東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Multidimensionality and item paramet...

Li, Xin.

FindBook

Google Book

Amazon

博客來

Multidimensionality and item parameter drift: An investigation of linking items in a large -scale certification test.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Multidimensionality and item parameter drift: An investigation of linking items in a large -scale certification test./
作者:	Li, Xin.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2008,
面頁冊數:	120 p.
附註:	Source: Dissertations Abstracts International, Volume: 70-10, Section: A.
Contained By:	Dissertations Abstracts International70-10A.
標題:	Educational tests & measurements. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3331964
ISBN:	9780549837534

Multidimensionality and item parameter drift: An investigation of linking items in a large -scale certification test.
Li, Xin.

Multidimensionality and item parameter drift: An investigation of linking items in a large -scale certification test. - Ann Arbor : ProQuest Dissertations & Theses, 2008 - 120 p.

Source: Dissertations Abstracts International, Volume: 70-10, Section: A.

Thesis (Ph.D.)--Michigan State University, 2008.

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

Common items are widely used to equate test scores of multiple forms or to link test scores at different grade levels. Violations of the assumptions of IRT models may distort the parameter invariance requirement for the use of linking items for repeated use over time. Unidimensionality is one of the primary assumptions. However, linking items are meant to be representative of the whole test and are likely to be sensitive to multiple dimensions if that is the intent of the test. Parameter drift occurs if the statistical properties of items change over time (Goldstein, 1983). In this study, a large-scale certification test dataset was used that included 30 linking items that were administered three times within six years. Simulated responses for these common items were generated from the "true" item parameters from a concurrent calibrated using the data from three years of administration. In addition, samples were randomly selected from the real data. Item parameters were estimated and linked to the same scale using both unidimensional and multidimensional techniques. Compared to the simulated data under the assumption of no drift in item parameters, the property of parameter invariance for the real data samples was examined for different latent trait distributions at different time points. The results confirmed that the potential effect of multidimensionality was associated with the item parameter drift (IPD) for the same set of common items using four different dimensional compositions. Also discussed were limitations for the artificially constructed compositions of actual item response and robustness of IPD detection. Future areas of research are suggested.

ISBN: 9780549837534Subjects--Topical Terms:

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

Item parameter drft

Multidimensionality and item parameter drift: An investigation of linking items in a large -scale certification test.
LDR:02979nmm a2200361 4500 001 2280540
005 20210830065253.5
008 220723s2008 ||||||||||||||||| ||eng d
020 $a 9780549837534
035 $a (MiAaPQ)AAI3331964
035 $a AAI3331964
040 $a MiAaPQ $c MiAaPQ
100 1 $a Li, Xin. $3 999627
245 1 0 $a Multidimensionality and item parameter drift: An investigation of linking items in a large -scale certification test.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2008
300 $a 120 p.
500 $a Source: Dissertations Abstracts International, Volume: 70-10, Section: A.
500 $a Publisher info.: Dissertation/Thesis.
500 $a Advisor: Reckase, Mark D.
502 $a Thesis (Ph.D.)--Michigan State University, 2008.
506 $a This item must not be sold to any third party vendors.
506 $a This item must not be added to any third party search indexes.
520 $a Common items are widely used to equate test scores of multiple forms or to link test scores at different grade levels. Violations of the assumptions of IRT models may distort the parameter invariance requirement for the use of linking items for repeated use over time. Unidimensionality is one of the primary assumptions. However, linking items are meant to be representative of the whole test and are likely to be sensitive to multiple dimensions if that is the intent of the test. Parameter drift occurs if the statistical properties of items change over time (Goldstein, 1983). In this study, a large-scale certification test dataset was used that included 30 linking items that were administered three times within six years. Simulated responses for these common items were generated from the "true" item parameters from a concurrent calibrated using the data from three years of administration. In addition, samples were randomly selected from the real data. Item parameters were estimated and linked to the same scale using both unidimensional and multidimensional techniques. Compared to the simulated data under the assumption of no drift in item parameters, the property of parameter invariance for the real data samples was examined for different latent trait distributions at different time points. The results confirmed that the potential effect of multidimensionality was associated with the item parameter drift (IPD) for the same set of common items using four different dimensional compositions. Also discussed were limitations for the artificially constructed compositions of actual item response and robustness of IPD detection. Future areas of research are suggested.
590 $a School code: 0128.
650 4 $a Educational tests & measurements. $3 3168483
653 $a Item parameter drft
653 $a Item response theory
653 $a Large-scale certification test
653 $a Multidimensionality
690 $a 0288
710 2 $a Michigan State University. $3 676168
773 0 $t Dissertations Abstracts International $g 70-10A.
790 $a 0128
791 $a Ph.D.
792 $a 2008
793 $a English
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3331964