Outliers in Questionnaire Data: Can They Be Detected and Should They Be Removed?

Outliers in questionnaire data are unusual observations, which may bias statistical results, and outlier statistics may be used to detect such outliers. The authors investigated the effect outliers have on the specificity and the sensitivity of each of six different outlier statistics. The Mahalanob...

Full description

Saved in:
Bibliographic Details
Published inJournal of Educational and Behavioral Statistics Vol. 36; no. 2; pp. 186 - 212
Main Authors Zijlstra, Wobbe P., van der Ark, L. Andries, Sijtsma, Klaas
Format Journal Article Book Review
LanguageEnglish
Published Los Angeles, CA SAGE Publications 01.04.2011
American Educational Research Association
Subjects
Bias
Coefficients
Data
Datasets
Methods
Observational research
Outliers
Proportions
Questionnaires
Rating Scales
Sampling bias
Standard deviation
Statistical Bias
Statistical variance
Statistics
Validity
outlier detection in questionnaire data
outlier detection methods for questionnaire data
contaminant observations
discordancy testing
suspected patterns of item scores
Online AccessGet full text

Cover

More Information
Summary:Outliers in questionnaire data are unusual observations, which may bias statistical results, and outlier statistics may be used to detect such outliers. The authors investigated the effect outliers have on the specificity and the sensitivity of each of six different outlier statistics. The Mahalanobis distance and the item-pair based outlier statistics were found to have the best combination of specificity and sensitivity. Next, it was investigated how outliers influenced the bias in the percentile rank score, Cronbach's alpha, and the validity coefficient. Outliers due to random responding and faking produced considerable bias, and outliers due to extreme responding produced little bias. Finally, the influence of removing discordant observations on bias was studied. Removing observations due to random responding identified by means of the Mahalanobis distance, the local outlier factor, and the item-pair based outlier statistic reduced bias.
ISSN:1076-9986
1935-1054
DOI:10.3102/1076998610366263