A Study on the Reflection of Reading Component of CET4
on Chinese College Students’ Reading Ability
items testing inference of word meanings from context, items having to do with
everyday reading skills, and essay items. He further argues that “both a full-information
factor analysis and a test of unidimensionality developed by Rosenbaum (1985)
supported the same conclusion: Despite their dissimilarities, the reading items used
could be regarded as measures of a single dimension” (taken from Cornoldi & Oakhill
1996:4).
220 German second graders who were learning to read their first language were
chosen by Rost and experimented on to find out if a differential assessment of reading
comprehension is possible. He gave “38 different reading tests, a spelling test, and a
speed-of-information processing test to them, and intercorrelated their results” (Rost
1989; taken from Cornoldi & Oakhill 1996:4). Rost’s study yielded 780 correlation
coefficients. 77% of the total variance was extracted and interpreted. The high
inter-correlations between various subtests which supposedly tested various subskills
indicated failure in operationalising different sub-skills in reading comprehension. His
experiments prove the reading items used could be regarded as measures of a single
dimension (Weir, Yang & Jin 2000).
2.1.2 Bi-divisible view
The bi-divisible view of reading denotes that the ability of reading consists of at
least two theoretically isolable and distinguishable factors: vocabulary and general
reading competence. It indicates the quick and conscious reading process of using
strategies for catching the meaning efficiently and quickly. A bi-divisible view of
reading has been evidenced in studies that extracted two factors. Examples include
Berkoff (1979), Carver (1992), Farr (1968), Guthrie & Kirsch (1987) and Weir & Porter
(1994). In these studies, what turned out to be the second factor apart from the general
reading competence is vocabulary. Weir and Porter (1994: 5) note that “the phenomenon
of vocabulary loading on a separate factor is not uncommon”. Quantitative evidence
(e.g., Berg 1973, Davis 1944, Rosenshine 1980, Spearritt 1972 etc.) suggests that in
general it may not be consistently possible to identify multiple, separate reading skill
components, “there does seem to be a strong case for considering vocabulary as a
component separate from reading comprehension in general” (Weir and Porter 1994: 5).
On this basis, Weir, Yang and Jin (2000: 22) argue that “if some putative skills and
strategies function in a statistically similar manner, and so load fairly heavily on one