Estimation of premorbid general fluid intelligence using traditional Chinese reading performance in Taiwanese samples

• Published in: Psychiatry and clinical neurosciences 2009-08, Vol.63 (4), p.500-507
• Main Authors: Chen, Ying‐Jen, Ho, Meng‐Yang, Chen, Kwan‐Ju, Hsu, Chia‐Fen, Ryu, Shan‐Jin
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; Age Distribution; Aged; Aged, 80 and over; Asian Continental Ancestry Group - psychology; Asian Continental Ancestry Group - statistics & numerical data; Biological and medical sciences; Chinese Graded Word Reading Test; Chinese reading ability; Educational Status; Female; Humans; Intelligence Tests - statistics & numerical data; Language; Male; Medical sciences; Middle Aged; premorbid general fluid intelligence; Psychology. Psychoanalysis. Psychiatry; Psychometrics; Psychopathology. Psychiatry; Raven's Standard Progressive Matrices; Reading; Regression Analysis; Reproducibility of Results; Taiwan; Wechsler Scales - statistics & numerical data
• ISSN: 1323-1316; 1440-1819
• DOI: 10.1111/j.1440-1819.2009.01970.x

Aims:  The aims of the present study were to (i) investigate if traditional Chinese word reading ability can be used for estimating premorbid general intelligence; and (ii) to provide multiple regression equations for estimating premorbid performance on Raven's Standard Progressive Matrices (RSPM), using age, years of education and Chinese Graded Word Reading Test (CGWRT) scores as predictor variables. Methods:  Four hundred and twenty‐six healthy volunteers (201 male, 225 female), aged 16–93 years (mean ± SD, 41.92 ± 18.19 years) undertook the tests individually under supervised conditions. Seventy percent of subjects were randomly allocated to the derivation group (n = 296), and the rest to the validation group (n = 130). Results:  RSPM score was positively correlated with CGWRT score and years of education. RSPM and CGWRT scores and years of education were also inversely correlated with age, but the declining trend for RSPM performance against age was steeper than that for CGWRT performance. Separate multiple regression equations were derived for estimating RSPM scores using different combinations of age, years of education, and CGWRT score for both groups. The multiple regression coefficient of each equation ranged from 0.71 to 0.80 with the standard error of estimate between 7 and 8 RSPM points. When fitting the data of one group to the equations derived from its counterpart group, the cross‐validation multiple regression coefficients ranged from 0.71 to 0.79. There were no significant differences in the ‘predicted–obtained’ RSPM discrepancies between any equations. Conclusion:  The regression equations derived in the present study may provide a basis for estimating premorbid RSPM performance.