Search Results

Author: Allison, Paul D.
Resulting in 2 citations.
1. Allison, Paul D.
Fixed Effects Regression Methods In SAS®
Presented: San Francisco, CA, SAS Users Group International Conference (SUGI31), March 2006
Cohort(s): Children of the NLSY79
Publisher: SAS Institute Inc.
Keyword(s): Longitudinal Data Sets; Methods/Methodology; Modeling, Fixed Effects; Statistical Analysis; Statistics

Permission to reprint the abstract has not been received from the publisher.

Fixed effects regression methods are used to analyze longitudinal data with repeated measures on both independent and dependent variables. They have the attractive feature of controlling for all stable characteristics of the individuals, whether measured or not. This is accomplished by using only within-individual variation to estimate the regression coefficients. This paper surveys the wide variety of fixed effects methods and their implementation in SAS, specifically, linear models with PROC GLM, logistic regression models with PROC LOGISTIC, models for count data with PROC GENMOD, and survival models with PROC PHREG.
Bibliography Citation
Allison, Paul D. "Fixed Effects Regression Methods In SAS®." Presented: San Francisco, CA, SAS Users Group International Conference (SUGI31), March 2006.
2. Allison, Paul D.
Missing Data Techniques for Structural Equation Modeling
Journal of Abnormal Psychology 112,4 (November 2003): 545–557.
Cohort(s): Children of the NLSY79
Publisher: American Psychological Association (APA)
Keyword(s): Behavior Problems Index (BPI); Home Observation for Measurement of Environment (HOME); Longitudinal Data Sets; Methods/Methodology; Missing Data/Imputation; Modeling, Growth Curve/Latent Trajectory Analysis; Modeling, Structural Equation; Statistical Analysis

[Editor's Note]: The data set used in this article came from a “1997 SRCD symposium that focused on modern approaches to longitudinal data analysis. The goal of this symposium was to compare and contrast three recently developed methods for analyzing developmental change over time. A single developmental data set was provided to all of the symposium participants with the instructions to analyze the data in any way they wished using a data analytic approach of their own choosing.” “The sample consisted of N=405 children drawn from the Children of the National Longitudinal Survey of Youth, about half of which were missing one or more of the repeated measures on aggression or reading ability.“

As with other statistical methods, missing data often create major problems for the estimation of structural equation models (SEMs). Conventional methods such as listwise or pairwise deletion generally do a poor job of using all the available information. However, structural equation modelers are fortunate that many programs for estimating SEMs now have maximum likelihood methods for handling missing data in an optimal fashion. In addition to maximum likelihood, this article also discusses multiple imputation. This method has statistical properties that are almost as good as those for maximum likelihood and can be applied to a much wider array of models and estimation methods.

Bibliography Citation
Allison, Paul D. "Missing Data Techniques for Structural Equation Modeling ." Journal of Abnormal Psychology 112,4 (November 2003): 545–557.