Educational Statistics for Longitudinal Research

by

Robin K. Henson, Krystal Hinerman

• LAST REVIEWED: 29 July 2020
• LAST MODIFIED: 26 August 2013
• DOI: 10.1093/obo/9780199756810-0083

Introduction

Broadly defined, longitudinal research involves analyzing data collected on the same variables for the same sample of individuals or comparable individuals over two or more time periods, or waves. Longitudinal data analysis may focus on interindividual (between-subjects variation) and/or intraindividual (within-subject variation) comparisons between variables at different time periods. Longitudinal research in education can help address a wide range of research questions involving such topics as describing and defining developmental change, predicting event occurrence, identifying treatment effects, assigning causality, measuring durations between events, and describing occurrences of phenomena over different time periods. In order to appropriately design, conduct, analyze, and interpret longitudinal research, researchers must first and foremost grasp the theoretical underpinnings of the longitudinal research question. It is with theory in mind that all decisions regarding what type of sampling, measurement, time periods, and methodologies are made. The following sections do not pretend to be comprehensive but instead attempt to provide a starting point to help researchers make informed decisions when conducting longitudinal research. The sections include coverage of General Overviews, Longitudinal Methodological and Design Considerations, Factorial Invariance over Time, Attrition and Missing Data, Traditional Methods of Longitudinal Data Analysis, Modeling Growth Curves, Multilevel Modeling, Structural Equation Modeling, Mixture Modeling, categorical and discrete time variables (Analyzing Event Occurrence), and Software Texts. It is important to note the topics are divided into discrete sections for organizational purposes only. The subjects presented often overlap and inform each other.

General Overviews

The texts in this section provide overviews and introductions to major concepts to consider while conducting longitudinal research. Written for the beginning researcher, Taris 2000 provides an abbreviated introduction to the vocabulary and definitions pertaining to longitudinal research, while Anstey and Hofer 2004 offers a somewhat extended introduction to many facets of longitudinal data analysis. For those seeking more in-depth coverage, each of the sections in Laursen, et al. 2012 begins with a chapter explaining the topic’s relevance and then continues with further explanatory chapters. Menard 2007 offers several useful chapters on longitudinal design issues, while Collins and Sayer 2001 includes chapters by several notable authors. Little, et al. 2000 provides practical explanations and examples of applied longitudinal research. In addition to more thorough coverage, Diggle, et al. 2002 includes several references for further reading, while Walls and Schafer 2006 explores the possibilities and intricacies of analyzing data collected over multiple waves.

• Anstey, K. J., and S. M. Hofer. 2004. Longitudinal designs, methods and analysis in psychiatric research. Australian and New Zealand Journal of Psychiatry 38.3: 93–104.

  DOI: 10.1080/j.1440-1614.2004.01343.x

  Written with a psychiatric focus but applicable to educational research. Provides a brief yet encompassing overview of longitudinal designs and methods including introducing and explaining relevant issues such as cross-sectional versus longitudinal data, continuous versus discrete variables, autocorrelation, attrition, and the most prevalent statistical methods for analyzing both longitudinal and cross-sectional data.

• Collins, L. M, and A. G. Sayer, eds. 2001. New methods for analysis of change. Washington, DC: American Psychological Association.

  DOI: 10.1037/10409-000

  A collection of essays by notable authors such as Raudenbush, Muthén, Curan, Bollen, and Kenny; covers relevant topics such as planned missing data, categorical latent variables, factorial invariance, and missingness.

• Diggle, P. J., P. Heagerty, K.-Y. Liang, and S. L. Zeger. 2002. Analysis of longitudinal data. 2d ed. Oxford: Oxford Univ. Press.

  Written mainly for biological and health sciences, provides a comprehensive introduction to the many facets of longitudinal data analysis. Presented in an accessible tone, but offers thorough coverage of topics and offers several references for further reading.

• Laursen, B., T. D. Little, and N. A. Card, eds. 2012. Handbook of developmental research methods. New York: Guilford.

  Collection of chapters pertaining to all aspects of longitudinal research sectioned into seven parts: measurement and design, approaches to data collection, interindividual analysis, intraindividual analysis, combining interindividual and intraindividual analysis, nonindependent data analysis, and special topics. Each section begins with a chapter on foundational issues that overviews relevant concepts within the field.

• Little, T. D., K.-U. Schnabel, and J. Baumert, eds. 2000. Modeling longitudinal and multilevel data: Practical issues, applied approaches, and specific examples. Mahwah, NJ: Lawrence Erlbaum.

  A collection of chapters narrowing the vast array of issues concerning longitudinal data analysis down to the more practical considerations; provides applied examples of modeling longitudinal and multilevel data. Includes working examples in LISREL, EQS, MX, and AMOS.

• Menard, S., ed. 2007. Handbook of longitudinal research: Design, measurement, and analysis. Burlington, MA: Elsevier.

  Written from a sociology perspective, includes several relevant chapters presenting comprehensive coverage of longitudinal data analysis considerations such as respondent recall, minimizing attrition, pooling cross-sectional and time series data, causal analysis with nonexperimental design, and discrete-time survival analysis.

• Taris, T. 2000. A primer in longitudinal data analysis. Thousand Oaks, CA: SAGE.

  Provides an accessible overview of the theory and methodology behind longitudinal data analysis. Includes chapters on design, nonresponse, and stability in addition to discrete-time panel analysis, repeated measures, and analyzing durations and sequences.

• Walls, T. A., and J. L. Schafer. 2006. Models for intensive longitudinal data. New York: Oxford Univ. Press.

  DOI: 10.1093/acprof:oso/9780195173444.001.0001

  Not intended for the beginning user; provides a comprehensive collection of chapters for understanding the intricacies of analyzing data gathered over many occasions. Covers advanced ILD topics such the increased complexity of modeling growth curves over a large number of time points, curve registration, effects reflected in the covariance structure, the importance of time-varying covariates, and autocorrelation.