Open AccessModeling Growth With Adaptive Longitudinal Large‐Scale Assessments
Author(s) -
Qian Jiahe
Publication year - 2018
Publication title -
ets research report series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.235
H-Index - 5
ISSN - 2330-8516
DOI - 10.1002/ets2.12220
Subject(s) - item response theory , latent growth modeling , scale (ratio) , longitudinal study , psychology , growth model , longitudinal data , logistic regression , set (abstract data type) , statistics , mathematics , psychometrics , computer science , geography , cartography , data mining , mathematical economics , programming language
The 2‐parameter logistic multidimensional item response theory (MIRT) model was employed to model growth for the National Education Longitudinal Study of 1988 (NELS:88). The 3 measurement waves of NELS:88 (base year, first follow‐up, and second follow‐up) represented 3 dimensions. The inquiry aimed to improve modeling performance growth based on examinees' responses to the test items in each wave, with item location parameters set to be invariant across the 3 waves (instead of using item changes) and the latent mean of the first wave set to 0. The yielded scores of 3 waves were thus placed on approximately the same scale; the changes of the scores across waves could be measured. Moreover, the growth models for longitudinal data were improved by using auxiliary information such as gender, race, school location, and parents' education. In the study, the results of the growth pattern were compared with those yielded by the Embretson models.