Dear 507ers,
I've graded (finally) HW6 and returned them to you. You all did
pretty well on it. As soon as I get permission from the student, I'll
send you an exemplary solution.
But the HW8's were not done well at all, especially part 8(b). It
appears that I did not make something clear. To use multiple regression
you need to code the independent variables: Gender has 2 level; so
there needs to be 1 coded vector. SES has 3 levels so you need 2 coded
vectors. And then there will be 2 coded vectors for the interaction. A
total of 5 vectors. So far so good. Most of you did that. But then to
test for the Gender Effect you need to get the R-squared (or SS) from a
FULL MODEL with all 5 vectors and a RESTRICTED MODEL leaving out the
Gender Variable (so only 4 vectors remain). And THEN you use THE F
formula (R-squared full - R-sqaured restricted etc) to calculate the F.
You cannot read the F from the printout because it is NOT over and
above. Also, for the SES effect you need to use the same FM with all 5
vectors and leave out the 2 SES vectors and compute the F using THE
formula.
Most of you noticed that your answers in (b) were different than in
(a), but didn't figure out why. We have always emphasized the FM and
RM and how things not in the RM go into the error term. But you didn't
show me you knew this concept.
In part (c) you used the MR approach correctly (with FM and RM).
The answers in (c) should be slightly different than your answers in (a)
because the dataset is slightly different.
So here's my question: What should we do about the concepts that
you should have learned for HW8? Should we postpone the final until
next week - both the WebCT part and the in-class part?
I really think we should talk about this tomorrow in class and have
a thorough review for both parts of the exam (the WebCT and the
in-class).
So we need to vote. Send me an e-mail note by 6 pm today. In the
SUBJECT indicate when you want the Final (AS SCHEDULED - both parts on
Thursday 12/6 or NEXT WEEK - we can talk about this, but it will
probably have to be on Thursday 12/13).
I'd like to review tomorrow and have the exam next week. I feel you
would all do much better. But vote today. Put either "AS SCHEDULE" or
"NEXT WEEK" in the SUBJECT line. If you want, you could put your
rationale or explanation in the body of the note. The results will be
posted very soon after 6 pm so that you know whether there will be a
FINAL tomorrow or next week. If you get this note, contact your friends
in the class via phone (in case they haven't checked their e-mail) and
have them vote. One vote for each person will be counted.
I'm sick about this. I want you all to do so well, but I didn't
make something clear. I'm leaving the country Saturday morning (12/15)
and I wanted to have this all done this week. But your education is
more important than my schedule. I'll grade the projects while I wait
for your votes to come in today.
I'm sorry.
Dennis
P.S. I've rescored Quiz 9. Most of you will be pleased with the new
scores.
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
