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

Two studies related to statistical e...

The Johns Hopkins University.

Linked to FindBook

Google Book

Amazon

博客來

Two studies related to statistical education.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Two studies related to statistical education./
Author:	McGready, John B.
Description:	401 p.
Notes:	Adviser: Ron Brookmeyer.
Contained By:	Dissertation Abstracts International69-04B.
Subject:	Biology, Biostatistics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3309720
ISBN:	9780549580898

Two studies related to statistical education.
McGready, John B.

Two studies related to statistical education. - 401 p.

Adviser: Ron Brookmeyer.

Thesis (Ph.D.)--The Johns Hopkins University, 2008.

Background. This work is a collection of two studies designed to answer fundamental questions germane to statistical education in the high technology era. The first study was a head-to-head comparison of course outcomes between an online and on-campus offering of the same course. The second study compared classical and bootstrap-based approaches to confidence interval estimation for a mean difference between two populations based on small sample comparisons.

ISBN: 9780549580898Subjects--Topical Terms:

1018416
Biology, Biostatistics.

Two studies related to statistical education.
LDR:03390nmm 2200337 a 45 001 867094
005 20100802
008 100802s2008 ||||||||||||||||| ||eng d
020 $a 9780549580898
035 $a (UMI)AAI3309720
035 $a AAI3309720
040 $a UMI $c UMI
100 1 $a McGready, John B. $3 1035798
245 1 0 $a Two studies related to statistical education.
300 $a 401 p.
500 $a Adviser: Ron Brookmeyer.
500 $a Source: Dissertation Abstracts International, Volume: 69-04, Section: B, page: 2035.
502 $a Thesis (Ph.D.)--The Johns Hopkins University, 2008.
520 $a Background. This work is a collection of two studies designed to answer fundamental questions germane to statistical education in the high technology era. The first study was a head-to-head comparison of course outcomes between an online and on-campus offering of the same course. The second study compared classical and bootstrap-based approaches to confidence interval estimation for a mean difference between two populations based on small sample comparisons.
520 $a Methods. Study 1 utilized study supplied information via confidential online surveys couple with instructor compiled information on course outcomes. Study 2 was done via simulations comparing the coverage properties and relative average widths of intervals by classical-t and bootstrap (bootstrap-t, bootstrap-percentile, BCa) approaches to confidence interval estimation for mean differences. These simulations were performed for several families of data generating distributions (normal, gamma, lognormal, discrete ordinal) for multiple small sample size combinations.
520 $a Results. Study 1 found no statistically, nor qualitatively significant differences in course exam or homework outcomes, despite substantial differences in the demographics and academic histories of students enrolled in the two sections compared. There was some evidence that quiz outcomes differed by a notable average amount, and further research is planned to investigate. The primary finding of the second study was that there was no tenable advantage to using bootstrap-based approached to confidence interval estimate for a differences in means across all scenarios considered. Secondary findings included results related to the sampling bias of the sample standard deviation for difference sample generating distributions.
520 $a Conclusions. Study 1 adds evidence to a small, but growing number of studies demonstrated equivalencies for measurable outcomes between online and more traditional classroom based approaches to basic statistical education. Study 2 answers a nagging question regarding the optimal choice of methods for creating confidence intervals for a mean difference based on small sample comparisons for a variety of sample generating distributions. The results clearly demonstrate that bootstrap based approaches are not a panacea for correcting deficiencies in the classical-based approach, and nearly uniformly these bootstrap-based approaches underperformed relative to the former approaches.
590 $a School code: 0098.
650 4 $a Biology, Biostatistics. $3 1018416
650 4 $a Education, Mathematics. $3 1017588
650 4 $a Education, Technology of. $3 1018012
650 4 $a Health Sciences, Public Health. $3 1017659
690 $a 0280
690 $a 0308
690 $a 0573
690 $a 0710
710 2 $a The Johns Hopkins University. $3 1017431
773 0 $t Dissertation Abstracts International $g 69-04B.
790 $a 0098
790 1 0 $a Brookmeyer, Ron, $e advisor
791 $a Ph.D.
792 $a 2008
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3309720