Team D		 95/100

GURARY, ELLEN BARA        Senior HSS STA  L     9.0  egurary 
SODERHOLM, MARGARET ADELE Junior HSS PHI  L     9.0  msoderho 
SUNG, JEN                 Junior HSS ECO  L     9.0  jens     
JACKSON, BRUCE WILLIAM    Junior HSS STA  L     9.0  bjackson 
SWIERKOCKI, CHRISTINA HEL Junior HSS STA  L     9.0  cswierko 

OVERALL:

I prefer you to do the friendship/involvement survey.  There are some
feasibillity issues with the Craig Street survey, and I am concerned
that the results (which coffeshop is preferred) will depend more on
your sampling efforts than on "true" underlying preferences.

I like the idea of doing cluster sampling (sample classrooms, and then
survey everyone in each classroom).  If this doesn't work out then a
simple random sample or stratified sample based on C-book will be
fine.


1. What Determines Friendship/Involvement at Carnegie Mellon?
   [could expand to social network survey]

   A. Interest.	 10/10

      fine
   

   B. Research Questions.	10/10

      Good ideas here.
  
      I like the idea of comparing the sizes of respondents' virtual
      vs "real" social networkds

   C. References.	5/5

      fine

   D. Sampling Frame		6/8

      you haven't really described a sampling frame here.  Either you
      should give a list of students (like C-Book) or a list of
      classes/classroom (e.g. from the current schedule of classes at
      hub or something).

      actually I do like the idea of a sample where you randomly
      sample classrooms and then survey everyone in each classroom.
      That's an example of a clustered survey. 

      There will be the issue of gaining the instructor's permission
      for interrupting the class, etc.  But it might be well worth it!

   E. Target Population		8/8

      ok

   F. Mode of Data Collection	5/5

      ok. p&p surveys I guess

   G. Important Variables	4/4

      good.  Add sizes of variosu networks that the person is involved
      in, as well.


2. Kiva Han vs starbucks (ebb and flow of clientelle at businesses on
   craig street)

   A. Interest.		7/10

      This is interesting but you will need permission from both
      coffeeshops to proceed, especially because it is a "competitive"
      survey that pits one business agains the other.

      A problem you raise in E below is important to consider: if you
      only visit people at each coffeeshop, you are likely to get a
      group that prefers KH at Kiva Han, and a group that prefers
      Starbucks there.  Overall preferences then will be driven mostly
      by different sample sizes at the two shops (i.e. driven by your
      effort) rather than being driven by true consumer preferences.

   B. Research Questions.   10/10

      OK

   C. References.  5/5

      OK

   D. Sampling Frame		8/8

      since there isn't an explicit sampling frame, this is a good
      start in describing the frame.  

   E. Target Population		8/8

      OK.   see A above also.

   F. Mode of Data Collection	5/5

      OK

   G. Important Variables	4/4

      OK