Methods for Imputing Missing Item Responses in Psychometric Scale Construction

• Activity Number: 212
• Type: Contributed
• Date/Time: Tuesday, August 13, 2002 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section*
• Author(s): Ohidul Siddiqui*+
• Affiliation(s): Food and Drug Administration
• Keywords: missing data ; psychometric scale ; imputation ; non-response

Abstract

In psychiatric clinical trials, therapeutic potential of a new antidepressant drug is open-evaluated based on multi-item psychometric scales. In collecting the patient's response data from such trials, missing responses in some of the items are inevitable and, hence, it is a difficult task to calculate the total score of a psychometric scale. Following, two approaches are commonly used to calculate the total scores of a psychometric scale in presence of missing item responses. The first approach is that if a patient has missing responses in one or more items, his/her total score will be missing. The second approach is that the missing item responses will be imputed before calculating the total score of a scale. For the imputation of missing item responses, different methods are also used. This paper evaluates six imputation methods, which are commonly used in imputing the missing item responses when there are missing responses at one or more items, but not missing responses at more than 50% of the items of a scale. Simulation studies indicate that substituting the mean of the completed item responses of a scale for a given patient is generally the most desirable method for imputing both the random and non-random missing item responses in psychometric scale construction.