Three-Parameter Item Response Theory Analysis of the Multiple-Choice Items in PIRLS 2016

Mulyono, Mulyono; Widyana, Rahma; Kadir, PUSPA MIRANI; Masruria, Wulida Wahidatul; Baihaqi, M; Setiawan, Engeline Chelsya

doi:10.22034/ijlt.2025.502694.1414

Three-Parameter Item Response Theory Analysis of the Multiple-Choice Items in PIRLS 2016

Document Type : Original Research Article

Authors

Mulyono Mulyono ¹

Rahma Widyana ²

PUSPA MIRANI Kadir ³

Wulida Wahidatul Masruria ⁴

M Baihaqi ⁵

Engeline Chelsya Setiawan ⁶

¹ State University of Surabaya, Indonesia ORCID :https://orcid.org/ 0000000316116617

² Universitas Mercu Buana Yogyakarta, Indonesia

³ Universitas Padjadjaran, Indonesia

⁴ Boston University, US

⁵ UIN Sunan Ampel Surabaya, Indonesia

⁶ Department of Health Policy and Administration, Faculty of Public Health Universitas Airlangga, Surabaya 60115, Indonesia

10.22034/ijlt.2025.502694.1414

Abstract

Progress in International Reading Literacy Study (PIRLS) is an international assessment that measures the reading literacy of fourth-grade students (aged 9-10 years old). PIRLS aims to evaluate and compare the reading abilities of students across different countries. It assesses how well students can understand and interpret written texts, which is fundamental to their overall educational development. In this study, psychometric analyses were run on a portion of the multiple-choice items of PIRLS 2016 taken by 4th graders in the USA. The 3PL item response theory model was utilized to examine the test. Discrimination, difficulty, and guessing parameters were estimated along with the fit values, reliability, item characteristic curves, and item-person map. M2, CFI, TLI, and RMSE statistics showed that the test is reliable and the model, overall, fits the data. Item fit statistics outfit and infit showed that most of the items fit the 3PL model. Findings showed that while all the items have acceptable discrimination values, two items have unacceptable guessing parameters. Examination of the ICCs showed that graphical displays are important, in addition to numerical values, for examining item quality. Item-person map showed that items do not target the whole ability scale.

Keywords

3PL

guessing parameter

ICC

IRT

PIRLS