International Conference on Assessment and Learning (ICAL)

Publication Date

10-2022

Download

Download Full Text (691 KB)

Keywords

Behaviour disorders, Measurement, Response style (Tests)

Subjects

Behaviour disorders, Measurement, Response style (Tests)

Disciplines

Applied Behavior Analysis | Disability and Equity in Education

Comments

Paper presented at ICAL 2022 — Transforming Assessment and Learning: Making the System Work!

13 - 15 October 2022, Bali, Indonesia

Description

This study aims to determine whether there is a difference in detection rates between response model and response time model to detect aberrant behavior. Also, to determine examinee’s ability estimation accuracy of each model along with checking the strengths and weakness when there is an aberrant behavior in testing data especially CBT. This research is a simulation study where concentrate on test-length, sample-size, and aberrant level with 50 replications. Analyzing parameter recovery from replicated data to check the strengths and weaknesses of each model. Further, comparing the lz person-fit that using response’s data and response time’s data to see which one the most sensitive is from both models. Moreover, the estimation of examinee’s ability also compared to see the accuracy of each model. The result shows that increment of aberrant level would make item parameter estimation more bias for both models. Detection rates using response time (or response time model) indicates more sensitivity than detection rates using response (or response model). Both models also showed there was an increment estimation for examinee who doing aberrant behavior, so becomes bias and invalid if used as decision-making.

City

Jakarta, Indonesia

Publisher

ACER Indonesia

ISBN

978-1-74286-697-0

DOI

https://doi.org/10.37517/978-1-74286-697-0-14

Simulation of comparing the sensitivity between response model and response time model detecting aberrant behavior

Share

 
COinS