Intelligence Tests Compared: Stanford-Binet Intelligence Scale versus Otis-Lennon School Abilities Test: Part 4
Critical Review of Otis-Lennon School Ability Test
Demographic traits of the standardized samples include geographic region, socioeconomic status, urbanity, ethnicity, nonpublic schools, and students with disabilities. These characteristics are found in the Preliminary Technical Manual and are similar to the demographics given for the U.S school enrollment. (Goldman, 2004)
Standardized samples of examinees were attained through a stratified random sampling technique. This was done to attain normative and descriptive data from students throughout the nation regarding the Otis-Lennon School Ability Test. Students who took part in these samples were also administered the Stanford Achievement Test Ninth Edition. During the Spring and Fall of 1995, approximately 463,000 students took part in this sample. The standardized samples and statistical data can be found in the OLSAT Preliminary Test Manual and the OLSAT Test Manual.
Reliability coefficients for Verbal, Nonverbal, and Total scores for each grade are mostly in the .80s and .90s. Internal consistency is shown with these coefficients, rather than test-retest stability. (Goldman, 2004) Measures of stability of the OLSAT scores over a period of time are not available in the Manual. (Destafano, 2004)
Content, Criterion-related, and construct validity are presented in the OLSAT Technical Manual. Three issues are presented as primary evidence for validity. Correlations between the Sixth and Seventh Edition fall between .68 to .87, with the majority in the .70s and low .80s. This is statistically significant but if there were no significant change between the two editions then the correlation should be much higher. (Goldman, 2004)
In the next article we will compare both the Otis-Lennon School Abilties Test and the Stanford-Binet Intelligence Scale. Article 5 will look at pros and cons for both tests and we will conclude our series of articles.