Comparing Methods for Addressing Missingness in Longitudinal Modeling of Panel Data

Abstract Respondent attrition is a common problem in national longitudinal panel surveys. To make full use of the data, weights are provided to account for attrition. Weight adjustments are based on sampling design information and data from the base year; information from subsequent waves is typically not utilized. Alternative methods to address bias from nonresponse are full information maximum likelihood (FIML) or multiple imputation (MI). The effects on bias of growth parameter estimates from using these methods are compared via a simulation study. The results indicate that caution needs to be taken when utilizing panel weights when there is missing data, and to consider methods like FIML and MI, which are not as susceptible to the omission of important auxiliary variables.

[1]  Patricia A. Berglund,et al.  Applied Survey Data Analysis , 2010 .

[2]  L. Kish,et al.  SAMPLING ORGANIZATIONS AND GROUPS OF UNEQUAL SIZES. , 1965, American sociological review.

[3]  Karen Tourangeau,et al.  Early Childhood Longitudinal Study, Kindergarten Class of 1998-99 (ECLS-K). Fifth-Grade Methodology Report. NCES 2006-037. , 2005 .

[4]  Melissa B. Cominole,et al.  2008/12 Baccalaureate and Beyond Longitudinal Study (B&B:08/12). Data File Documentation. NCES 2015-141. , 2015 .

[5]  Sabrina Eberhart,et al.  Applied Missing Data Analysis , 2016 .

[6]  Robert Cudeck,et al.  Conditionally Linear Mixed-Effects Models With Latent Variable Covariates , 1999 .

[7]  Edward A Frongillo,et al.  Food insecurity affects school children's academic performance, weight gain, and social skills. , 2005, The Journal of nutrition.

[8]  Bengt Muthén,et al.  Bayesian Analysis Using Mplus: Technical Implementation , 2010 .

[9]  F. Danner.  A national longitudinal study of the association between hours of TV viewing and the trajectory of BMI growth among US children. , 2008, Journal of pediatric psychology.

[10]  Thanh Le,et al.  Early Childhood Longitudinal Study, Kindergarten Class of 1998-99 (ECLS-K): Combined User's Manual for the ECLS-K Eighth-Grade and K-8 Full Sample Data Files and Electronic Codebooks. NCES 2009-004. , 2009 .

[11]  D. Rubin,et al.  Multiple Imputation for Nonresponse in Surveys , 1989 .

[12]  Lena Osterhagen,et al.  Multiple Imputation For Nonresponse In Surveys , 2016 .

[13]  D. Crawford.  Introduction , 2008, CACM.

[14]  S. Pedlow,et al.  Baccalaureate and Beyond Longitudinal Study: 1993/97 Second Follow-up Methodology Report. Technical Report. , 1996 .

[15]  James E. Rogers,et al.  High School Longitudinal Study of 2009 (HSLS:09) Base Year to First Follow-Up Data File Documentation. NCES 2014-361. , 2013 .

[16]  James L. Arbuckle,et al.  Full Information Estimation in the Presence of Incomplete Data , 1996 .

[17]  A. O'connell,et al.  Growing readers : A hierarchical linear model of children's reading growth during the first 2 years of school , 2006 .

[18]  Rebekah Young,et al.  Methods for Handling Missing Secondary Respondent Data. , 2013, Journal of marriage and the family.

[19]  Jay R. Levinsohn,et al.  National Longitudinal Study: Base Year (1972) through Fourth Follow-Up (1979). Data File Users' Manual. Volume I (Includes Appendix A through Appendix C). , 1981 .

[20]  J. Schafer,et al.  A comparison of inclusive and restrictive strategies in modern missing data procedures. , 2001, Psychological methods.

[21]  David R. Johnson,et al.  Toward best practices in analyzing datasets with missing data: Comparisons and recommendations , 2011 .

[22]  Donald A. Rock,et al.  Early Childhood Longitudinal Study, Kindergarten Class of 1998?99 (ECLS-K). Psychometric Report for the Fifth Grade. NCES 2006?036. , 2005 .

[23]  D. Rubin,et al.  Inference from Iterative Simulation Using Multiple Sequences , 1992 .

[24]  Daniel J. Pratt,et al.  Education Longitudinal Study of 2002 (ELS:2002) Third Follow-up Data File Documentation. NCES 2014-364. , 2014 .

[25]  Craig K. Enders,et al.  Applied Missing Data Analysis , 2010 .

[26]  Michelle Najarian,et al.  Early Childhood Longitudinal Study, Kindergarten Class of 2010-11 (ECLS-K:2011). User's Manual for the ECLS-K:2011 Kindergarten Data File and Electronic Codebook, Public Version. NCES 2015-074. , 2015 .

[27]  Andrew Gelman,et al.  General methods for monitoring convergence of iterative simulations , 1998 .

[28]  P. Lewinsohn,et al.  Heterogeneous trajectories of depressive symptoms: adolescent predictors and adult outcomes. , 2013, Journal of affective disorders.

[29]  Roderick J. A. Little,et al.  Statistical Analysis with Missing Data: Little/Statistical Analysis with Missing Data , 2002 .

[30]  Jacob Cohen.  Statistical Power Analysis for the Behavioral Sciences , 1969, The SAGE Encyclopedia of Research Design.

[31]  Jeanne M. Poduska,et al.  Modeling growth in boys' aggressive behavior across elementary school: links to later criminal involvement, conduct disorder, and antisocial personality disorder. , 2003, Developmental psychology.

[32]  J. Bethlehem.  Weighting nonresponse adjustments based on auxiliary information , 2002 .

[33]  John W Graham,et al.  Planned missing data designs in psychological research. , 2006, Psychological methods.

[34]  Roderick J. A. Little,et al.  Bayes and Multiple Imputation , 2014, Statistical Analysis with Missing Data, Third Edition.

[35]  Ting Yan,et al.  The Relation Between Unit Nonresponse and Item Nonresponse: A Response Continuum Perspective , 2010 .

[36]  Craig K. Enders,et al.  A Primer on Maximum Likelihood Algorithms Available for Use With Missing Data , 2001 .

[37]  R Hardy,et al.  Methods for handling missing data , 2009 .

[38]  G. A. Marcoulides,et al.  Full Information Estimation in the Presence of Incomplete Data , 2013 .

[39]  J. Graham,et al.  Missing data analysis: making it work in the real world. , 2009, Annual review of psychology.