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

Polytomous factor analytic models in...

Bontempo, Daniel E.

FindBook

Google Book

Amazon

博客來

Polytomous factor analytic models in developmental research.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Polytomous factor analytic models in developmental research./
作者:	Bontempo, Daniel E.
面頁冊數:	186 p.
附註:	Source: Dissertation Abstracts International, Volume: 67-08, Section: B, page: 4754.
Contained By:	Dissertation Abstracts International67-08B.
標題:	Psychology, Developmental. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3231806
ISBN:	9780542842344

Polytomous factor analytic models in developmental research.
Bontempo, Daniel E.

Polytomous factor analytic models in developmental research. - 186 p.

Source: Dissertation Abstracts International, Volume: 67-08, Section: B, page: 4754.

Thesis (Ph.D.)--The Pennsylvania State University, 2006.

The relations among constructs in our analytical models of human development correspond to substantive theories, however it is often forgotten that there is a theory of measurement that connects our data to the constructs in our models. This work examines the use of factor analytic measurement models for test instruments with polytomous item response formats. Specific attention is given to issues of factorial invariance, the theorized causal direction between indicators and constructs, and whether the intended use of the test if for diagnostic screening or for individual differences research. When clinical checklists, which often have dichotomous (yes/no) or polytomous (never, sometimes, always) response formats are used in individual differences research, measurement issues need to be reviewed because decisions made during the development of a diagnostic instrument often differ from decisions made in the course of developing an instrument intended to quantify inter-individual and intra-individual differences. In addition, factorial invariance procedures for constructs with categorical (i.e., dichotomous or polytomous) indicators, and the associated confirmatory factor analysis (CFA) model are examined. These goals are advanced through the application of polytomous CFA modeling of two checklist instruments: The Developmental Behavior Checklist, and the short form of the Eysenck Personality Questionnaire. First, a general review is undertaken of the factorial invariance literature, causal directions in factor analytic models, and recent extensions that permit factor analytic models of polytomous test items. Next the factor structure of the Developmental Behavior Checklist is examined using polytomous CFA. Issues of causal direction, factorially complex items, and simple structure are discussed. Finally, polytomous CFA models are used to investigate the factorial invariance of the Eysenck Personality Questionnaire across gender and late-adult age groups. In polytomous factor models a latent continuous response variable underlying each observed categorical indicator is connected to the observed indicators by calculating the cutpoint (or threshold) needed to produce the observed response frequencies. Since these thresholds are all that connects the actual data to the factor model, it is essential that investigations of invariance test the threshold parameters in addition to the usual test of loadings and item-specific variances. However, the additional measurement model parameters in polytomous CFA presents issues for the specification and identification of an appropriate baseline model for nested invariance tests. This problem is addressed by using two interdependent nested sequences. One sequence permits testing of measurement thresholds and loadings, while assuming invariance of item-specific variance, and the other sequence permits testing of the item-specific variances by assuming threshold invariance.

ISBN: 9780542842344Subjects--Topical Terms:

1017557
Psychology, Developmental.

Polytomous factor analytic models in developmental research.
LDR:03832nmm 2200277 4500 001 1832724
005 20070807074156.5
008 130610s2006 eng d
020 $a 9780542842344
035 $a (UMI)AAI3231806
035 $a AAI3231806
040 $a UMI $c UMI
100 1 $a Bontempo, Daniel E. $3 1921443
245 1 0 $a Polytomous factor analytic models in developmental research.
300 $a 186 p.
500 $a Source: Dissertation Abstracts International, Volume: 67-08, Section: B, page: 4754.
500 $a Adviser: Scott Hofer.
502 $a Thesis (Ph.D.)--The Pennsylvania State University, 2006.
520 $a The relations among constructs in our analytical models of human development correspond to substantive theories, however it is often forgotten that there is a theory of measurement that connects our data to the constructs in our models. This work examines the use of factor analytic measurement models for test instruments with polytomous item response formats. Specific attention is given to issues of factorial invariance, the theorized causal direction between indicators and constructs, and whether the intended use of the test if for diagnostic screening or for individual differences research. When clinical checklists, which often have dichotomous (yes/no) or polytomous (never, sometimes, always) response formats are used in individual differences research, measurement issues need to be reviewed because decisions made during the development of a diagnostic instrument often differ from decisions made in the course of developing an instrument intended to quantify inter-individual and intra-individual differences. In addition, factorial invariance procedures for constructs with categorical (i.e., dichotomous or polytomous) indicators, and the associated confirmatory factor analysis (CFA) model are examined. These goals are advanced through the application of polytomous CFA modeling of two checklist instruments: The Developmental Behavior Checklist, and the short form of the Eysenck Personality Questionnaire. First, a general review is undertaken of the factorial invariance literature, causal directions in factor analytic models, and recent extensions that permit factor analytic models of polytomous test items. Next the factor structure of the Developmental Behavior Checklist is examined using polytomous CFA. Issues of causal direction, factorially complex items, and simple structure are discussed. Finally, polytomous CFA models are used to investigate the factorial invariance of the Eysenck Personality Questionnaire across gender and late-adult age groups. In polytomous factor models a latent continuous response variable underlying each observed categorical indicator is connected to the observed indicators by calculating the cutpoint (or threshold) needed to produce the observed response frequencies. Since these thresholds are all that connects the actual data to the factor model, it is essential that investigations of invariance test the threshold parameters in addition to the usual test of loadings and item-specific variances. However, the additional measurement model parameters in polytomous CFA presents issues for the specification and identification of an appropriate baseline model for nested invariance tests. This problem is addressed by using two interdependent nested sequences. One sequence permits testing of measurement thresholds and loadings, while assuming invariance of item-specific variance, and the other sequence permits testing of the item-specific variances by assuming threshold invariance.
590 $a School code: 0176.
650 4 $a Psychology, Developmental. $3 1017557
650 4 $a Psychology, Psychometrics. $3 1017742
690 $a 0620
690 $a 0632
710 2 0 $a The Pennsylvania State University. $3 699896
773 0 $t Dissertation Abstracts International $g 67-08B.
790 1 0 $a Hofer, Scott, $e advisor
790 $a 0176
791 $a Ph.D.
792 $a 2006
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3231806