Handling attrition in longitudinal studies: The case for refreshment samples

  title={Handling attrition in longitudinal studies: The case for refreshment samples},
  author={Yiting Deng and D. Sunshine Hillygus and Jerome P. Reiter and Yajuan Si and Siyu Zheng},
  journal={Quality Engineering},
Panel studies typically suffer from attrition, which reduces sample size and can result in biased inferences. It is impossible to know whether or not the attrition causes bias from the observed panel data alone. Refreshment samples—new, randomly sampled respondents given the questionnaire at the same time as a subsequent wave of the panel—offer information that can be used to diagnose and adjust for bias due to attrition. We review and bolster the case for the use of refreshment samples in… 

Figures and Tables from this paper

Semi-parametric Selection Models for Potentially Non-ignorable Attrition in Panel Studies with Refreshment Samples

A Bayesian approach to handle attrition in two-wave panels with one refreshment sample and many categorical survey variables is presented and a model to correct attrition bias is applied in an analysis of data from the 2007–08 Associated Press/Yahoo News election panel study.

Bayesian Latent Pattern Mixture Models for Handling Attrition in Panel Studies With Refreshment Samples

A Bayesian latent pattern mixture model is presented, in which an indicator for attrition and the survey variables are modeled jointly via a latent class model, which allows the multinomial probabilities within classes to depend on the attrition indicator.

Bias Correction for Replacement Samples in Longitudinal Research

Four ways to correct the bias introduced by replacement samples are proposed and evaluated: a parametric bootstrapping replacement sample correction, a non-parametricBootstrapping Replacement Sample Correction, a primary inverse probability reweighting correction, and a likelihood-based inverse probability Reweighting Correction.

Evaluation of supplemental samples in longitudinal research with non-normal missing data

Simulation results suggest that the addition of refreshment samples, but not replacement samples, is an effective way to respond to attrition in longitudinal research.

Evaluating Supplemental Samples in Longitudinal Research: Replacement and Refreshment Approaches

Analysis of how researchers may supplement the sample with additional participants suggests that refreshment samples lead to less relative bias, greater relative efficiency, and more acceptable coverage rates than replacement samples or not supplementing the missing participants in any way.

A Panel Quantile Approach to Attrition Bias in Big Data: Evidence from a Randomized Experiment

A mixed-effects estimating equation approach to nonignorable missing longitudinal data with Refreshment samples

Nonignorable missing data occur frequently in longitudinal studies and can cause biased estimations. Refreshment samples which draw new subjects randomly in subsequent waves from the original

Structural equation modelling with complex sampling designs and non-random attrition: A tutorial using lavaan and Mplus

Using an auto-regressive latent trajectory model with structured residuals (ALT-SR) to model the relations between relationship satisfaction and mental health in the Understanding Society study as an example is used and guidance on implementation of this approach in both R and Mplus is provided.

Dealing with Attrition and Missing Data in Longitudinal Studies: A Critique

Longitudinal studies in teenage pregnancy, like any other topical issue, have been greatly affected by the problem of participant drop out and missing data. Although existing evidence indicate that

Looking Beyond Demographics: Panel Attrition in the ANES and GSS

Longitudinal or panel surveys offer unique benefits for social science research, but they typically suffer from attrition, which reduces sample size and can result in biased inferences. Previous



Combining Panel Data Sets with Attrition and Refreshment Samples

The manner in which the presence of refreshment samples allows the researcher to test various models for attrition in panel data, including models based on the assumption that missing data are missing at random are described.

Methods to Account for Attrition in Longitudinal Data: Do They Work? A Simulation Study

A cohort of 300 subjects is simulated using 500 computer replications to determine whether regression imputation, individual weighting, or multiple imputation is useful to reduce attrition bias, and results are compared to a complete subject analysis.

Panel Conditioning and Attrition in the AP-Yahoo! News Election Panel Study

This paper examines the potential for attrition bias and panel conditioning in a longitudinal online election panel survey. The 2008 Associated Press-Yahoo! News Poll was conducted by Knowledge

Inference in panel data models under attrition caused by unobservables

Simple estimators for nonparametric panel data models with sample attrition

Adjusting for non-ignorable sample attrition using survey substitutes identified by propensity score matching: an empirical investigation using labour market data

This article assesses the potential for reducing attrition bias by replacing survey dropouts with substitutes drawn from the same population and identified using propensity score matching. By linking

Using Analysis of Gini (ANOGI) for Detecting Whether Two Subsamples Represent the Same Universe

A wildly discussed shortcoming of panel surveys is a potential bias arising from selective attrition. Based on data of the German Socio-Economic Panel Study (SOEP), the authors analyze potential

Attrition in Panel Data: The Effectiveness of Weighting

Although weighting seems to be a popular strategy for coping with panel attrition in sociological research it is not clear how effective it is in reducing attrition bias. This article aims to fill

The Multiple Adaptations of Multiple Imputation

Some of the main adaptations of the multiple-imputation framework, including missing data in large and small samples, data confidentiality, and measurement error, are described, and the combining rules for each setting are reviewed and explained.

Weighted Generalized Estimating Functions for Longitudinal Response and Covariate Data That Are Missing at Random

Longitudinal studies often feature incomplete response and covariate data. It is well known that biases can arise from naive analyses of available data, but the precise impact of incomplete data