TitleThe Use of Decision Trees for Adaptive Item Selection and Score Estimation
Publication TypeConference Paper
Year of Publication2011
AuthorsRiley, BB, Funk, R, Dennis, ML, Lennox, RD, Finkelman, M
Conference NameAnnual Conference of the International Association for Computerized Adaptive Testing
Date Published10/2011
Keywordsadaptive item selection, CAT, decision tree

Conducted post-hoc simulations comparing the relative efficiency, and precision of decision trees (using CHAID and CART) vs. IRT-based CAT.

  • Measure: Global Appraisal of Individual Needs (GAIN) Substance Problem Scale (16 items)
  • Past-year symptom count (SPSy)
  • Recency of symptom scale (SPSr)


Decision tree methods were more efficient than CAT

  • CART for dichotomous items (SPSy)
  • CHAID for polytomous items (SPSr)
  • Score bias was low in all conditions, particularly for decision trees using dichotomous items
  • In early stages of administration, decision trees provided slightly higher correlations with the full scale and lower RMSE values.


  • CAT outperformed decision tree methods in later stages of administration.
  • CAT also outperformed decision trees with respect to sensitivity to group differences as measured by effect size.


CAT selects items based on two criteria: Item location relative to current estimate of theta, Item discrimination

Decision Trees select items that best discriminate between groups defined by the total score.

CAT is optimal only when trait level is well estimated.
Findings suggest that combining decision tree followed by CAT item selection may be advantageous.