Open AccessA SAS MACRO FOR LOGLINEAR SMOOTHING: APPLICATIONS AND IMPLICATIONS
Author(s) -
Moses Tim P.,
Davier Alina A.
Publication year - 2006
Publication title -
ets research report series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.235
H-Index - 5
ISSN - 2330-8516
DOI - 10.1002/j.2333-8504.2006.tb02011.x
Subject(s) - bivariate analysis , univariate , log linear model , equating , macro , smoothing , mathematics , econometrics , statistics , bivariate data , multivariate statistics , linear model , computer science , rasch model , programming language
The two purposes of this paper are to provide a SAS IML macro that performs loglinear smoothing and to apply this macro to loglinear smoothing problems that have not been extensively discussed in the test‐equating literature. The SAS macro is demonstrated on univariate, bivariate, and trivariate smoothing problems. The univariate and bivariate examples reproduce published results (von Davier, Holland, & Thayer, 2004). The trivariate example extends the bivariate smoothing example to allow for comparisons of subgroups' univariate and bivariate distributions. The implications are that important questions about distribution differences and subpopulation invariance of equating functions can be considered through comparisons and evaluations of complex loglinear models that are easily fit with this SAS IML macro.