Naslov
In search for characteristics of test-takers as potential sources of construct-irrelevant variance

Autori
Ljubotina, Damir

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
22. Dani Ramira i Zorana Bujasa - knjiga sažetaka / Zarevski, P. ; Jurin, T. ; Modić Stanke, K. - Zagreb : Odsjek za psihologiju Filozofskog fakulteta Sveučilišta u Zagrebu, 2015, 64-64

ISBN
978-953-175-463-7

Skup
22. Dani Ramira i Zorana Bujasa

Mjesto i datum
Zagreb, Hrvatska, 16.04.2015. - 18.04.2015

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Domaća recenzija

Ključne riječi
test taking strategies, construct irrelevant variance, test theory

Sažetak
One source of threats to the validity of measurement lies in the construct irrelevant variance (CIV) that can be defined as a systematic variation in scores that is not associated with the measured construct but is in fact an error. Messick (1995) lists several broader groups of CIV: test characteristics, test administration conditions, respondent characteristics such as specific strategies and behaviours used in test situation. The main goal of this study was to develop and validate an instrument aimed to measure some characteristics, strategies and behaviours of respondents in test situations that could be regarded as potential sources of CIV, and to develop a model that would enable identification and integration of different sources of CIV. With that in mind, a questionnaire assessing six dimensions was developed. The first three dimensions refer to respondent strategies in test situation: overall speed of performance, persistence and tendency to guessing at random. Remaining dimensions refer to respondent’s general sensitivity in test situations and include: test anxiety, effect of limited time on efficacy and general sensitivity to distraction. Correlations between these new measures and following personality dimensions were assessed: conscientiousness, impulsivity, neuroticism, perfectionism and three dimensions of temperament along with performance on a knowledge based test administered under limited and unlimited time. Results were obtained on a sample of 220 high school and 450 university students. Preliminary results indicate a satisfactory reliability and factor structure of the new instruments. Significant correlations were found between temperament and the majority of the dimensions used. Conscientiousness and impulsiveness have a slightly higher correlation with test taking strategies, while perfectionism and neuroticism correlate with anxiety symptoms and general sensitivity to distraction in test situation.

Izvorni jezik
Engleski

Znanstvena područja
Psihologija

POVEZANOST RADA