Announcements -
check here for new
information!

How legitimate is
the Flynn effect for
the gifted?

Click here for a new report from John Wasserman

Need to download
this article?
Here it is in pdf
format,
How to Use the New
IQ Tests in
Selecting Gifted
Students
 

The new individual intelligence scales have raised many questions for school districts and schools for the gifted, as well as for state legislation, regarding selection of students for gifted programs. Our models for the identification of gifted students originated during simpler times, when one IQ test was popular throughout the United States, and it generated a single IQ score. The cut-off for admission to programs for the gifted was generally two standard deviations above the mean on the Stanford-Binet Intelligence Scale, around 132 IQ. When the Wechsler Intelligence Scale for Children (WISC) won over the American market in the 1960s (Lubin, Wallis & Paine, 1971), some states and districts determined that admission to gifted programs would be based upon Verbal, Performance or Full Scale IQ score in the gifted range (130 or above). Regardless of which IQ test was used, the IQ scores were relatively comparable and measured similar variables.

The new IQ tests have rendered all of these ideas obsolete. Different IQ tests no longer measure the same basic construct of intelligence. Each test identifies a different population as gifted. Gifted students may obtain widely varying scores on different instruments. Full Scale IQ scores are not the unitary constructs they once were and, often, they are not the best representation of the child’s intellectual capacities. The familiar Verbal and Performance IQ scores are gone. There are multiple ways of administering and scoring the new instruments. The designation of giftedness is now unclear. Under these circumstances, how should decisions be made about the use of IQ tests in the selection of students for gifted schools, programs and services?

The following recommendations have been drawn from the Symposium on Assessment Techniques in the Identification of Gifted Learners hosted by the World Council for Gifted Children 16 th Biennial Conference in New Orleans, Louisiana, August 7, 2005; the National Association for Gifted Children Task Force on Assessment, November 2, 2006; chapters in press in The International Handbook on Giftedness and Critical Issues in Equity and Excellence in Gifted Education Series, Volume 2: Alternative Assessment of Gifted Learners, and several studies presented at National Association for Gifted Children Conferences in the last three years. Please review the references for additional information on each recommendation.

Recommendations

1. Individual IQ tests provide better information for high-stakes decision making for gifted students than group tests. Group IQ tests are best used for general screening purposes (Rimm, Gilman & Silverman, in press).
   
2. Gifted students are a special needs population; therefore, comprehensive assessment is needed to determine strengths and weaknesses, information which should be used in their programming (VanTassel-Baska, & Baska, 1993).
   
3. When students are selected for gifted programs on the basis of achievement tests, grades and teacher recommendations, programs are likely to miss children who are economically disadvantaged, who have few books in their homes, and fewer role models for achievement. Individual ability tests provide greater access to programs for diverse cultural groups and children of low socio-economic status (Silverman & Miller, in press).
   
4. The Wechsler Intelligence Scale for Children, Fourth Edition (WISC-IV) and the Stanford-Binet Intelligence Scale, Fifth Edition (SB5) are the two most popular individual IQ tests used in the selection of students for gifted programs. Both tests offer several methods of scoring. The Full Scale IQ scores generated on these tests are not as cohesive measures of general intelligence as in prior editions. When there are extreme discrepancies between composite scores on the WISC-IV (23 points or more), the Full Scale IQ score should not be derived (Flanagan & Kaufman, 2004). Therefore, the Full Scale IQ score should not be the main score used to determine program selection (Rimm, Gilman & Silverman, in press).
   
5. It is not necessary to calculate a Full Scale IQ on the SB5; it is permissible to use either the Verbal IQ or the Nonverbal IQ independently to locate gifted children with different strengths. The highest index, composite or factor score is often the best predictor of success in the gifted program, if the program is responsive to the learning strengths of the students (Silverman, in press, a).
   
6. When using the WISC-IV, either the General Ability Index (GAI), which emphasizes reasoning ability, or the Full Scale IQ Score (FSIQ), should be acceptable for selection to gifted programs.  The GAI should be derived using the table provided by Harcourt Assessments (Technical Report 4).  The Verbal Comprehension Index (VCI) or the Perceptual Reasoning Index (PRI) are also independently appropriate for selection to programs for the gifted, especially for culturally diverse, bilingual, twice-exceptional students or visual-spatial learners. It is important that a good match be made between the strengths of the child and the attributes of the program.  Students who have special learning needs should be admitted to gifted programs, provided that there are other indications of giftedness and instructional modifications are made to fit the needs of the students (NAGC Task Force on Assessment, 2006).
   
7. Processing Speed and Working Memory often lower composite and Full Scale IQ scores for the gifted. The 63 gifted children in the norm sample reported in the WISC-IV Technical Manual had a mean Working Memory score of 112.5 and a mean Processing Speed score of 110.6 (Wechsler, 2003). Only 4 of 103 gifted children tested at the Gifted Development Center scored 130 or above on Processing Speed (Silverman, Gilman & Falk, 2004). The Family Achievement Center found Processing Speed to be in the high average range for 42 children (Rimm, Gilman & Silverman, in press). In another study, Wasserman (2006) reported that “over 70% of the students applying for gifted placement have Processing Speed Index scores in the average range or below” (Wasserman, 2006, p. 2). On the WISC-IV, the GAI omits both Processing Speed and Working Memory in the scoring. In using the SB5, a Gifted Composite Score and a Gifted Nonverbal Composite score can be generated in which Working Memory is eliminated (Silverman, in press, a).
   
8. Funds for assessing gifted students are usually limited. For this reason, many districts employ short forms of individual intelligence scales, such as the Wechsler Abbreviated Scale of Intelligence (WASI) (Wechsler, 1999). A reasonable alternative would be to administer only the six core subtests of the WISC-IV from which the General Ability Index (GAI) can be derived: Vocabulary, Similarities, Comprehension, Block Design, Matrix Reasoning and Picture Concepts (Silverman, Gilman & Falk, 2004). As most of these subtests are richly loaded in general intelligence (g), they are likely to locate the students who would be most successful in a gifted program.
   
9. The gifted validation sample reported in the Technical Manual of the WISC-IV achieved a mean Full Scale IQ score of 123.5 (Wechsler, 2003). The mean IQ score of 202 children in the gifted validation sample of the SB5 was 124. Therefore, cut-off scores for gifted programs should be lowered to 120, rather than 130 (Rimm, Gilman & Silverman, in press; Silverman, in press, a).
   
10. In selecting an instrument to use for assessing the gifted, it is necessary to keep in mind that the WISC-IV and the SB5 identify different students as gifted. As 30% of the WISC-IV measures abstract verbal reasoning, compared to 10% of the SB5, the WISC-IV is likely to find more highly verbal children. As 20% of the SB5 measures mathematically gifted children, compared to 0 – 10% of the WISC-IV (depending if Arithmetic, an optional test, is administered), the SB5 is likely to find more mathematically gifted children. While 20% of each test is devoted to the measurement of visual-spatial abilities, there may be more visual-spatial content in the SB5, so it may be preferable for locating spatially gifted children.
    
11. Those instruments or portions of instruments with the richest loadings on general intelligence (g) are the most useful for locating gifted children. Raven’s Progressive Matrices, the Stanford-Binet scales and the Wechsler scales were all founded on the conception of intelligence as abstract reasoning (g). Abstract reasoning and general intelligence (g)are synonymous. Giftedness is high abstract reasoning. Therefore, g could as easily stand for giftedness as for general intelligence (Silverman, in press, a).

g-loadings on the WISC-IV

Good Measures of g

Vocabulary (.82)
(Information) (.79)
Similarities (.79)
(Arithmetic) (.74)
(Word Reasoning) (.70)
Comprehension (.70)

Fair Measures of g

Matrix Reasoning (.68)
Block Design (.67)
(Picture Completion) (.63)
Letter-Number Sequencing (.60)
Symbol Search (.58)
Picture Concepts (.57)
Digit Span (.51)

Poor Measures of g

Coding (.48)
(Cancellation) (.25)

*Items in parentheses are optional tests.
(Flanagan & Kaufman, 2004, p. 309)

Recommendations for Differentiating Children at Higher Levels of Intelligence

12. Exceptionally gifted children are among the highest risk gifted populations (Rimm, Gilman & Silverman, in press). An international group determined nomenclature for the higher IQ levels. Their results were published in a chapter on “Assessment of Intellectual Functioning” by John Wasserman (2003).

Table 5
Levels of Giftedness

Profoundly Gifted

Exceptionally Gifted

Highly Gifted

Gifted

above 175

160-174

145-159

130-144

+5 SD

+4 SD

+3 SD

+2 SD

13. The low ceilings on both the WISC-IV and the SB5 make it difficult to locate highly, exceptionally and profoundly gifted children. To document that the child’s abilities exceed the measuring tool, Betty Meckstroth created a method of tracking the number of raw score points earned beyond the minimum score required to attain a 19 (the highest possible subtest score). Children have been found who scored 13 raw score points beyond the ceiling on Vocabulary (19 + 13) and 8 extra points on Similarities (19 + 8) on the WISC-IV (Rimm, Gilman & Silverman, in press). It is obvious that the abilities of such children are not fully tapped on the WISC-IV.

14. As the distributions at the extremes do not follow the normal curve (J. C. Raven, 1959; J. Raven, 1983; Terman, 1925; Wechsler, 1939), derivation of ratio IQs is needed at the extremes. Ratio IQ scores are growing in popularity, even in nonverbal tests (Wasserman, 2005).

15. The Interpretive Manual of the SB5 (Roid, 2003) offers a table of Extended IQ scores for children who score above 150 IQ or below 40 IQ. Scores range as low as 10 and as high as 225 IQ. Based on Rasch scoring, the examinee is credited with all raw score points beyond the requirement to obtain the ceiling score of 19. As few score above 150, Rasch-Ratio scores may also be derived by hand for students who score 130 and above on the SB5 (Carson & Roid, 2004). It is recommended that students be selected for gifted programs who attain a score of 120 or above on the SB5, and that Rasch-Ratio scores be derived to qualify students for services for the highly and exceptionally gifted or to determine the degree of acceleration needed (Silverman, in press, a).

16. Rimm Ratios can be derived by utilizing test-age equivalents for subtests provided in the WISC-IV manual (p. 253), converted into months, to determine a child’s mental age. Age-equivalent scores reflect all correct items for each subtest. The child’s mental age is divided by the chronological age and multiplied by 100 to derive a Rimm Ratio (Rimm, Gilman & Silverman, in press). This method is most appropriate for children 10 and under, due to the low ceiling of the test-age equivalents.

17. The two-step process employed in the Talent Searches for differentiating the most able students at 12 or 13 years of age is the best model for locating exceptionally gifted children. Talent Searches provide out-of-level testing to children who score at or above the 95 th or 97 th percentile in reading and mathematics. Since 1989, younger children in these ranges have been found by using a combination of two different measures: one comparing them with others their own age and one with a higher ceiling, like the SAT, that compares their abilities to those of older children. Because it is organized by age levels, with increasing levels of difficulty, all the way to Superior Adult III, the “Binet-type age scale might be considered the original examination suitable for extensive out-of-level testing” (Stanley, 1990, p. 167).

18. Examiners who assess the exceptionally gifted offer out-of-level testing to children who achieve on a standard IQ test (e.g., the WISC-IV, WPPSI-III, SB5, DAS-2, KABC-2, etc.) at or above the 99 th percentile on at least two subtests (particularly those subtests that are good measures of g). In the case of the Raven’s Progressive Matrices, a score at the 97 th percentile would probably warrant an out-of-level supplementary test. The Stanford-Binet Intelligence Scale (Form L-M) is given purely as a supplemental measure to test the limits of children’s abilities when they achieve ceiling-level scores on tests with lower ceilings (Silverman & Kearney, 1989; see also, Silverman & Kearney, 1992; Wasserman, in press). When the child exceeds the scores in the norm table of the SBL-M, a formula IQ is derived according to the instructions on page 339 in the manual (Terman & Merrill, 1973). The formula IQ is a ratio metric.

19. It is permissible to use the SBL-M as a supplemental test, as long as examiners acknowledge that the scores are on a different metric and, therefore, not comparable to deviation IQs (Carson & Roid, 2004). The publisher prefers that the SBL-M be co-administered with the SB5 so that three types of composite scores might be contrasted: standard scores, Rasch-Ratio scores, and SBL-M scores (Carson & Roid, 2004).

20. The Flynn Effect is the most frequently cited reason for not using a test with older norms. Newer studies suggest that the Flynn effect may be have tapered off at the beginning of the 1990s (Teasdale & Owen, 2005). John Wasserman (2007) recently studied the Flynn Effect and writes:

My January, 2007 examination of psychological research databases suggests that the Flynn effect has not yet been adequately demonstrated for all levels of ability … there is no substantive evidence for its validity with high ability individuals (particularly those who are intellectually gifted). … I have yet to see any sound empirical studies of the Flynn effect in gifted samples. (p. 1)

For more information on the recommendations listed above, please see [(Rimm, Gilman & Silverman, in press; Silverman, in press (a, b); Silverman & Miller, in press; Wasserman, in press.]

*Silverman, L. K. (in press). The measurement of giftedness. In L. Shavinina (Ed.), The International Handbook on Giftedness. Amsterdam: Springer Science.

References

Carson, D. & Roid, G.  (2004).  Acceptable use of the Stanford-Binet Form L-M:  Guidelines for the professional use of the Stanford-Binet Intelligence Scale, Third Edition (Form L-M).  Itasca, IL:  Riverside Publishing.  

Flanagan, D. P., & Kaufman, A. S. (2004). Essentials of WISC-IV assessment. Hoboken, NJ: John Wiley & Sons.

Lubin, B., Wallis, R. R., & Paine, C. (1971). Patterns of psychological test usage in the United States: 1935 – 1969. Professional Psychology, 2, 70-74.

National Association for Gifted Children (NAGC) Task Force on Assessment. (2006, November). Minutes of the meeting. National Association for Gifted Children, Charlotte, NC.

Raven, J. C. (1959). A note on Burt’s “The distribution of intelligence.” British Journal of Psychology, 47, 95-100.

Raven, J. (1983). Raven Coloured Progressive Matrices Intelligence Test in Thailand and in Denmark: A response. School Psychology International, 4, 173-176.

Rimm, S., Gilman, B. J., & Silverman, L. K. (in press). Non-traditional applications of traditional testing. In J. VanTassel-Baska (Ed.), Critical issues in equity and excellence in gifted education series,Volume 2: Alternative assessment of gifted learners. Waco, TX: Prufrock Press.

Roid, G. H. (2003). Stanford-Binet Intelligence Scales interpretive manual: Expanded guide to the interpretation of SB5 test results. Itasca, IL: Riverside.

Silverman, L. K. (in press, a). The measurement of giftedness. In L. Shavinina (Ed.), The International Handbook on Giftedness. Amsterdam: Springer Science.

Silverman, L. K. (in press, b). A new era in identification of the gifted. Gifted Education Communicator, 8(1).

Silverman, L. K., Gilman, B. J., & Falk, R. F. (2004, November). Who are the gifted using the new WISC-IV? Paper presented at the 51 st annual convention of the National Association for Gifted Children, Salt Lake City, UT

Silverman, L. K., & Kearney, K. (1989). Parents of the extraordinarily gifted. Advanced Development, 1, 41-56.

Silverman, L .K., & Kearney, K. (1992). The case for the Stanford-Binet L-M as a supplemental test. Roeper Review, 15, 34-37.

Silverman, L. K. & Miller, N. B. (in press). A feminine perspective of giftedness. In L. Shavinina (Ed.), The International Handbook on Giftedness. Amsterdam: Springer Science.

Stanley, J. C. (1990). Leta Hollingworth’s contributions to above-level testing of the gifted. Roeper Review, 12(3), 166-171.

Teasdale, T.W., & Owen. D.R. (2005). A long-term rise and recent decline in intelligence test performance: The Flynn Effect in reverse. Personality and Individual Differences, 39, 837 – 843.

Terman, L. M. (1925). Genetic studies of genius, Vol. 1: Mental and physical traits of a thousand gifted children. Stanford, CA: Stanford University Press.

Terman, L. M., & Merrill, M.A. (1973). The Stanford -Binet Intelligence Scale: 1973 norms edition. . Boston: Houghton Mifflin.

VanTassel-Baska, J. & Baska, L. (1993). The role of educational personnel in counseling the gifted. In L. Silverman (Ed.) Counseling the gifted & talented. Denver: Love.

Wasserman, J. (2003). Assessment of intellectual functioning. In J. R. Graham & J. A. Naglieri (Eds.), Handbook of psychology, Volume 10: Assessment psychology (pp. 417-442). Hoboken, NJ: Wiley.

Wasserman, J. (2005, November). Utility of nonverbal measures with the gifted. Paper presented at the 16 th Biennial Conference of the World Council for Gifted and Talented Children, New Orlean, LA.

Wasserman, J. (2006). Tips for parents: Intellectual assessment of exceptionally and profoundly gifted children. Unpublished manuscript available online at
http://www.gt-cybersource.org/Record.aspx?NavID=2_0&rid=14107

Wasserman, J. D. (2007). The Flynn effect in gifted samples: Status as of 2007. Unpublished manuscript available here.

Wasserman, J. (in press). Intellectual assessment of exceptionally and profoundly gifted children. In K. Kay, D. Robson, & J. F. Brenneman (Eds.), High IQ kids: A manual for adults who care. Minneapolis, MN: Free Spirit.

Wechsler, D. (1999). Wechsler Abbreviated Scale of Intelligence. San Antonio, TX: The Psychological Corporation.

Wechsler, D. (2003). The WISC-IV technical and interpretive manual. San Antonio, TX: Psychological Corporation.