across each age group, which named as crystallized intelligence factor, fluid intelligence factor, and memory-attention factor. (2) The total score of ISCA related to that of HDS-R and NTBE were 0.33,0.49, respectively, which were statistically significant (P<0.05 and P<0.01, respectively). Compared with those of the normal group, the IQ scores and Index scores of clients with cerebral infarction were lower significantly (P<0.01).Conclusion:The ISCA has generated good items, reliability, and validity. The national norms of ISCA for the elderly over 65 years old have been developed and can be applied to assess cognitive functions of the elderly fully.
|
PDF Full Text Request