Moderator Under- and Overspecification in Multiple Regression Analysis

• In Event: Saturday Roundtable Session 3:05 pm
  In Roundtable Session: Regression Modeling for Educational Research (Table 21)

Abstract

Research regarding under- and overspecification within the context of multiple regression models has always focused on situations involving individual predictors. Little research has been done to see if the same results replicate in the cases of under- or overspecifying multiple regression models through the exclusion or inclusion of moderators. The present study conducts a series of Monte Carlo simulations to assess the consequences for biases in parameter estimates, type-I error rate, and power for these under- and overspecified models relative to correctly specified models. Results are discussed in the context of exploratory versus confirmatory research studies.

Authors

• Noah Koehler, Texas A&M University

• Wen Luo, Texas A&M University

• Olukayode E. Apata, Texas A&M University