This report records the efforts undertaken to use imputation to reduce the bias caused by nonresponse in the NELS:88 dataset. First, a simulation study was conducted to evaluate two different imputation procedures (a model-based random imputation method called PROC IMPUTE and a within-class random hot-deck imputation). Second PROC IMPUTE (which performed better in the study) was used to impute missing 2nd follow-up cognitive test scores in four subject areas: math, science, reading, and history/citizenship/geography.
Online Availability
