Tímarit um menntarannsóknir - 01.01.2013, Side 26

Anna Helga Jónsdóttir, Freyja Hreinsdóttir, Guðrún Geirsdóttir, Rögnvaldur G. Möller og Gunnar Stefánsson
Abstract
Status exam in mathematics at the University of Iceland
- results and predictive value
The dropout and failure rates in the fírst
courses in calculus in the School of Eng-
ineering and Natural Sciences (SENS) at
the University of Iceland are of great con-
cern. In recent years, only about one-third
of students who started a calculus course
completed it. For example, of 721 students
who started calculus courses in fall 2011,
37% completed a course, 17% did not pass
a fínal exam and 46% stopped before the
semester ended.
To investigate this, a status exam for
freshmen in SENS was conducted in the
second week of fall semester 2011. The goal
was to test whether and to what extent the
students met SENS standards in mathema-
tics. The students were also asked about
their background, such as gender, which
secondary school they attended, year of
graduation from secondary school, year of
the last math course they took and num-
ber of semesters studying math. The status
exam was shown to mathematics inst-
ructors in secondary schools who agreed
that it was representative of what their
students learn.
Four calculus courses are taught in
SENS: a theoretical course for math and
physics students, a slightly more applied
version for engineering students and two
practical courses for students of other su-
bjects in SENS. 80% of the students in the
course for math and physics students were
male and around 70% of the students in
the course for engineering students were
also male. The gender proportions in the
other two courses were more equal.
Performance on the status exam was
poor. Not even half of the students gave
the right answer to 50% of the problems
on the exam (median 44%). The problems
on the exam covered numbers and func-
tions, basic algebra, equation of a straight
line, trigonometry, differentiation and in-
tegration, vectors and complex numbers.
Performance was poorest on the problems
testing knowledge in basic algebra, tri-
gonometry, differentiation and integration.
There was a great difference in perform-
ance of students from school to school.
Elowever, it should be noted that the aut-
hors are not trying to evaluate the quality
of teaching in schools: that would only be
possible if data were available about the
ability of the students before they enrolled
in secondary school. The data also show
average performance decreasing with the
time since students last took a math course
in secondary school and increasing with
the number of semesters in math courses.
Analysis of variance was used to exam-
ine which factors were associated with per-
formance on the status exam in a signifíc-
ant manner. Parameters were estimated in
a model containing the factors secondary
school, year of graduation from secondary
school, year of last math course, number of
semesters in mathematics, SENS calculus
course, study program in the university
and gender. Factors that were not signific-
ant were removed one by one, ending up
with a model including secondary school,
24