Page 215 - Contributed Paper Session (CPS) - Volume 7

CPS2065 Lina Schelin et al.
4. The project
As mentioned, the main aim of the project was to reach a higher level of
consensus among the teachers at our department, especially for the
examination of the Statistics A1 course. The two distinct parts of the project
will be described in more detail next.

Focus group
The aim of the focus group discussions was to use so called “social
moderation” to better understand how teachers in the assessment teams
categorize different mistakes in students’ solutions of exam tasks. Social
moderation, here, imply teachers working together when assessing and
grading (Thornberg & Jönsson, 2015). All participants were asked to mark
and grade two students’ solutions (of the whole exam) before the focus
group meeting. Not all participants corrected the same solutions. In total,
the group corrected six different exams. The participants were also asked
to write down how they were thinking during their correction process, to
better take use of the two hours allocated for the focus group discussions.
We closely followed the discussions in the group to understand when the
teachers were agreeing or not, and if they came to a consensus on the
different student solutions or not.

Examination guide
Based on the discussions in the focus group an examination guide was
developed. Particular focus was given to those things that was difficult to
reach consensus about during the discussions. A preliminary examination
guide was tested, in a pilot study, during the marking of the exams in the
course given in the second part of the spring semester 2018. A sample of
the exams was marked using the instructions in the examination guide, in
parallel to the ordinary marking in the course.

5. Result and discussion
The results of the focus group discussions were that the participants
usually agreed on the overall grade of the exams, but that the exact points for
each task could differ quite a lot. Some of our teachers are using a holistic
approach when correcting, while others use a more fragmented approach
(giving points to specific parts without an overall perspective of the solution).
Since the grades are awarded based on the total score on exam, students that
do not reach the expected learning outcomes are sometimes awarded the
grade Pass, or are close to being awarded grade Pass, when the more
fragmented approach is used. This problem might be reduced by not allowing
such a fine resolution of point increments as we do now. With only a few
predefined levels (or points) the teachers are forced to categorize the

202 | I S I W S C 2 0 1 9