東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Effect Size Measures for Differentia...

Feng, Yanan.

Linked to FindBook

Google Book

Amazon

博客來

Effect Size Measures for Differential Item Functioning in Cognitive Diagnostic Models.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Effect Size Measures for Differential Item Functioning in Cognitive Diagnostic Models./
Author:	Feng, Yanan.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2021,
Description:	243 p.
Notes:	Source: Dissertations Abstracts International, Volume: 82-08, Section: B.
Contained By:	Dissertations Abstracts International82-08B.
Subject:	Educational tests & measurements. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28315047
ISBN:	9798569966240

Effect Size Measures for Differential Item Functioning in Cognitive Diagnostic Models.
Feng, Yanan.

Effect Size Measures for Differential Item Functioning in Cognitive Diagnostic Models. - Ann Arbor : ProQuest Dissertations & Theses, 2021 - 243 p.

Source: Dissertations Abstracts International, Volume: 82-08, Section: B.

Thesis (Ph.D.)--Indiana University, 2021.

This item must not be sold to any third party vendors.

This dissertation aims to investigate the effect size measures of differential item functioning (DIF) detection in the context of cognitive diagnostic models (CDMs). A variety of DIF detection techniques have been developed in the context of CDMs. However, most of the DIF detection procedures focus on the null hypothesis significance test. Few studies investigated the practical significance of DIF items. The probability-based effect size indices (signed and unsigned probability difference: SPD and UPD) were proposed in this study. The estimation accuracy of SPD and UPD, as well as the classification of DIF items, were examined under various conditions. The accuracy of the SPD and UPD estimates was largely satisfactory in the studied conditions. The SPD and UPD indices were used together with the Wald test in detecting DIF in the CDM context. The performance of three DIF detection procedures were compared: the Wald method, the Mantel-Haenszel method matched on the total score (MH-T), and the Mantel-Haenszel method matched on the attribute mastery profiles (MH-P). This study also examined the impact of including effect size measures on the Type I error and power rates of DIF detection. The results showed that the Wald method outperformed the two MH procedures in most conditions. The two MH procedures yielded inflated Type I error when group distributions were unequal, and the power rates were also impaired. Including the effect size measures could substantially reduce the Type I error rates of all three methods. In the meantime, including effect size measures also led to a reduction in power for detecting smaller DIF, while the power rates of large DIF were not affected. Moreover, the Trends of International Mathematics and Science Study (TIMSS) 2007 fourth grade mathematics assessment was used to illustrate how the effect size measures and the significance test are combined in detecting DIF. Implications and recommendations for detecting DIF in the CDM context are provided.

ISBN: 9798569966240Subjects--Topical Terms:

3168483
Educational tests & measurements.
Subjects--Index Terms:

Cognitive diagnostic models

Effect Size Measures for Differential Item Functioning in Cognitive Diagnostic Models.
LDR:03197nmm a2200373 4500 001 2285143
005 20211129123950.5
008 220723s2021 ||||||||||||||||| ||eng d
020 $a 9798569966240
035 $a (MiAaPQ)AAI28315047
035 $a AAI28315047
040 $a MiAaPQ $c MiAaPQ
100 1 $a Feng, Yanan. $0 (orcid)0000-0002-8409-8716 $3 3564430
245 1 0 $a Effect Size Measures for Differential Item Functioning in Cognitive Diagnostic Models.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2021
300 $a 243 p.
500 $a Source: Dissertations Abstracts International, Volume: 82-08, Section: B.
500 $a Advisor: Svetina, Dubravka.
502 $a Thesis (Ph.D.)--Indiana University, 2021.
506 $a This item must not be sold to any third party vendors.
520 $a This dissertation aims to investigate the effect size measures of differential item functioning (DIF) detection in the context of cognitive diagnostic models (CDMs). A variety of DIF detection techniques have been developed in the context of CDMs. However, most of the DIF detection procedures focus on the null hypothesis significance test. Few studies investigated the practical significance of DIF items. The probability-based effect size indices (signed and unsigned probability difference: SPD and UPD) were proposed in this study. The estimation accuracy of SPD and UPD, as well as the classification of DIF items, were examined under various conditions. The accuracy of the SPD and UPD estimates was largely satisfactory in the studied conditions. The SPD and UPD indices were used together with the Wald test in detecting DIF in the CDM context. The performance of three DIF detection procedures were compared: the Wald method, the Mantel-Haenszel method matched on the total score (MH-T), and the Mantel-Haenszel method matched on the attribute mastery profiles (MH-P). This study also examined the impact of including effect size measures on the Type I error and power rates of DIF detection. The results showed that the Wald method outperformed the two MH procedures in most conditions. The two MH procedures yielded inflated Type I error when group distributions were unequal, and the power rates were also impaired. Including the effect size measures could substantially reduce the Type I error rates of all three methods. In the meantime, including effect size measures also led to a reduction in power for detecting smaller DIF, while the power rates of large DIF were not affected. Moreover, the Trends of International Mathematics and Science Study (TIMSS) 2007 fourth grade mathematics assessment was used to illustrate how the effect size measures and the significance test are combined in detecting DIF. Implications and recommendations for detecting DIF in the CDM context are provided.
590 $a School code: 0093.
650 4 $a Educational tests & measurements. $3 3168483
650 4 $a Statistics. $3 517247
650 4 $a Cognitive psychology. $3 523881
653 $a Cognitive diagnostic models
653 $a Differential item functioning
653 $a Effect size
653 $a Mantel-Haenszel method
653 $a Wald test
690 $a 0288
690 $a 0463
690 $a 0633
710 2 $a Indiana University. $b School of Education. $3 1020202
773 0 $t Dissertations Abstracts International $g 82-08B.
790 $a 0093
791 $a Ph.D.
792 $a 2021
793 $a English
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28315047