Mitchell Dayton

Learn More
Methodologists have criticized the use of significance tests in the behavioral sciences but have failed to provide alternative data analysis strategies that appeal to applied researchers. For purposes of comparing alternate models for data, information-theoretic measures such as Akaike AIC have advantages in comparison with significance tests.(More)
Title of Dissertation: FACTORS INFLUENCING THE MIXTURE INDEX OF MODEL FIT IN CONTINGENCY TABLES SHOWING INDEPENDENCE Xuemei Pan, Doctor of Philosophy, 2006 Dissertation directed by: Professor C. Mitchell Dayton Department of Measurement, Statistics & Evaluation Traditional methods for evaluating contingency table models based on chi square statistics or(More)
Title of Dissertation: EXAMINING DIFFERENTIAL ITEM FUNCTIONING FROM A LATENT CLASS PERSPECTIVE Karen Samuelsen, Doctor of Philosophy, 2005 Dissertation directed by: C. Mitchell Dayton Department of Measurement, Statistics and Evaluation Current approaches for studying differential item functioning (DIF) using manifest groups are problematic since these(More)
Several competing computational techniques for dealing with sampling zeros were evaluated when estimating the two-point mixture model index, * π , in contingency tables under an independence assumption. Also, the performance of the estimate and associated standard errors were studied under various combinations of conditions. Introduction Traditional methods(More)
  • 1