ISSUES IN THE PRE-ANALYSIS OF PANEL DATA

Abstract An important phase of longitudinal data research is the pre-analysis of the data to determine the nature and extent of biases attributable to missing data. Loss of information takes many forms, but can be classified into the broad categories of loss of item data (i.e. a specific variable for some units of observation) and loss of unit data (i.e. the loss of an entire observation due to refusal or movement out of the sampled population). In this paper we discuss various ways of correcting for identified bias. The methods outlined are separated into those statistical procedures which are suitable in testing and correcting at the aggregate level, those relevant at the individual unit level, and nonstatistical procedures such as unit tracing. Since correction at the unit level is especially important for household panel data that is to be used in econometric modelling, we look in detail at the relationship between imputation and weighting methods of correcting for bias due to attrition.

[1]  Russell S. Winer,et al.  Attrition Bias in Econometric Models Estimated with Panel Data , 1983 .

[2]  Gary G. Koch,et al.  Analyzing Panel Data with Uncontrolled Attrition , 1974 .

[3]  Ryuichi Kitamura,et al.  Analysis of attrition biases and trip reporting errors for panel data , 1987 .

[4]  David A. Hensher,et al.  LONGITUDINAL SURVEYS IN TRANSPORT: AN ASSESSMENT , 1983 .

[5]  David A. Hensher,et al.  Dimensions of Automobile Demand: An Overview of an Australian Research Project , 1986 .

[6]  P. Schmidt,et al.  Limited-Dependent and Qualitative Variables in Econometrics. , 1984 .

[7]  G. Duncan,et al.  Conceptions of longitudinal households: fertile or futile? , 1985 .

[8]  Donald B. Rubin,et al.  Maximum-Likelihood Estimation in Panel Studies with Missing Data , 1980 .

[9]  David A. Hensher An assessment of attrition in a multi wave panel of households , 1989 .

[10]  Neil Wrigley,et al.  The Cardiff Consumer Panel: methodological aspects of the conduct of a long-term panel survey , 1985 .

[11]  J. Hay,et al.  Occupational choice and occupational earnings : selectivity bias in a simultaneous Logit-OLS model , 1980 .

[12]  B. Clarridge,et al.  Tracing Members of a Panel: A 17-Year Follow-Up. , 1978 .

[13]  D. Rubin Formalizing Subjective Notions about the Effect of Nonrespondents in Sample Surveys , 1977 .

[14]  Greg J. Duncan,et al.  Issues of design and analysis of surveys across time , 1987 .

[15]  D. Rubin,et al.  The central role of the propensity score in observational studies for causal effects , 1983 .

[16]  Lung-fei Lee Generalized Econometric Models with Selectivity , 1983 .

[17]  F. Mannering,et al.  A DYNAMIC EMPIRICAL ANALYSIS OF HOUSEHOLD VEHICLE OWNERSHIP AND UTILIZATION , 1985 .

[18]  Frank W. Milthorpe,et al.  Selectivity correction in discrete-continuous choice analysis: With Empirical Evidence for Vehicle Choice and Use , 1987 .