*Direct correspondence to David A. Penn, Business and Economic Research Center, Middle Tennessee State University, 1301 E. Main St., Box 102, Murfreesboro, TN 37132 〈email@example.com〉. SAS coding and data will be shared on request. The article benefits from the insightful comments and suggestions offered by the reviewers.
Estimating Missing Values from the General Social Survey: An Application of Multiple Imputation*
Version of Record online: 19 APR 2007
Social Science Quarterly
Volume 88, Issue 2, pages 573–584, June 2007
How to Cite
Penn, D. A. (2007), Estimating Missing Values from the General Social Survey: An Application of Multiple Imputation. Social Science Quarterly, 88: 573–584. doi: 10.1111/j.1540-6237.2007.00472.x
- Issue online: 19 APR 2007
- Version of Record online: 19 APR 2007
Objectives. Most researchers who use survey data must grapple with the problem of how best to handle missing information. This article illustrates multiple imputation, a technique for estimating missing values in a multivariate setting.
Methods. I use multiple imputation to estimate missing income data and update a recent study that examines the influence of parents' standard of living on subjective well-being. Using data from the 1998 General Social Survey, two ordered probit models are estimated: one using complete cases only, and the other replacing missing income data with multiple imputation estimates.
Results. The analysis produces two major findings: (1) parents' standard of living is more important than suggested by the complete cases model, and (2) using multiple imputation can help reduce standard errors.
Conclusions. Multiple imputation allows a researcher to use more of the available data, thereby reducing biases that may occur when observations with missing data are simply deleted.