Off-campus WSU users: To download campus access dissertations, please use the following link to log into our proxy server with your WSU access ID and password, then click the "Off-campus Download" button below.
Non-WSU users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
Access Type
WSU Access
Date of Award
January 2018
Degree Type
Dissertation
Degree Name
Ed.D.
Department
Education Evaluation and Research
First Advisor
Shlomo S. Swilowsky
Abstract
Commercial statistical packages such as SPSS which include multiple imputation(MI) options make replacement of missing values easier. Use of such data handling utilities risks biased data resulting from artificially wide or narrow confidence intervals (CIs) however. Convenience samples of sizes 20, 40, 50, 100, 200, and 500 were drawn from the 2015 Consumer Expenditure Survey (United States Department of Labor, 2016) and evaluated using linear regression. Random deletion at levels of 10%, 20%, 30%, and 40% was performed on 6 of 8 selected variables. MI of 20, 40, 60, and 100 imputations were performed. Comparison of confidence interval widths between intact and imputed data showed an overall pattern of widened CIs in imputed data sets as compared to intact data. Larger sample sizes were significantly associated with narrower CIs. A suggestive association between CIs and number of imputations was found. Thirteen of 24 original data sets returned patterns of significant association between variables. Eight of 24 data sets showed migration either above or below nominal alpha thresholds post imputation, suggesting MI results should not immediately be taken at face value.
Recommended Citation
Kavanagh, Maurice, "The Effect Of Number Of Imputations On Parameter Estimates In Multiple Imputation With Small Sample Sizes" (2018). Wayne State University Dissertations. 2035.
https://digitalcommons.wayne.edu/oa_dissertations/2035