INTERACTION WITH TEAM

* Was the frequency of meetings with the Stat Practice team about
  right? No.

  Both sides would probably have benefited from slightly more frequent
  interaction.  There were several meetings prior to the delivery of
  the alumni data in order to clarify goals for the project and to
  determine the necessary data to analyze.  However, after the Office
  of Alumni Relations supplied the data, there was minimal contact by
  the students until the project was nearly at its end.

* Did the team make a genuine effort to understand your problem from
  your point of view, and translate it into statistical tasks?  Yes.

  Especially given the time constraints put on them (by a delay in the
  delivery of the data by the Office of Alumni Relations), the
  students appear to have worked hard to understand the needs of our
  office.  They clearly attempted to gain insights into the business
  operations and programs of not only of Alumni Relations but the
  entire University Advancement division.  They also attempted to
  provide useful data analysis and recommendations.  However, the
  results of their analysis are relatively limited in terms of their
  ability to help us make programmatic decisions.

* Was the team able to translate the results of their statistical work
  into terms that are/were helpful to you?  No.

  Unfortunately, much of their analysis is focused on determining
  frequency distributions for various alumni data rather than on
  detailed analysis to determine predictors for alumni activity.
  Although their logistic regression results are aimed at doing this
  and validate some of our existing assumptions (on the correlation
  between activity and variables such as class year, degree type,
  number of degrees, etc.), we had hoped that the students could
  provide some more detailed insight. However, we realize that some of
  this might be due to apparent flaws in the data.

  Also, their suggestions for the design of the survey panel, although
  helpful, are very general in nature.  We had hoped for more specific
  recommendations on what methodology to use in developing and
  maintaining such a panel.  In particular, we had hoped that the
  students might use the existing data to simulate different sampling
  methods and compare how well those samples represent the population.

* Did the team keep you informed of its progress?  Yes.

  As mentioned above, more frequent progress updates would have been
  helpful in heading off problems in analyzing the data and in
  ensuring that all of the objectives of the project were met.  It
  would have allowed us to provide input on on-going basis and to
  inform to of which pieces of their data analysis would be most
  helpful.


* Did the students accept your suggestions gracefully?  Yes.

  Yi and Zhanwu were a pleasure to work with and they were very open
  to suggestions, especially at the front end of the project.
  However, as mentioned above, the lack of contact during the middle
  of the project period made it difficult to give input on their work.

DOCUMENT

* Do you have an acceptable final paper from your team?  Yes.

  As mentioned above, we had hoped for some more detailed analysis on
  predictors for alumni activity and suggestions for the design of a
  survey panel.  However, given the research and analysis that the
  students completed, their final paper appears to include all of the
  pertinent information.

* Does the document achieve your goals?  No.  

  See comments in section above.

* Are you satisfied with the document?  Yes.

FINAL PRESENTATION

* Was the team able to highlight the most important aspects of their
  work with you? Yes.

  Overall, the students did well in presenting their findings,
  especially given the fact that both are non-native speakers.
  However, it would have been helpful if the students had spent more
  time discussing their analysis of alumni activity data and their
  research on survey panel design.  They focused too much of their
  time on the exploratory analysis of the data which is less useful to
  us (considering that many of the frequency distributions they
  provides are ones that we are already well aware of).
 
OVERALL 

* Would you comment on strengths or weaknesses you perceived in any
  member of the team you worked with?

  Both Yi and Zhanwu are very hard-working and insightful individuals.
  They made a conscious effort to learn about our office and the
  concept of alumni relations, and we very much enjoyed working them.

  The only true weakness we perceived is related to their ability to
  communicate (in terms of clarifying objectives, requesting data to
  analyze, pointing out challenges in analyzing data, and presenting
  the results of their analysis).

* Would you comment on any aspect of the design of the Statistical
  Practice course that you would like to see changed?

  Although we understand that a significant aspect of the class is
  learning to work independently with clients to determine their
  problems/needs (as well as the appropriate analytical methods to
  solve those problems), it might be advisable to have the instructor
  be slightly more involved in the discussions that the students are
  having with the various clients.  This could help ensure that the
  students are using the appropriate statistical techniques and that
  they are communicating with their clients about data needs,
  difficulties in their analysis, preliminary findings, etc.

=========================================