Robert & others,
It can be useful when trying to analyze patterns like this, to identify
the length of time and number of tests involved in the data
gathered.
Because of my long association with Larry, I am aware that the
"cumulative scores" he cites represents data from iRATs and
gRATs added up at the end of a whole semester, i.e., 15 weeks and usually
6 to 7 RATs.
Robert: Have many RATs and weeks were represented by your
data?
Was that
just from one RAT or the combined data from several RATs?
Dee
At 09:44 AM 11/24/2005, Larry Michaelsen wrote:
My explanation is a combination
of those that you offer. The data you
present is typical of: 1) a single (and probably short) test, 2)
given
to newly-formed groups--not yet teams, 3) not using IFAT answer
sheets.
Thus, your results are a likely to be a combination of the
explanations
you offer plus the "lucky guess," option could just as easily
be called,
"unreliable test"--which is characteristic of almost ANY short
and new
test. (You might want to check out Watson, W. E., Michaelsen, L. K.
&
Sharp, W. (1991). Member competence, group interaction and
group
decision-making: A longitudinal study. Journal of Applied
Psychology,
76, 801-809 to see a fuller explanation and some supporting
empirical
evidence.) Based on cumulative scores (i.e., increased reliability
from
a longer test), between 1986 and 2003 (when I started using IFATs),
I
had data from over 1,100 teams and all but 1 team scored higher than
its
own best member and by an average of nearly 11%. Since I've started
using IFATs (which provide teams with some within-test feedback, I
haven't had any team fail to beat its best member and the average
gain
has been 20+%.
Larry
Larry K. Michaelsen
Professor of Management
Central Missouri State University
Dockery 400G
Warrensburg, MO 64093
660/543-4124 voice
660/543-8465 fax
>>> "Philpot, Robert J."
<[log in to unmask]> 11/23/05 3:11 PM >>>
Hello All,
I have been intrigued by the comparisons of team scores on the gRATs
to
the High, Low and Mean scores on the iRATs. Lately I've had the
opportunity to keep track of scores following different team
learning
experiences. It struck me as a little odd that some teams actually
score
lower than the highest member on that team. I've attributed this to
1.)
inexperience working as a team, 2.) withholding by the brighter team
member (for whatever reason), and 3.) lucky guessing by unprepared
students that cannot help their team experience the same success.
Perhaps someone could posit another cause?
Low
High
Mean
Team Score
Gain above high
Gain above mean
Gain above low
Team1
56%
67%
61%
72%
6%
11%
17%
Team 2
50%
89%
78%
94%
6%
17%
44%
Team 3
67%
89%
78%
94%
6%
17%
28%
Team 4
61%
83%
78%
89%
6%
11%
28%
Team 5
50%
89%
67%
78%
-11%
11%
28%
Team 6
56%
89%
72%
83%
-6%
11%
28%
Team 7
72%
89%
78%
94%
6%
17%
22%
Team 8
67%
78%
75%
89%
11%
14%
22%
Team 9
72%
89%
78%
89%
0%
11%
17%
Team 10
56%
78%
67%
83%
6%
17%
28%
My real question, however, revolves around the analysis of this data
once it is collected. Has anyone used a reasonable statistical test
to
compare individual scores on the iRAT to the team scores on the
gRAT? I
have been considering ways to compare performance of several teams
on
gRATs (dependent variable) following the use of an educational
intervention (independent variable). All of the students will have
taken
the iRAT prior to the intervention so I could compare team scores to
high, low and mean individual scores for each group also. Exp.
(iRAT->X1->gRAT) vs. control (iRAT->gRAT->X2).
'Looking forward to hearing your ideas and experiences,
Bob
Robert Philpot Jr., PhD, PA-C
Clinical Assistant Professor
Associate Clinical Coordinator
University of Florida College of Medicine
Physician Assistant Program
Gainesville, FL 32610-0176
352-265-7955 w
352-871-5053 mobile
[log in to unmask]
http://medinfo.ufl.edu/pa/faculty/Bob/
"Someday, after mastering the winds, the waves, the tides and
gravity,
we shall harness for God the energies of love, and then, for a
second
time in the history of the world, man will have discovered
fire."
Pierre Teilhard de Chardin
!DSPAM:1576,4384e50a18453946398363!
* * * * * * * * * * * * * * * * * * * * * * *
*
L. Dee Fink, Instructional
Consultant
Phone:
405-364-6464
in Higher
Education
Email:
[log in to unmask]
234 Foreman
Ave
FAX:
405-364-6464
Norman, OK
73069
Website:
www.finkconsulting.info
**Author of: Creating Significant Learning Experiences
(Jossey-Bass, 2003)
**Immediate Past President of the POD Network [Professional and
Organizational Development] in Higher Education
**Founding director (now retired), Instructional Development Program,
University of Oklahoma (1979-2005)