Predictive validity of the intellectual constructs identified on the universal nonverbal intelligence test and the woodcock-johnson psychoeducational battery-revised with achievement

The predictive validity of cognitive constructs taken from two major theories of intelligence- Cattell-Horn's Model and Jensen's theory. The Horn-Cattell model of intellectual processing is an extension of Gf-Gc theory (fluid and crystallized abilities). Jensen posits a general intelligence factor and categorization of mental abilities into memory and reasoning constructs. Gf-Gc cognitive constructs were measured using the Woodcock-Johnson-Revised Tests of Cognitive Ability (WJ-R; Woodcock & Johnson, 1989, 1990) and include: processing speed, fluid reasoning, acculturation-knowledge, short-term memory, long term retrieval, auditory processing, and visual processing. Operational measures of Jensen's two level theory were assessed through the Universal Nonverbal Intelligence Test (UNIT; Bracken & McCallum, 1992). The predictive utility of the WJ-R aptitude cluster scores was also examined. Scores from the Comprehensive Test of Basic Skills (CTB; Macmillan/McGraw Hill, 1989) were used as the criterion measures. The instruments were administered according to standardized procedures to 104 students.

Multiple regression equations were calculated to determine the relative predictability of the WJ-R, Gf-Gc constructs and Aptitude Cluster scores for specific achievement areas. In equations using the WJ-R Gf-Gc constructs, various combinations of the Comprehension-knowledge, Fluid Reasoning, and Processing Speed variables were consistently found to be the best predictors of achievement. Multiple Rs ranged from the 60s to .705. Equations using the WJ-R Aptitude Cluster scores produced multiple correlations of similar magnitude in the prediction of achievement.

Multiple regression analyses of the UNIT revealed among global scores the Symbolic, Memory, and Reasoning Quotients were consistently the best predictors to all academic achievement areas. In regression analyses of UNIT subtests Analogic Reasoning, Cube Design, and Symbolic Memory were identified as the best predictors of achievement. Evidence of the UNIT's predictive validity to achievement as equivalent to or better than other nonverbal assessment measures was also demonstrated. Multiple Rs ranged from the .40s to .605.

Overall, results provide evidence for the importance of cognitive constructs across verbal and nonverbal assessment instruments to achievement. Determination of how particular constructs contribute to various areas of achievement is potentially useful for program planning and aptitude treatment interaction research.