Title: Loglinear models as a general framework for stochastic curtailment

Author: Niels Smits

Affiliation: VU University Amsterdam

Abstract: Stochastic Curtailment (SC, Finkelman, Smits, Kim, & Riley, 2012;
Fokkema, Smits, Finkelman, Kelderman, & Cuijpers, 2014) is a new method for
efficient computerized testing for classification into two classes.  It has
been noted that its construction phase is rather laborious; the main issue
is dealing with structural zeros and dichotomizing total scores.  In this
paper a solution is sought by using loglinear models under
quasi-independence.  This approach uses a single model for the observed
cross-tabulations of cumulative and total scores, resulting in expected
cell probabilities under the model.  These probabilities may then be used
as an input for determining early stopping rules, either for classification
or obtaining total scores.

References

Finkelman, M. D., Smits, N., Kim, W., & Riley, B. (2012). Curtailment and
stochastic curtailment to shorten the CES-D. Applied Psychological
Measurement, 36, 632-658.

Fokkema, M., Smits, N., Finkelman, M. D., Kelderman, H., & Cuijpers, P.
(2014). Curtailment: A method to reduce the length of self-report ques-
tionnaires while maintaining diagnostic accuracy. Psychiatry Research,
215 (2), 477-482.