Hi Linsey
For the past six years I have been using peer assessment (the last three
years using a branched version of WebPA) to generate 75% of marks for an
industry-based capstone software engineering design course where 3rd and
4th year students are required to work for two semesters in small teams
of 5-7 students developing a software solution to a real-world problem
for their client. At four points during the course team progress is
assessed using gateway reviews which involve the team, the client, the
team mentor and course academics. Peer assessment is used to calculate a
moderator which is used to generate individual marks from the team mark.
Students like using peer assessment in this way as it empowers them. It
enables students who want to work hard and get a good mark while not
rewarding those students who have less drive and seems to overcome many
(if not all) of the problems associated with social loafing and free-riding.
The assessment scheme makes it clear, and we are at pains to explain to
students how we use peer assessment. We are also explicit that we
reserve the right to amend the marks generated by peer assessment if we
feel they are not a true reflection of contribution and that the marks
they receive therefore may not be exactly as generated by WebPA. We have
rarely had to intervene and amend the peer assessed marks. We do,
however, exclude self-assessment when calculating the peer assessed
rating as otherwise students can affect their own results in a way that
is not necessarily reflective of contribution. We collect written
feedback as well as the numeric ratings (using a Likert scale of 1-5).
The written feedback provides valuable insight into team contribution
and validates the numeric ratings. The peer assessment data we collect
usually validates the performance we see in the gateway review process
and what mentors observe while supporting teams. Like Neil we use SVN
repositories and collaborative sites so the academics involved have
access to all artefacts and team communication which enables us to
verify claims if necessary.
Since the introduction of peer assessment, the complaints about marks
have dropped to almost zero. After all who better to know what is going
on in a team than the members of the team itself. Now with minimal
effort by the academics we can award defensible individual marks for
group work. The peer assessment data collected can also be used to
mentor students to improve their performance - we have some great
evidence of this working extremely well.
The course forms part of a 4-year accredited software engineering degree
and our accrediting body have been very accepting of the benefits the
students draw through the use of peer assessment to generate individual
marks from group artefacts.
Cheers
Lynette
On 25/04/2013 01:13, Linsey Duncan-Pitt wrote:
> Thank you Neil, Keith and Paul
>
> This is really helpful information. I saw a presentation on this a little while ago but gleaned from that and the WebPA resources that there were checks and balances. However the real experience of people using it at the degree classification module level is what I was really interested in. I like the safety net options you mentioned Neil.
>
> Best Regards
> Linsey
--
====================================================
Lynette Johns-Boast
Lecturer
Research School of Computer Science
The ANU College of Engineering and Computer Science
CS&IT Building 108
The Australian National University
Canberra ACT 0200 Australia
T: +61 2 6125 4526
F: +61 2 6125 0010
M: +61 405 611 859
W: http://cecs.anu.edu.au
Available Tuesday, Wednesday & Thursday
|