Measuring response styles in likert items

Ulf Böckenholt*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

44 Scopus citations


The recently proposed class of item response tree models provides a flexible framework for modeling multiple response processes. This feature is particularly attractive for understanding how response styles may affect answers to attitudinal questions. Facilitating the disassociation of response styles and attitudinal traits, item response tree models can provide powerful process tests of how different response formats may affect the measurement of substantive traits. In an empirical study, 3 response formats were used to measure the 2-dimensional Personal Need for Structure traits. Different item response tree models are proposed to capture the response styles for each of the response formats. These models show that the response formats give rise to similar trait measures but different response-style effects. Translational Abstract It is well-known that labels of items and their response categories can elicit response processes that may distort the measurement of interest. For example, some respondents may avoid using extreme response categories even if they feel strongly about a topic, whereas other respondents may prefer to select extreme response categories. In this article, I show that item response tree models are well suited to measure response styles as an integral part of a response process. I investigate the effects of response styles in an empirical study that utilizes 3 distinct item formats to measure the same personality construct. When applying item response tree models tailored to the different response formats, I find that 2 of the 3 response conditions give rise to substantial format-specific response style effects. This work illustrates that it may be time to move beyond the view of response styles as nuisance effects and to treat them as psychological phenomena that reveal interesting insights on how respondents differ in their reaction to different questionnaire formats.

Original languageEnglish (US)
Pages (from-to)69-83
Number of pages15
JournalPsychological methods
Issue number1
StatePublished - Mar 1 2017


  • Attitudinal measurement
  • Item formats
  • Item-response tree models
  • Response style

ASJC Scopus subject areas

  • Psychology (miscellaneous)


Dive into the research topics of 'Measuring response styles in likert items'. Together they form a unique fingerprint.

Cite this