19.01.2013 Views

The Journal of the World Council for Gifted and Talented Children

The Journal of the World Council for Gifted and Talented Children

The Journal of the World Council for Gifted and Talented Children

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

invariance is in fact systemic regarding low-income <strong>and</strong> non low-income students, <strong>the</strong> instrument<br />

would not be appropriate <strong>for</strong> making across-group (low-SES to non-low SES) comparisons. Put<br />

ano<strong>the</strong>r way, if non-invariance was found using <strong>the</strong> NNAT with students from low-income families,<br />

<strong>the</strong> use <strong>of</strong> this measure in identifying students from low-income backgrounds <strong>for</strong> gifted education<br />

programs would be inappropriate given <strong>the</strong> fact that <strong>the</strong> instrument does not measure what its<br />

supposed to measure in <strong>the</strong> same fashion <strong>for</strong> both groups <strong>of</strong> students. As it st<strong>and</strong>s, <strong>the</strong> invariance <strong>of</strong><br />

<strong>the</strong> NNAT with regard to SES groups, as is true with many o<strong>the</strong>r instruments, remains unknown.<br />

Implications<br />

Test authors should evaluate measurement invariance be<strong>for</strong>e an instrument is ever put into<br />

widespread use <strong>and</strong> <strong>the</strong> results from such analyses should be presented in <strong>the</strong> instrument’s<br />

technical manual. If such in<strong>for</strong>mation is not available, school practitioners <strong>and</strong> academic researchers<br />

should be wary <strong>of</strong> using an instrument with diverse groups, as <strong>the</strong>y cannot be sure <strong>the</strong> instrument<br />

measures <strong>the</strong> same thing <strong>for</strong> people in various groups. However, a secondary analysis <strong>of</strong> a test, as<br />

presented with <strong>the</strong> NNAT example, should also be conducted in order to assure previous validity<br />

evidence is still supported over time <strong>and</strong> across new groups.<br />

A look at <strong>the</strong> test manuals <strong>of</strong> larger, more established instruments such as individual<br />

intelligence tests or computer-adaptive achievement tests will reveal detailed sections on how MI/E,<br />

bias, or differential item functioning were evaluated in <strong>the</strong> development phase or in instrument<br />

revision. Interested readers do not have to be too concerned with how to evaluate such in<strong>for</strong>mation<br />

when presented, because it is so rarely presented with regard to instruments commonly used in<br />

gifted <strong>and</strong> talented research. However, even today some measures do include this in<strong>for</strong>mation. For<br />

example, <strong>the</strong> technical manual <strong>of</strong> <strong>the</strong> Scales <strong>for</strong> Identifying <strong>Gifted</strong> Students (Ryser & McConnell,<br />

2004) does present DIF validity evidence across racial/ ethnic <strong>and</strong> gender groups, although no<br />

comparisons were made <strong>for</strong> income groups (Mat<strong>the</strong>ws, 2007). Still, this type <strong>of</strong> reporting should<br />

become st<strong>and</strong>ard in any field where across-group comparisons are common.<br />

Researchers interested in conducting MI/ E from <strong>the</strong> SEM perspective should consult Brown<br />

(2006), who includes a step-by-step process as well as statistical code <strong>for</strong> several computer<br />

programs. A CFA can be run in most statistical programs. LISREL <strong>and</strong> MPlus allow <strong>for</strong> <strong>the</strong> widest<br />

range <strong>of</strong> estimation procedures <strong>and</strong> examination <strong>of</strong> partial (sub-item) MI/E. EQS, AMOS, <strong>and</strong> SAS<br />

will also run CFA analyses, but have a limited variety <strong>of</strong> estimation procedures (although this should<br />

not cause a problem <strong>for</strong> most research). Brown (2006) is an excellent source <strong>for</strong> example code <strong>for</strong> all<br />

<strong>of</strong> <strong>the</strong>se procedures. <strong>The</strong> MPlus technical manual (Muthén & Muthén, 2007) also presents code <strong>for</strong> a<br />

variety <strong>of</strong> different CFA situations.<br />

References<br />

American Education Research Association. (1999). St<strong>and</strong>ards <strong>for</strong> educational <strong>and</strong> psychological testing.<br />

Washington, DC: Author.<br />

Brown, T. A. (2006). Confirmatory factor analysis <strong>for</strong> applied research. New York: <strong>The</strong> Guil<strong>for</strong>d Press.<br />

Carman, C. A., & Taylor, D. K. (2010). Socioeconomic status effects on using <strong>the</strong> Naglieri Nonverbal Ability Test<br />

(NNAT) to identify <strong>the</strong> gifted/talented. <strong>Gifted</strong> Child Quarterly, 54, 75-84.<br />

French, B. F. (2005). [Review <strong>of</strong> Naglieri Nonverbal Ability Test-Individual Administration]. <strong>The</strong> Sixteenth Mental<br />

Measurements Yearbook. Lincoln, NE: Buros Institute.<br />

French, B. F., & Finch, W. H. (2006). Confirmatory factor analytic procedures <strong>for</strong> <strong>the</strong> determination <strong>of</strong><br />

measurement invariance. Structural Equation Modeling, 13, 378-402.<br />

Hu, L., & Bentler, P. M. (1999). Cut<strong>of</strong>f criteria <strong>for</strong> fit indices in covariance structure analysis: Conventional criteria<br />

versus new alternatives. Structural Equation Modeling, 6, 1-55.<br />

Joint Committee on Testing Practices. (2005). Code <strong>of</strong> fair testing practices in education. Washington, DC:<br />

American Psychological Association.<br />

Lord, F. M. (1980). Applications <strong>of</strong> item response <strong>the</strong>ory to practical testing problems. Mahwah, NJ: Erlbaum.<br />

Mat<strong>the</strong>ws, M. S. (2007). <strong>The</strong> Scales <strong>for</strong> Identifying <strong>Gifted</strong> Students by G. R. Ryser & K. McConnell [Test review].<br />

Pp. 729-731 in K. F. Geisinger, R. A. Spies, J. F. Calrson, & B. S. Plake (Eds.), <strong>The</strong> Seventeenth Mental<br />

Measurements Yearbook. Lincoln, NE: <strong>The</strong> Buros Institute <strong>of</strong> Mental Measurements.<br />

Muthén, L. K., & Muthén, B. O. (2007). Mplus User’s Guide. Fifth Edition. Los Angeles, CA: Muthén <strong>and</strong> Muthén.<br />

Naglieri, J. A. (2003). Naglieri Nonverbal Ability Test. San Antonio, TX: Psychological Corporation.<br />

Naglieri, J. A., & Ford, D. Y. (2003). Addressing underrepresentation <strong>of</strong> gifted minority children using <strong>the</strong> Naglieri<br />

Nonverbal Ability Test (NNAT). <strong>Gifted</strong> Child Quarterly, 47, 155-160.<br />

<strong>Gifted</strong> <strong>and</strong> <strong>Talented</strong> International – 26(1), August, 2011; <strong>and</strong> 26(2), December, 2011. 103

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!