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

New applications in testing: Using r...

Bovaird, James Alexander, IV.

Linked to FindBook

Google Book

Amazon

博客來

New applications in testing: Using response time to increase the construct validity of a latent trait estimate.

Record Type:	Electronic resources : Monograph/item
Title/Author:	New applications in testing: Using response time to increase the construct validity of a latent trait estimate./
Author:	Bovaird, James Alexander, IV.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2002,
Description:	261 p.
Notes:	Source: Dissertations Abstracts International, Volume: 65-03, Section: B.
Contained By:	Dissertations Abstracts International65-03B.
Subject:	Psychological tests. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3082643
ISBN:	9780496304936

New applications in testing: Using response time to increase the construct validity of a latent trait estimate.
Bovaird, James Alexander, IV.

New applications in testing: Using response time to increase the construct validity of a latent trait estimate. - Ann Arbor : ProQuest Dissertations & Theses, 2002 - 261 p.

Source: Dissertations Abstracts International, Volume: 65-03, Section: B.

Thesis (Ph.D.)--University of Kansas, 2002.

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

Cognitive psychology has extensively used response time as a measure of processing. Following the cognitive paradigm shift in psychometrics, item level response time can be unobtrusively collected in a computer-based testing situation and used to illuminate the cognitive processing of examinees. Assuming that different strategies require different amounts of processing time, response time can be used to infer the existence of examinee strategies. Current test models assume that examinees use the same strategy for item completion, but differential strategy usage necessitates multiple task models. Test administration conditions further impact the strategy used by examinees by changing the speed-accuracy relationship. In three studies, response time data from three samples administered under distinct time conditions were used to increase the construct validity of the abstract reasoning ability trait estimate. In Study 1, response time was used to detect multiple examinee strategies to investigate the construct representation aspect of construct validity by determining a need for multiple processing models. Latent profile analysis was used to detect examinee classes based on response time patterns. The mean latent class response time per item was plotted along with accuracy. Examinee classes were described as solution-seeking characterized by longer-than-average latency and high accuracy or rapid-guessing with faster-than-average latency and low accuracy. Study 2 used mixture distribution modeling to determine the proportion of examinees that exhibited rapid-guessing behavior on each item. Item position and item difficulty were used to account for individual item differences in the amount of rapid-guessing that occurred on each item. The speededness of the test was also discussed. Study 3 addressed the nomothetic span aspect of construct validity and showed that excluding responses made by rapid-guessers resulted in increased item difficulty estimates, decreased standard errors of the difficulty estimate, and decreased item discrimination estimates. Solution-seeking examinees had a stronger relationship between abstract reasoning ability and verbal and mathematical ability than did the rapid-guessing examinees, but rapid-guessers had a stronger relationship between abstract reasoning ability and performance on an air-traffic controller task. Implications of the research to construct validity research and methodological considerations were also discussed.

ISBN: 9780496304936Subjects--Topical Terms:

519068
Psychological tests.
Subjects--Index Terms:

Abstract reasoning

New applications in testing: Using response time to increase the construct validity of a latent trait estimate.
LDR:03767nmm a2200361 4500 001 2278290
005 20210628082439.5
008 220723s2002 ||||||||||||||||| ||eng d
020 $a 9780496304936
035 $a (MiAaPQ)AAI3082643
035 $a AAI3082643
040 $a MiAaPQ $c MiAaPQ
100 1 $a Bovaird, James Alexander, IV. $3 3556666
245 1 0 $a New applications in testing: Using response time to increase the construct validity of a latent trait estimate.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2002
300 $a 261 p.
500 $a Source: Dissertations Abstracts International, Volume: 65-03, Section: B.
500 $a Publisher info.: Dissertation/Thesis.
500 $a Advisor: Embretson, Susan E.
502 $a Thesis (Ph.D.)--University of Kansas, 2002.
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 Cognitive psychology has extensively used response time as a measure of processing. Following the cognitive paradigm shift in psychometrics, item level response time can be unobtrusively collected in a computer-based testing situation and used to illuminate the cognitive processing of examinees. Assuming that different strategies require different amounts of processing time, response time can be used to infer the existence of examinee strategies. Current test models assume that examinees use the same strategy for item completion, but differential strategy usage necessitates multiple task models. Test administration conditions further impact the strategy used by examinees by changing the speed-accuracy relationship. In three studies, response time data from three samples administered under distinct time conditions were used to increase the construct validity of the abstract reasoning ability trait estimate. In Study 1, response time was used to detect multiple examinee strategies to investigate the construct representation aspect of construct validity by determining a need for multiple processing models. Latent profile analysis was used to detect examinee classes based on response time patterns. The mean latent class response time per item was plotted along with accuracy. Examinee classes were described as solution-seeking characterized by longer-than-average latency and high accuracy or rapid-guessing with faster-than-average latency and low accuracy. Study 2 used mixture distribution modeling to determine the proportion of examinees that exhibited rapid-guessing behavior on each item. Item position and item difficulty were used to account for individual item differences in the amount of rapid-guessing that occurred on each item. The speededness of the test was also discussed. Study 3 addressed the nomothetic span aspect of construct validity and showed that excluding responses made by rapid-guessers resulted in increased item difficulty estimates, decreased standard errors of the difficulty estimate, and decreased item discrimination estimates. Solution-seeking examinees had a stronger relationship between abstract reasoning ability and verbal and mathematical ability than did the rapid-guessing examinees, but rapid-guessers had a stronger relationship between abstract reasoning ability and performance on an air-traffic controller task. Implications of the research to construct validity research and methodological considerations were also discussed.
590 $a School code: 0099.
650 4 $a Psychological tests. $3 519068
653 $a Abstract reasoning
653 $a Construct validity
653 $a Latent trait estimate
653 $a Response time
690 $a 0632
710 2 $a University of Kansas. $3 626626
773 0 $t Dissertations Abstracts International $g 65-03B.
790 $a 0099
791 $a Ph.D.
792 $a 2002
793 $a English
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3082643