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

Accuracy of effect size calculation ...

Glaser, Renita R.

FindBook

Google Book

Amazon

博客來

Accuracy of effect size calculation methods for repeated measures and ANCOVA data.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Accuracy of effect size calculation methods for repeated measures and ANCOVA data./
作者:	Glaser, Renita R.
面頁冊數:	155 p.
附註:	Source: Dissertation Abstracts International, Volume: 63-02, Section: B, page: 1070.
Contained By:	Dissertation Abstracts International63-02B.
標題:	Psychology, Experimental. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3042525
ISBN:	0493564144

Accuracy of effect size calculation methods for repeated measures and ANCOVA data.
Glaser, Renita R.

Accuracy of effect size calculation methods for repeated measures and ANCOVA data. - 155 p.

Source: Dissertation Abstracts International, Volume: 63-02, Section: B, page: 1070.

Thesis (Ph.D.)--The University of Memphis, 2002.

The primary goal of the project was to explore differences in effect size estimates between Cohen's <italic>d</italic> and estimates using repeated measure and ANCOVA methods. Monte Carlo simulations using the GAUSS program (Aptech system, Inc., 1992) were used to compare Cohens <italic>d</italic> to the 14 effect size estimation methods of interest and to explore the variance of these methods. The computer simulations allowed investigation of whether the means and variances of the 14 effect size estimates matched the expected means and variances programmed into the simulations. The findings of this study raise caution to meta-analysts when using effect size calculation methods not known to be equivalent to Cohen's <italic>d</italic>. Across the 14 effect size calculation methods used to approximate Cohen's <italic>d</italic> from either repeated measures or ANCOVA data, 25.4% of the average effect sizes were significant indicating that, as a whole, the tested methods produced different effect size estimates in comparison to the true, population parameter estimate, δ. All but three methods were heavily influenced by sample size and the intraclass correlation (<italic>r</italic>i) or partial correlation coefficient (<italic>r</italic>w). The magnitude of the bias when these methods are used in actual meta-analyses would depend on the number of effect sizes calculated using methods not known to be equivalent to Cohen's <italic>d</italic> as well as the type of effect size calculation method used. Implications for meta-analysts currently using these methods are discussed and suggestions for future research are made.

ISBN: 0493564144Subjects--Topical Terms:

517106
Psychology, Experimental.

Accuracy of effect size calculation methods for repeated measures and ANCOVA data.
LDR:02555nmm 2200277 4500 001 1854998
005 20040609162024.5
008 130614s2002 eng d
020 $a 0493564144
035 $a (UnM)AAI3042525
035 $a AAI3042525
040 $a UnM $c UnM
100 1 $a Glaser, Renita R. $3 1942821
245 1 0 $a Accuracy of effect size calculation methods for repeated measures and ANCOVA data.
300 $a 155 p.
500 $a Source: Dissertation Abstracts International, Volume: 63-02, Section: B, page: 1070.
500 $a Major Professor: William R. Shadish.
502 $a Thesis (Ph.D.)--The University of Memphis, 2002.
520 $a The primary goal of the project was to explore differences in effect size estimates between Cohen's <italic>d</italic> and estimates using repeated measure and ANCOVA methods. Monte Carlo simulations using the GAUSS program (Aptech system, Inc., 1992) were used to compare Cohens <italic>d</italic> to the 14 effect size estimation methods of interest and to explore the variance of these methods. The computer simulations allowed investigation of whether the means and variances of the 14 effect size estimates matched the expected means and variances programmed into the simulations. The findings of this study raise caution to meta-analysts when using effect size calculation methods not known to be equivalent to Cohen's <italic>d</italic>. Across the 14 effect size calculation methods used to approximate Cohen's <italic>d</italic> from either repeated measures or ANCOVA data, 25.4% of the average effect sizes were significant indicating that, as a whole, the tested methods produced different effect size estimates in comparison to the true, population parameter estimate, δ. All but three methods were heavily influenced by sample size and the intraclass correlation (<italic>r</italic>i) or partial correlation coefficient (<italic>r</italic>w). The magnitude of the bias when these methods are used in actual meta-analyses would depend on the number of effect sizes calculated using methods not known to be equivalent to Cohen's <italic>d</italic> as well as the type of effect size calculation method used. Implications for meta-analysts currently using these methods are discussed and suggestions for future research are made.
590 $a School code: 1194.
650 4 $a Psychology, Experimental. $3 517106
650 4 $a Statistics. $3 517247
690 $a 0623
690 $a 0463
710 2 0 $a The University of Memphis. $3 1025952
773 0 $t Dissertation Abstracts International $g 63-02B.
790 1 0 $a Shadish, William R., $e advisor
790 $a 1194
791 $a Ph.D.
792 $a 2002
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3042525