The
aim of the present study has
been to investigate
difficult KPG reading comprehension
test items across two levels
of language proficiency,
namely B1 and C1 level, in
order to shed light a) on
the cognitive processes in which
test-takers engage in order
to respond to reading test
items and b) on the
cognitive factors which
contribute to the difficulty
of an item. My ultimate
purpose has been to identify
whether significant
differences there exist
across the two levels of
language proficiency with
respect to the cognitive
load variables that are
inherent in the test items.
To this effect, I
systematically analyzed a
body of thirty-six (36) B1 and
C1 level multiple-choice
items extracted from the KPG
past exam papers over a
period of two years, during
four different
administration periods. My
item corpus consisted of
items with a low easiness
index, specifically ei ≤ 50,
which means that I
considered the items that
proved to be difficult for
50% of the test takers or
more at each of the two levels
(B1 or C1) of language
proficiency. The systematic
analysis and critical
approach of the items in
focus has been conducted
with the help of two members
of the KPG test development
team working at the RCeL
with extensive experience in
pre-testing the KPG exams.
The findings of the
research suggest that both
test items assessing comprehension
of information stated in the
text and inferential test questions
can be cognitively demanding
for B1 and C1 level
examinees. However, there
are qualitative differences
across levels with reference to
the cognitive load factors
entailed in each type of
test items, ones that point
to the fact that C1 reading
comprehension test items are cognitively
more demanding than the B1
items. In addition, test-wiseness strategies
were found to hold a
prominent role in test-takers’
cognition, regardless of
their level of proficiency.
The findings of the present research
may be of interest to test
item writers as well as
English language instructors.
Suggestions for further
research also provided.
