Stanford-Binet Intelligence Scale - Form L-M, Third Edition

The Stanford-Binet was created by Lewis M. Terman and Maud A. Merrill and was published in 1972. It is a wide-range individual test, assessing intelligence from age two through the superior adult level. It is an age scale, requiring subjects to solve problems, give definitions, memorize new material, and use some visual-motor skills at various age levels; and is used for assessment in educational, research, and clinical settings.

12 Sub-Tests of the Standford-Binet

• vocabulary
• comprehension
• absurdities
• verbal relations
• pattern analysis
• copying
• matrices
• paper folding and cutting
• quantitative
• bead memory
• memory for sentences
• memory for digits

The examiner must establish a basal age of the child and continue testing until a ceiling is reached. Interpretation is based primarily on objective scoring.

Population: Ages 2 to adult.

Score: Scaled scores, area scores, and a Composite Score.

Time: 30 - 90 minutes.

Authors: Lewis M. Terman and Maud A. Merrill.

Publisher: The Riverside Publishing Co.

Reliability: Reliability of the scale varies at different ages and different IQ ranges. From age 2.5 to 5.5, the reliability coefficients range from 0.83 for IQS 140-149 to 0.91 for IQS 60-69. For ages 6 to 13, the coefficients are 0.91 to 0.79 respectively, and for ages 14-18 the coefficients range from 0.95 to 0.98 respectively.

Validity of the Standford-Binet Scale Depends on Three Sources:

1) the choice of items according to mental age on the 1937 scale assures that the new scale is measuring the same thing that was measured in the original 1916 scale.

2) regular increases in mental age from one age to another agreed with increases in percent passing from one chronological age to the next in both forms of the 1937 scale.

3) biserial correlations were computed for each item of Forms L and M of the 1937 scale.

The retention of an item for the 1960 scale partly depended on its correlation with the total score. The mean biserial correlation for the 1960 scale is 0.66 range from 0.61 at young age levels to 0.73 for the adult levels.

The standardization group consisted of a representative sample of 2,100 children, with approximately 100 subjects at each Stanford-Binet year level. Unlike the 1960 norms, which did not include nonwhites in the standardization group, the 1972 norms contained non whites (including black and Spanish-surnamed individuals) and whites. Subjects were, however, excluded from the normative sample if English was not the primary language spoken in the home.

Scoring: Raw scores are converted into scaled scores (M = 10, SD = 3) using age appropriate tables. The scaled scores are summed for Nonverbal, Verbal and Full Scale IQ as well as for the five factor index scores. These are all normalized standard scores (M = 100, SD = 15). Percentile rank equivalent and the confidence intervals are also obtained. Computerized scoring is faster, provides greater consistency of raw score conversion and is recommended for use whenever available.

Reliability: Using the split half method, and correcting with the Spearman-Brown formula, reliability coefficients were extremely high for the Full Scale Score (0.98). The Nonverbal (0.95) and Verbal (0.96), showed excellent stability, and the Abbreviated Battery (0.91) is also considered excellent as it contains only two subtests. The five factor index scores were all above 90, and were higher than the subtest scales, which were however comparable to other cognitive tests with ranges from 0.84 to 0.89.

Validity: As with the SB: FE, several studies were done to investigate the validity of the SB5. The first such studies explored the relationship between the SB5 and the SB: FE and the Form L-M. There was a high correlation (0.90) between the SB5 Full Scale and the SB: FE Composite Scale. The higher scores of the Composite Scale are explained by the differences in the SD used and the countrywide changes in the IQ (0.3 points per year, as documented by Flynn (1985, 1987). The difference in the Full Scale Score of the SB5 and the one score of the Form L-M was in the direction, but not as great as predicted. Differences in scoring and nonverbal aspects of the tests, as the Full Scale and Verbal Score correlations are high (0.85 and 0.88). Comparisons were also made with other tests such as the WPPSI-R, the WISC-III, the WAIS-III, and the Woodcock-Johnson III test of Cognitive Ability and W-J III Test of Achievement. Correlations ranged from 0.78 to 0.84. for Full Scale or Verbal IQ and comparable scores. This extensive analysis revealed a high correlation between the composite IQ scores of the SSB5 and the composite scores of previous SB edition and all of the major IQ batteries used for all populations.

Norms: The standardization consisted of 4,800 subjects ages 2 to 85+. Care was taken to assure that the sample was as representative of the US population Census 2001. Stratification variables were age, sex, race / ethnicity, geographic region, and socioeconomic level.

Suggested use: The SB5 are used in diagnosis of mental retardation, learning disabilities, and developmental cognitive delays in young children and for placement in academic programs for the intellectually gifted.