Team C		98/100

SINDLER, STEPHANIE        Senior HSS ECO  L     9.0  ssindler 
BONCEK, EMILY LOUISE      Junior CMU SHS  L     9.0  eboncek 
CHANG, KELLY LAN          Junior CIT ECE  L     9.0  klchang 
MOSCA, LEANDRA FAITH      Senior HSS SDS  L     9.0  lmosca   
CHANG, CHRISTOPHER J.     Junior CFA MUS  L     9.0  cjchang 

Overall:

(a) Both surveys are well-described.  I prefer you to do the first
survey however.  It is interesting, and much more feasible than trying
to sample area hs classrooms.

(b) What is going on with Leandra Mosca?  Her name doesn't appear in
either of your proposals.

1. Survey of Carnegie Mellon Faculty Regarding Attendance and Student
   Performance (On-Campus)

   A. Interest.		10/10

      still a nice topic

   B. Research Questions.	10/10

      this is fine.  You may still have to consider  whether
      you want to ask faculty about *current* classes, or *past*
      classes.  For *past* classes faculty may be able to provide
      final grade distributions or summaries.  For *current* classes,
      faculty can only provide midterm grades (safe to assume you will
      do the survey after midterm).

   C. References.	5/5

      ok.  again, thinking about these references when you write up
      the results will be good.

   D. Sampling Frame		8/8

      good to think of both the directory and dept websites as sources
      of email addresses.  To simplify your life, you may want to
      restrict attention to one of these (e.g. CMU directory).

   E. Target Population		8/8

      good.  note that some instructors will not be listed on either
      websites or the cmu directory (e.g. last-minute hires).  This is
      a source of coverage error.  Rather than try to "fix" it, I
      would be inclined to accept it and try to show that the coverage
      error will not be very big.

   F. Mode of Data Collection	5/5

      I think professors will be less likely to participate if the
      questions are in an email or on a website that they have to go
      to.  You may find it necessary to do face-to-face interviews
      with those who you have randomly selected into your sample.

   G. Important Variables	4/4

      converting some of these into useful survey questions
      (e.g. "grade distribution") will require some thought.


2. Pittsburgh High School Students' Interest in Local Colleges

   A. Interest.	   10/10

      Reformulation of research question is fine.

   B. Research Questions.	10/10

      okk    

   C. References.     	  3/5

      seem fine.  aren't references #1 and #4 the same? 

      The city of pittsburgh and pittsburgh public schools have
      developed the "Promise" program, which is basically a college
      scholarship based on years of attendance in pgh schools.  There
      should be some data on college attendance patterns in
      promotional or reporting materials for the "promise" program, at
      the pps website, the city website, etc.  This would make good
      additional reference material.

   D. Sampling Frame		7/8

      how will you assemble this sampling frame?

   E. Target Population		7/8

      good comments on non-response, non-sampling error, measurement
      error.

      what about coverage error?  How could your sampling frame of
      "accessible students" fail to cover the whole target population
      of hs jrs and srs?

   F. Mode of Data Collection	5/5

      paper surveys in classrooms seems fine to me.

      actually it sounds to me like you might actually make a random
      sample of *classrooms* and then follow your plan of surveying
      everyone in each sampled classroom.  This would be an example of
      *clustered sampling*, which we will talk about later in the
      semester.


   G. Important Variables	4/4

      ok