z-logo
Premium
Assessing item‐feature effects with item response tree models
Author(s) -
Böckenholt Ulf
Publication year - 2019
Publication title -
british journal of mathematical and statistical psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.157
H-Index - 51
eISSN - 2044-8317
pISSN - 0007-1102
DOI - 10.1111/bmsp.12163
Subject(s) - midpoint , feature (linguistics) , computer science , construct (python library) , item response theory , extension (predicate logic) , tree (set theory) , missing data , class (philosophy) , econometrics , artificial intelligence , machine learning , data mining , statistics , psychometrics , mathematics , mathematical analysis , philosophy , linguistics , geometry , programming language
Recent applications of item response tree models demonstrate that this model class is well suited to detect midpoint and extremity response style effects in both attitudinal and personality measurements. This paper proposes an extension of this approach that goes beyond measuring response styles and allows us to examine item‐feature effects. In a reanalysis of three published data sets, it is shown that the proposed extension captures item‐feature effects across affirmative and reverse‐worded items in a psychological test. These effects are found to affect directional responses but not midpoint and extremity preferences. Moreover, accounting for item‐feature effects substantially improves model fit and interpretation of the construct measurement. The proposed extension can be implemented readily with current software programs that facilitate maximum likelihood estimation of item response models with missing data.

This content is not available in your region!

Continue researching here.

Having issues? You can contact us here