Test Equating with the Rasch Model to Compare Pre-test and Post-test Measurements

UZUN, Zeynep; Öğretmen, Tuncay

doi:10.21031/epod.957614

Cited by 3 publications

(3 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Each student examined must be based on the same test item, so that changes in students' abilities on the pre-test and post-test can be measured. Measures of student ability on each pre-test and the post-test item can be compared because the data is analyzed in one single measurement but produces an item size for each student and one measure for each item (Laliyo, 2021;Uzun & Öğretmen, 2021).…”

Section: Methodsmentioning

confidence: 99%

Measuring students’ statistical reasoning abilities using flipped classroom model with SPSS and STATCAL

Ramadhani¹,

Evans

2022

J. Honai Math

View full text Add to dashboard Cite

The development of Technology, Information and Communication (ICT) in education has been growing considerably. Use of technologies contribute to supporting student learning and play a vital role in the learning process. The Covid-19 pandemic, since early 2020, has impacted the implementation of the learning process, especially in higher education. The learning instruction has shifted from face-to-face to online learning (e-learning). It is undeniable that the learning process experiences disruptions, resulting in significant changes in the learning environment. Thus, the flipped classroom model is likely to be on model for educators in higher education. This study aims to investigate an increase in the statistical reasoning abilities of engineering students taught using the flipped classroom model assisted by SPSS and STATCAL applications analyzed based on gender. Participants involved 55 first-year informatics engineering students. Data were collected using a test in the form of essay questions and then analyzed using the stacking method of the Rasch measurement model. The results showed that the statistical reasoning abilities of students improved after being taught using the flipped classroom model with the SPSS and STATCAL applications. This finding implies that the flipped classroom model assisted by SPSS and STATCAL applications is effectively implemented in technology-integrated statistics learning.

show abstract

Section: Methodsmentioning

confidence: 99%

Measuring students’ statistical reasoning abilities using flipped classroom model with SPSS and STATCAL

Ramadhani¹,

Evans

2022

J. Honai Math

View full text Add to dashboard Cite

show abstract

“…The IRT model makes the basic assumption that a test's components measure only one ability, a concept known as unidimensionality (Hambleton, et al in Uzun & T, 2021). Because cognitive, personality, and test-taking factors constantly affect test results, at least somewhat, this premise cannot be strictly satisfied.…”

Section: Rasch Modelmentioning

confidence: 99%

Analysis of Mathematics Questions on Theme 9 at Grade v of Raudhatul Rahmah Islamic Elementary School Pekanbaru

Siskowati

Ramadhan

2023

PJR

View full text Add to dashboard Cite

This study aims to evaluate learning through test techniques, which are conducted on 5th grade elementary school students, Mathematics subject. This research is a type of quantitative descriptive research using the ex post facto method. The instrument used in this study was in the form of math subject matter questions consisting of 20 items collected through a written test technique on students. The data obtained were analyzed using the Rasch Model assisted by the Winstep program through an empirical validity process including validity, reliability and difficulty level tests. The results showed that in the validity test, items were invalid, in the reliability test there were 2 types, namely Person Reliability which was categorized as very good, while Item Reliability was categorized as weak. Whereas on the results of the difficulty level of the questions, the category is very difficult. The implication of this research is that this research can be used as a reference for other researchers in conducting research on the preparation and analysis of item items. At the time of designing the questions, it is hoped that a review will be carried out regarding the material that will be discussed in the test.

show abstract

“…Currently, there has never been a study that reveals the equivalence and accuracy problems between the mean/mean method and mean/sigma based on the estimation results between the BILOG program and the R program. There are studies that compared equating methods, but they used the IRTEQ Program [32], [51] or Rasch model [52], [53]. Other studies focused more on comparing equating methods that focused on the ability distribution estimation and sample size [54], comparing equating methods with the use of different test formats [55], investigating the effect of the number of biased items and their distribution in the equated form on equating error [56] and examining the impact of anchor items in the mean/sigma method on true score equating using IRT [57].…”

Section: Introductionmentioning

confidence: 99%

Which one is more accurate, BILOG or R program? (a comparison for score test equating)

Purnama,

Hadi,

Sukirno

et al. 2024

IJERE

View full text Add to dashboard Cite

Evaluation may be carried out using different tests that are not necessarily parallel. Students with lower abilities may get higher scores while those with higher abilities get lower scores. Measurement errors caused by this condition require test equating. Several computer programs, including Bilog and the R program, can be used for test equating. Each program has a different level of accuracy, and the accuracy of the equating results will affect the standard errors of equating. This study aimed to find out the most accurate equating test method and the accuracy of the estimated BILOG and R program. This research used two sets of tests with equivalent group designs. The determination of the most accurate equating method was based on the root mean square deviation (RMSD) value. Equating test packages, A to package B with BILOG program estimation on the mean and Sigma method resulted in RMSD value of 0.320. In the mean and mean method, RMSD value is 0.250. Meanwhile, equating package A to package B using R program on the mean and sigma resulted in RMSD value of 0.300, and the mean and mean method with the RMSD value is 0.272. The mean and mean yield RMSD values smaller than the mean and sigma methods. Therefore, the mean and mean method is more accurate and applicable. Moreover, the estimation results of the BILOG program are more accurate and can be used in test equating.

show abstract

Test Equating with the Rasch Model to Compare Pre-test and Post-test Measurements

Cited by 3 publications

References 10 publications

Measuring students’ statistical reasoning abilities using flipped classroom model with SPSS and STATCAL

Measuring students’ statistical reasoning abilities using flipped classroom model with SPSS and STATCAL

Analysis of Mathematics Questions on Theme 9 at Grade v of Raudhatul Rahmah Islamic Elementary School Pekanbaru

Which one is more accurate, BILOG or R program? (a comparison for score test equating)

Contact Info

Product

Resources

About