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

Estimation and psychometric analysis...

Grochowalski, Joseph H.

FindBook

Google Book

Amazon

博客來

Estimation and psychometric analysis of component profile scores via multivariate generalizability theory.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Estimation and psychometric analysis of component profile scores via multivariate generalizability theory./
作者:	Grochowalski, Joseph H.
面頁冊數:	214 p.
附註:	Source: Dissertation Abstracts International, Volume: 77-07(E), Section: B.
Contained By:	Dissertation Abstracts International77-07B(E).
標題:	Quantitative psychology. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10013383
ISBN:	9781339479316

Estimation and psychometric analysis of component profile scores via multivariate generalizability theory.
Grochowalski, Joseph H.

Estimation and psychometric analysis of component profile scores via multivariate generalizability theory. - 214 p.

Source: Dissertation Abstracts International, Volume: 77-07(E), Section: B.

Thesis (Ph.D.)--Fordham University, 2015.

Component Universe Score Profile analysis (CUSP) is introduced in this paper as a psychometric alternative to multivariate profile analysis. The theoretical foundations of CUSP analysis are reviewed, which include multivariate generalizability theory and constrained principal components analysis. Because CUSP is a combination of generalizability theory and principal components analysis, CUSP was analyzed for accuracy and precision for small sample sizes via a simulation analysis. The stability and accuracy of CUSP analysis was also assessed using a large sample of examinee data (n=5,000) from the College Board Advanced Placement (AP) Statistics subject test. The simulation showed that CUSP reliability estimates and coordinate estimates are generally unbiased when sample sizes were larger than n=100. Subtest covariance structure caused significant bias in reliability and coordinate estimates when the covariance matrix was compound symmetric. Coordinate standard error estimates were significantly biased when sample sizes were very small (n<100). The AP data analysis, like the simulation, showed that profile estimates were consistent and stable for larger samples, but profile scores were inconsistent for the small sample condition (n=50). CUSP external analysis illustrated the use of an external variable (gender) to define and predict AP Statistics profiles. The estimated external gender profile was not useful for classifying examinees by gender, achieving accuracy approximately equal to chance assignment.

ISBN: 9781339479316Subjects--Topical Terms:

2144748
Quantitative psychology.

Estimation and psychometric analysis of component profile scores via multivariate generalizability theory.
LDR:02455nmm a2200289 4500 001 2070585
005 20160617121819.5
008 170521s2015 ||||||||||||||||| ||eng d
020 $a 9781339479316
035 $a (MiAaPQ)AAI10013383
035 $a AAI10013383
040 $a MiAaPQ $c MiAaPQ
100 1 $a Grochowalski, Joseph H. $3 3185644
245 1 0 $a Estimation and psychometric analysis of component profile scores via multivariate generalizability theory.
300 $a 214 p.
500 $a Source: Dissertation Abstracts International, Volume: 77-07(E), Section: B.
500 $a Adviser: Se-Kang Kim.
502 $a Thesis (Ph.D.)--Fordham University, 2015.
520 $a Component Universe Score Profile analysis (CUSP) is introduced in this paper as a psychometric alternative to multivariate profile analysis. The theoretical foundations of CUSP analysis are reviewed, which include multivariate generalizability theory and constrained principal components analysis. Because CUSP is a combination of generalizability theory and principal components analysis, CUSP was analyzed for accuracy and precision for small sample sizes via a simulation analysis. The stability and accuracy of CUSP analysis was also assessed using a large sample of examinee data (n=5,000) from the College Board Advanced Placement (AP) Statistics subject test. The simulation showed that CUSP reliability estimates and coordinate estimates are generally unbiased when sample sizes were larger than n=100. Subtest covariance structure caused significant bias in reliability and coordinate estimates when the covariance matrix was compound symmetric. Coordinate standard error estimates were significantly biased when sample sizes were very small (n<100). The AP data analysis, like the simulation, showed that profile estimates were consistent and stable for larger samples, but profile scores were inconsistent for the small sample condition (n=50). CUSP external analysis illustrated the use of an external variable (gender) to define and predict AP Statistics profiles. The estimated external gender profile was not useful for classifying examinees by gender, achieving accuracy approximately equal to chance assignment.
590 $a School code: 0072.
650 4 $a Quantitative psychology. $3 2144748
650 4 $a Educational tests & measurements. $3 3168483
650 4 $a Behavioral psychology. $3 2122788
690 $a 0632
690 $a 0288
690 $a 0384
710 2 $a Fordham University. $b Psychology. $3 3183238
773 0 $t Dissertation Abstracts International $g 77-07B(E).
790 $a 0072
791 $a Ph.D.
792 $a 2015
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10013383