I've used similar rating systems with students who work in groups and
teams, and I have no problem justifying giving a lower grade to a low
performing member of a high performing group, as in Debra Connelley's first
example below. If the group paper received, say, a 95 out of a possible
100 points, the low performing member might receive, say, an 80, based upon
the evaluations provided by other group members.
However, I've often been puzzled by the reverse -- how to assign a grade to
a high performing student in a low performing group? As I see it, there
are at least three different arguments that can be made, each resulting in
different grading implications for the high performing group member.
One argument is that the high performing student should receive a grade NO
HIGHER than the grade of the paper, on the ground that the grade of the
paper establishes an upper limit on grades that are possible for any of the
group members. Thus, by this argument, if a paper received, say, a 75, the
high performer would receive a 75. I've always followed this reasoning in
assigning grades to members of groups and teams, on the belief that it
would be unfair to assign a grade higher than the upper limit on the work
done.
Another argument, one that Debra suggests in her second example below, is
that the high performing student should receive a HIGHER grade than the
grade assigned to the paper, on the ground that the student's performance
was hampered by the low-performing members of the group. By this argument,
the high performing student might receive, say, an 85 on a paper that
received a 75. Personally, I find this argument troubling, even though I
can see the reasoning behind it.
But a more curious argument is that the high performing student in a low
performing group should receive a LOWER grade, not a HIGHER one, on the
ground that this was the student who performed most of the work that
produced an inferior paper. By this argument, the high performing student
might receive, say, a 70, on a paper that received a 75. I want to be
clear that I'm not advocating argument #3 reasoning; I just find this
argument to be a curious way of framing the impact of the high performer in
a low performance group.
I'll probably continue to follow the reasoning of argument #1 when
assigning grades to group members, and my guess is that Debra will probably
continue to follow the reasoning of argument #2. But I sure would welcome
others' thoughts on the merits of these arguments, particularly argument
#3. Thanks.
Best,
Larry Pate
University of Wisconsin-Madison
At 04:05 PM 3/4/99 -0400, you wrote:
>At 02:26 PM 3/4/99 -0500, you wrote:
> The only way your weighting system allows for the team paper
>>grade to be used fully is if the peer evaluation is 100. Don't you
>>think you are inducing rater error here, maybe leniency? A 90 as a peer
>>evaluation seems like a good grade, but it punishes the team paper
>>grade.
>
>I have had similar concerns in grading teams. While such a system punishes
>poor performers, what happens to the superior performer? I have developed a
>system where I calculate a mean and standard deviation for each team (I have
>24 in 2 MBA core courses) based on a standard numerical peer evaluation
>form. I weight the team grade by the standard deviation in the following
>manner: + or - 25% of the team grade for individuals who score + or - 1
>standard deviation, and + or - 10% for those who score between .5 an 1
>standard deviation above or below the mean. Everyone else receives the team
>grade.
>
>The advantage, obviously, is that social loafers are not rewarded, so a low
>performer in a high performing team will not get a free ride.(ex. team grade
>= 90, individual's peer evaluation score is 1 SD below the mean for that
>team, so his/her team grade = 68.25.) Similarly a high performer in a low
>performing team will not be screwed.(team grade = 80, individual's grade =
>80*1.25=100) For those teams who refuse to distinguish between high and low
>performers, everyone gets the same grade, which, as it turns out, is often
>just an average performance.
>
>Though this may sound like a lot of trouble, it is very easy to automate on
>an excel spreadsheet.
>
>Debra Connelley
>Assistant Professor, Organizational Behavior
>State University of New York at Buffalo
>