20–22 Nov 2026
The WINC Aichi
Asia/Tokyo timezone

Best Test Design Using Rasch Measurement

21 Nov 2026, 15:05
30m
The WINC Aichi

The WINC Aichi

Research-oriented Presentation (30-minutes) TEVAL: Testing and Evaluation Room 1108

Speakers

David Beglar (Temple University, Japan Campus)Dr Tomoko Nemoto (Temple University, Japan Campus)

Description

In this presentation, we demonstrate how to design high-quality tests characterized by good targeting of test-takers, high reliability, and precise measurement. We focus on three aspects of test design: height, width, and length (Wright & Stone, 1979). Height is the average difficulty of the test items, width indicates the range of item difficulties, and length concerns the number of test items. By adjusting these factors, short, reliable tests can be created.

Abstract section 5: References

Linacre, J. M. (2025). Winsteps (Version 5.10.0) [Computer software]. winsteps.com
Rasch, G. (1960). Probabilistic models for some intelligence and attainment tests. Danmarks Paedogogiske Institute.
Wright, B. D., & Stone, M. H. (1979). Best test design: Rasch measurement. MESA.

Abstract section 3: Content/method

A vocabulary size test was administered to 178 Japanese university students. The data were calibrated with the dichotomous Rasch model (Rasch, 1960), using Winsteps version 5.10.0 (Linacre, 2025). Two tests were created: one highlighting errors in test design and another demonstrating a test designed using the test-design aspects of height, width, and length.

Abstract section 2: Contribution/research questions

  1. Demonstrate mistakes that can be made with test design and the impact of those mistakes on student measurement.
  2. Explain and demonstrate how to use the three aspects of test design—height, width, and length—to develop an effective test instrument.

Abstract section 1: Relevance

Test design is generally not covered in undergraduate or graduate education courses; therefore, many teachers lack a clear understanding of how to create entrance, placement, or proficiency tests. By learning to design tests that account for height, width, and length, educators can develop higher-quality assessments that are more likely to yield fair and accurate results.

Abstract section 4: Outcomes/results

The two tests demonstrate that good test design makes a difference compared to poor design. Based on the three principles of height, width, and length, improvements in reliability, measurement accuracy, test-taker focus, and test length are clearly shown in the well-designed test. Participants will learn how to develop short, effective tests that deliver highly reliable and precise measurements.

Title Best Test Design Using Rasch Measurement
Teaching Context College and university education

Author

David Beglar (Temple University, Japan Campus)

Presentation materials

There are no materials yet.