Modified logistic regression approaches to eliminating the impact of response styles on DIF detection in likert-type scales

Hui Fang CHEN, Kuan Yu JIN, Wen Chung WANG

Research output: Contribution to journalArticlespeer-review

7 Citations (Scopus)

Abstract

Extreme response styles (ERS) is prevalent in Likert- or rating-type data but previous research has not well-addressed their impact on differential item functioning (DIF) assessments. This study aimed to fill in the knowledge gap and examined their influence on the performances of logistic regression (LR) approaches in DIF detections, including the ordinal logistic regression (OLR) and the logistic discriminant functional analysis (LDFA). Results indicated that both the standard OLR and LDFA yielded severely inflated false positive rates as the magnitude of the differences in ERS increased between two groups. This study proposed a class of modified LR approaches to eliminating the ERS effect on DIF assessment. These proposed modifications showed satisfactory control of false positive rates when no DIF items existed and yielded a better control of false positive rates and more accurate true positive rates under DIF conditions than the conventional LR approaches did. In conclusion, the proposed modifications are recommended in survey research when there are multiple group or cultural groups. Copyright © 2017 Chen, Jin and Wang.
Original languageEnglish
Article number1143
JournalFrontiers in Psychology
Volume8
DOIs
Publication statusPublished - Jul 2017

Citation

Chen, H.-F., Jin, K.-Y., & Wang, W.-C. (2017, July). Modified logistic regression approaches to eliminating the impact of response styles on DIF detection in likert-type scales. Frontiers in Psychology, 8, Article 1143. Retrieved July 11, 2017, from http://dx.doi.org/10.3389/fpsyg.2017.01143

Keywords

  • Extreme response styles
  • Logistic regression
  • Likert scale
  • Differential item functioning
  • Mild response style

Fingerprint

Dive into the research topics of 'Modified logistic regression approaches to eliminating the impact of response styles on DIF detection in likert-type scales'. Together they form a unique fingerprint.