Rf. Waugh et Pa. Addison, A RASCH MEASUREMENT MODEL ANALYSIS OF THE REVISED APPROACHES TO STUDYING INVENTORY, British journal of educational psychology, 68, 1998, pp. 95-112
Background. The Revised Approaches to Studying Inventory (Entwistle &
Tait, 1994) comprises 38 self-report items designed to measure student
approaches to learning in a higher education context. The items have
been conceptualised and designed from five learning orientations, corr
esponding to five subscales of the Inventory: 'a deep approach','a sur
face approach','a strategic approach','a lack of direction' and 'acade
mic self-confidence'. Aims. The study aims to create an interval level
scale for the Inventory and analyse its psychometric properties using
a modern measurement model, the Extended Logistic Model of Rasch (And
rich, 1988a, 1988b; Rasch, 1980), and investigate the conceptual desig
n of the Inventory. Sample. The sample was 346 students (170 females,
176 males, 212 less than 23 years and 134 older than 23), studying in
first year Faculty of Business, at a university in Perth, Western Aust
ralia. Method. A scale was created for the Inventory and analysed for
reliability, fit to the model, meaning and validity. The Inventory was
analysed separately for each of four subgroups (females, males, young
er and older students) to test the invariance of the scale. The five s
ubscales were analysed separately to test the conceptual design and me
aning of the Inventory. Results. The Inventory has satisfactory psycho
metric properties, when items 20 and 33 are deleted. Easier and harder
items need to be added to better target the student measures. Its con
ceptual design from the five learning orientations is confirmed. Only
15 items have satisfactory invariance across the four subgroups. The p
sychometric properties of three subscales (deep approach, surface appr
oach and strategic approach) are only moderately satisfactory and the
item separabilities of the other two (lack of direction and academic s
elf-confidence) are unsatisfactory.