Methods for handling missing data for multiple-item questionnaires
Publisher
University of Tennessee at Chattanooga
Place of Publication
Chattanooga (Tenn.)
Abstract
Missing data is a common problem, especially in the social and behavioral sciences. Modern missing data methods are underutilized in the industrial/organizational psychology and human resource management literature. This topic has gained increasing attention due to technological advancements in statistical software, although recommendations for handling missing data and default options in software packages often use outdated, suboptimal methods for missing data. Resulting analyses tend to be biased, underpowered, or both. Best practice recommends for the handling of missing data includes the use of multiple imputation (MI) methods, in which missing values are filled in multiple times with predicted values, analyzed, and combined to produce one overall dataset. However this method is often ignored in favor of more convenient methods. For industrial/organizational psychologists, missing data is particularly problematic on multiple-item questionnaires, such as the Counterproductive Work Behavior Checklist (CWB-C). This commonly occurs when participants choose not to respond to a certain item or items on a questionnaire. Person mean imputation is one of the most common methods used to handle missing data on multiple-item questionnaires. This method involves filling in missing data with an average score from the other items for a single dimension and participant. However, it makes strong assumptions about the missing data mechanism and the underlying factor structure of a measure and should be avoided, particularly if data has a high rate of non-response. MI does not make the same assumptions as person mean imputation and may be a superior method when items are missing from a multiple-item questionnaire. Therefore, this study aims to provide recommendations via Monte Carlo simulations with data that is both missing at random (MAR) and missing completely at random (MCAR) for the use of MI methods or person mean imputation when investigating correlations among scale scores for multi-item questionnaires such as the CWB-C.
Date
10-22-2016
Subject
Industrial and organizational psychology
Document Type
posters
Language
English
Rights
http://rightsstatements.org/vocab/InC/1.0/
Methods for handling missing data for multiple-item questionnaires
Missing data is a common problem, especially in the social and behavioral sciences. Modern missing data methods are underutilized in the industrial/organizational psychology and human resource management literature. This topic has gained increasing attention due to technological advancements in statistical software, although recommendations for handling missing data and default options in software packages often use outdated, suboptimal methods for missing data. Resulting analyses tend to be biased, underpowered, or both. Best practice recommends for the handling of missing data includes the use of multiple imputation (MI) methods, in which missing values are filled in multiple times with predicted values, analyzed, and combined to produce one overall dataset. However this method is often ignored in favor of more convenient methods. For industrial/organizational psychologists, missing data is particularly problematic on multiple-item questionnaires, such as the Counterproductive Work Behavior Checklist (CWB-C). This commonly occurs when participants choose not to respond to a certain item or items on a questionnaire. Person mean imputation is one of the most common methods used to handle missing data on multiple-item questionnaires. This method involves filling in missing data with an average score from the other items for a single dimension and participant. However, it makes strong assumptions about the missing data mechanism and the underlying factor structure of a measure and should be avoided, particularly if data has a high rate of non-response. MI does not make the same assumptions as person mean imputation and may be a superior method when items are missing from a multiple-item questionnaire. Therefore, this study aims to provide recommendations via Monte Carlo simulations with data that is both missing at random (MAR) and missing completely at random (MCAR) for the use of MI methods or person mean imputation when investigating correlations among scale scores for multi-item questionnaires such as the CWB-C.
Department
University of Tennessee at Chattanooga. Dept. of Psychology