ERIC Number: ED339755
Record Type: RIE
Publication Date: 1991-Nov
Reference Count: 0
Four Methods of Handling Missing Data with the 1984 General Social Survey.
Witta, Lea; Kaiser, Javaid
When survey data are statistically analyzed, many times some of the data is missing. If the missing values are not correctly handled, results of the analysis may be dubious and publication may jeopardize the credibility of the organization preparing the report. This study examined four of the more commonly used methods of handling missing data. The following techniques were compared: (1) listwise deletion; (2) pairwise deletion; (3) mean substitution; and (4) regression imputation of missing data. Comparisons were made using a sample selected from the General Social Survey--1984 of the National Opinion Research Center. The sample of 829 cases was randomly divided into two sample groups: Sample 1, with 415 cases; and Sample 2, with 414 cases, which was reduced to only non-missing cases at 283. Sample 1 was used to develop regression equations after treatment by each technique. Sample 2 was used to compare the efficiency of these regression equations in predicting the criterion variable by comparing the actual criterion mean to the predicted mean using Dunnett's test for contrasts. There was a statistically significant difference between the actual mean and the mean predicted by mean substitution with the significance level at 0.01. The other methods exhibited no significant differences. Mean substitution appears inappropriate as a way of handling missing data. A seven-item list of references is included. Three data tables are provided. (SLD)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers: General Social Survey; Missing Data
Note: Paper presented at the Annual Meeting of the Mid-South Educational Research Association (20th, Lexington, KY, November 12-15, 1991).