The standardized mean-squared difference (SMSD) has been used for summarizing the bias of parameter estimates in the three-parameter logistic (3PL) model. Due to the indeterminacy problem of the 3PL model, researchers must select a common scale for comparing the theoretical and estimated parameters. The use of different scales can yield noncomparable SMSD values, which in turn can affect the comparison of bias between different parameters. This research used three methods for selecting the common scale. Through a simulation, the three scaling methods were used to numerically demonstrate their effect on SMSD values. Index terms:
Daniels, R.W. (1978). An introduction to numerical methods and optimization techniques. New York: Elsevier North-Holland .
2.
Drasgow, F., & Parsons, C.K. (1983). Application of unidimensional item response theory to multidimensional data. Applied Psychological Measurement , 7, 189-199.
3.
International Mathematical & Statistical Libraries (1982). Reference manual (9th ed.). Houston TX: IMSL, Inc.
4.
Lord, F.M. (1980). Application of item response theory to practical testing problems. Hillsdale NJ: Erlbaum.
5.
Lord, F.M. (1983). Unbiased estimators of ability parameters, of their variance, and of their parallel-form reliability. Psychometrika , 48, 233-245.
6.
Samejima, F. (1984). Comparison of the estimated item parameters of Shiba's word/phrase comprehension tests obtained by LOGIST-5 and those by the tetrachoric method (Res. Rep. 84-2). Knoxville TN: University of Tennessee, Department of Psychology.
7.
Samejima, F. (1986, April). Effect of guessing parameter on the estimation of the item discrimination and difficulty parameters when three-parameter logistic model is assumed. Paper presented at the meeting of the American Educational Research Association, San Francisco CA, U.S.A.
8.
Scales, L.E. (1985). Introduction to non-linear optimization. New York: Springer-Verlag.
9.
Stocking, M.L., & Lord, F.M. (1983). Developing a common metric in item response theory . Applied Psychological Measurement, 7, 201-210.
Yen, W.M. (1984). Obtaining maximum likelihood trait estimates from number-correct scores for the three-parameter logistic model. Journal of Educational Measurement, 21, 91-111.
