Effect size measures for longitudinal growth analyses: Extending a framework of multilevel model R-squareds to accommodate heteroscedasticity, autocorrelation, nonlinearity, and alternative centering strategies.
{"title":"Effect size measures for longitudinal growth analyses: Extending a framework of multilevel model R-squareds to accommodate heteroscedasticity, autocorrelation, nonlinearity, and alternative centering strategies.","authors":"Jason D Rights, Sonya K Sterba","doi":"10.1002/cad.20387","DOIUrl":null,"url":null,"abstract":"<p><p>Developmental researchers commonly utilize multilevel models (MLMs) to describe and predict individual differences in change over time. In such growth model applications, researchers have been widely encouraged to supplement reporting of statistical significance with measures of effect size, such as R-squareds (R<sup>2</sup> ) that convey variance explained by terms in the model. An integrative framework for computing R-squareds in MLMs with random intercepts and/or slopes was recently introduced by Rights and Sterba and it subsumed pre-existing MLM R-squareds as special cases. However, this work focused on cross-sectional applications, and hence did not address how the computation and interpretation of MLM R-squareds are affected by modeling considerations typically arising in longitudinal settings: (a) alternative centering choices for time (e.g., centering-at-a-constant vs. person-mean-centering), (b) nonlinear effects of predictors such as time, (c) heteroscedastic level-1 errors and/or (d) autocorrelated level-1 errors. This paper addresses these gaps by extending the Rights and Sterba R-squared framework to longitudinal contexts. We: (a) provide a full framework of total and level-specific R-squared measures for MLMs that utilize any type of centering, and contrast these with Rights and Sterba's measures assuming cluster-mean-centering, (b) explain and derive which measures are applicable for MLMs with nonlinear terms, and extend the R-squared computation to accommodate (c) heteroscedastic and/or (d) autocorrelated errors. Additionally, we show how to use differences in R-squared (ΔR<sup>2</sup> ) measures between growth models (adding, for instance, time-varying covariates as level-1 predictors or time-invariant covariates as level-2 predictors) to obtain effects sizes for individual terms. We provide R software (r2MLMlong) and a running pedagogical example analyzing growth in adolescent self-efficacy to illustrate these methodological developments. With these developments, researchers will have greater ability to consider effect size when analyzing and predicting change using MLMs.</p>","PeriodicalId":47745,"journal":{"name":"New Directions for Child and Adolescent Development","volume":null,"pages":null},"PeriodicalIF":3.4000,"publicationDate":"2021-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1002/cad.20387","citationCount":"14","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"New Directions for Child and Adolescent Development","FirstCategoryId":"102","ListUrlMain":"https://doi.org/10.1002/cad.20387","RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2021/1/29 0:00:00","PubModel":"Epub","JCR":"Q1","JCRName":"PSYCHOLOGY, DEVELOPMENTAL","Score":null,"Total":0}
引用次数: 14
Abstract
Developmental researchers commonly utilize multilevel models (MLMs) to describe and predict individual differences in change over time. In such growth model applications, researchers have been widely encouraged to supplement reporting of statistical significance with measures of effect size, such as R-squareds (R2 ) that convey variance explained by terms in the model. An integrative framework for computing R-squareds in MLMs with random intercepts and/or slopes was recently introduced by Rights and Sterba and it subsumed pre-existing MLM R-squareds as special cases. However, this work focused on cross-sectional applications, and hence did not address how the computation and interpretation of MLM R-squareds are affected by modeling considerations typically arising in longitudinal settings: (a) alternative centering choices for time (e.g., centering-at-a-constant vs. person-mean-centering), (b) nonlinear effects of predictors such as time, (c) heteroscedastic level-1 errors and/or (d) autocorrelated level-1 errors. This paper addresses these gaps by extending the Rights and Sterba R-squared framework to longitudinal contexts. We: (a) provide a full framework of total and level-specific R-squared measures for MLMs that utilize any type of centering, and contrast these with Rights and Sterba's measures assuming cluster-mean-centering, (b) explain and derive which measures are applicable for MLMs with nonlinear terms, and extend the R-squared computation to accommodate (c) heteroscedastic and/or (d) autocorrelated errors. Additionally, we show how to use differences in R-squared (ΔR2 ) measures between growth models (adding, for instance, time-varying covariates as level-1 predictors or time-invariant covariates as level-2 predictors) to obtain effects sizes for individual terms. We provide R software (r2MLMlong) and a running pedagogical example analyzing growth in adolescent self-efficacy to illustrate these methodological developments. With these developments, researchers will have greater ability to consider effect size when analyzing and predicting change using MLMs.
期刊介绍:
The mission of New Directions for Child and Adolescent Development is to provide scientific and scholarly presentations on cutting edge issues and concepts in the field of child and adolescent development. Each issue focuses on a specific new direction or research topic, and is peer reviewed by experts on that topic. Any topic in the domain of child and adolescent development can be the focus of an issue. Topics can include social, cognitive, educational, emotional, biological, neuroscience, health, demographic, economical, and socio-cultural issues that bear on children and youth, as well as issues in research methodology and other domains. Topics that bridge across areas are encouraged, as well as those that are international in focus or deal with under-represented groups. The readership for the journal is primarily students, researchers, scholars, and social servants from fields such as psychology, sociology, education, social work, anthropology, neuroscience, and health. We welcome scholars with diverse methodological and epistemological orientations.