Málfríður - 15.03.2010, Blaðsíða 16

proficiency may contribute towards the high drop-
out rate at the University of Iceland; (2) Phillipson’s
claim that Danes are not as proficient in English as
the status of the language in Denmark would lead
one to believe; (3) Hellekjær’s findings that many
Norwegian university students have difficulty read-
ing academic English; it is clear that we need to
assess how well Icelandic university students are
able to read academic English, and not be lulled by
Icelanders’ apparent comfort and skill in talking
casually.
To test the reading proficiency of Icelandic uni-
versity students, reading passages were used from
students’ own courses. I did this first because, as
Peretz and Shoham found (1990), “EFL students pre-
fer texts on topics that are, or appear to be, related to
their field of study”, which they rate to be “easier”.
(Although Peretz and Shoham go on to point out
that students’ “hunches” about what is easy “are
not always a reliable index of their performance”.)
In other words, I wanted students to be motivated
to read texts which they would consider relevant.
Second, choosing reading material from students’
courses meant that the tests would allow predictions
to be made of students’ success in dealing with their
actual English reading requirements, although of
course using different tests meant that comparisons
could not be made between faculties.
The effectiveness of using reading tests contain-
ing students’ background knowledge is disputed.
Carrell (1983) found that “nonnative readers show
virtually no significant effects of background knowl-
edge,” as did Clapham (1990). On the other hand,
(Just et al, 1982) found that if “readers are already
familiar with the content, then the comprehension
test may be probing previous knowledge rather than
comprehension.” In any case, I ensured that the read-
ing test passages were texts that the students had not
yet read, so although they were in the students’ cho-
sen field, they were likely to contain some unknown
information or an unconsidered perspective.
I began by carrying out a small-scale pilot study
with third-year undergraduate students in an
American culture course. The reading section of
the test included an off-the-shelf TOEFL-like test
on slavery (a topic that the students had already
encountered), and a reading comprehension test that
I had developed, based on an academic text (and
topic) that the class had not yet considered, the rela-
tive lack of social mobility in post-Reagan America
(Weir, 2007). I also administered a listening test, on
which I will not report here, except to mention that
the short-answer test items of this test proved to be
more reliable than its multiple choice test items.
The TOEFL-like reading test scores were shown
to be unreliable, with many students scoring per-
fectly, so these scores were disregarded. (Of course
as a summative course test, this would be the result
one would hope for, but as a research instrument,
students’ wholesale “acing” of this test on a familiar
topic was not useful.) On the other hand, the results
of the reading comprehension test on the rise of
“Class in America” since the 1980s, a new concept
for the students, were more interesting. The average
score was 69%, with about half the students scor-
ing above 60% and about half scoring below 60%.
Considering the fact that these were third year stu-
dents taking the pilot test, I was relieved that almost
everyone passed. It was clear from the pilot that
it was in fact possible to administer a reading test
based on a course-related text, as long as the topic
had not been widely covered, and it also became
clear that test reliability using short-answer ques-
tions was indeed possible.
Incoming students in Science and Engineering
engage in a combined orientation session spanning
a couple of days in September. During this time, one
of the many activities they carry out is reading and
discussing an English text which highlights some
of the differences between science and engineering,
such as the different place that theory holds within
each discipline, and the two fields’ different con-
cepts of what “knowledge” is. This year’s text was
to be “The Wisdom of Engineers” (McCarthy, 2009),
administered to the students who attended on day
two of the orientation, numbering over 300.
I cut two theoretical sections from the original text,
largely to edit its size from 2,500 words to just over
2,300. Students were allotted 50 minutes to read the
text and answer the 10 questions with short English
or Icelandic answers. Answers were worth 1 or 2,
depending on the complexity of the questions, and
there were no half points given.
As students began finishing the test (or appear-
ing to finish the test) after only 25 minutes, the ori-
entation administrator suddenly announced to the
students that they would have five more minutes
to complete the test. Unfortunately, having only 30
minutes to answer 10 questions on a rather com-
plex academic text meant that too few students (just
under 20%) completed the whole test. Therefore, in
discussing these students’ reading proficiency, only
answers to the first six questions will be analysed
here, since most participants did complete these first
six answers.
1 MÁLFRÍÐUR