Background

Statistical competence demands solid grounding in theory from which statistical methods derive and familiarity with the application of these methods. This exam requires demonstration of the ability to apply statistical methods and communicate results. It consists of two parts: (1) analyzing data and submitting written reports, and (2) answering oral questions on the data analysis.

In the References there are a number of items which present informative case studies and these may be selectively and critically read. Cox and Snell is particularly appropriate, since many of their examples derive from the annual practical exam at Imperial College. The Canadian Journal of Statistics has a section which presents worked out analyses of consulting type problems.

There is no substitute for an active involvement in statistical consulting over a relatively long period. It is expected that candidates will have participated in courses on statistical consulting and had "hands-on" experience with live consulting problems. This experience should have involved real clients, computing, and report writing.

It will be assumed that all candidates will have background corresponding to Biostat/Stat 571 and 572. In some situations it may be useful to apply methods taught in other applied statistics courses. However, these courses are not required for background for the exam.

Written Reports

The exam requires a written report on the analysis of one of three given datasets. A copy of each dataset will be made available on the Hilbert computer. The reports should be prepared for a critical client who is unacquainted with statistical theory, computer technology, and statistical jargon. Each report should be limited to at most 20 pages including appendices and figures and should include most of the following sections (you have some flexibility here; what is suggested is a compromise with the recommendations of Ehrenburg, 1982):

1. Abstract: This should consist of a brief statement of the results of your investigation (not the objectives).
2. Introduction: This should include a clear statement of the examinee's interpretation of the scientific questions addressed by the data. Background provided with the data may need to be amplified. The goal of the statistical analysis should be clear to all who read the introduction.
3. Methodology: Describe precisely the models used, any theory developed to implement these models, the methods used to analyze these data assumptions, and so on. Explain why the methods are appropriate.
4. Results: Results of analyses should be presented simply and clearly. Use graphical displays tables and discussion. Undigested computer output is not appropriate here. Reference to labeled points in computer output in the appendix is suitable.
5. Discussion: This section should describe the scientific and statistical issues raised by the results described in the preceding sections. Limitations in data, scientific background and statistical expertise should be pointed out. Suggestions for further analysis or other data are appropriate. Summarize (again) your conclusions about the issues of the scientific concern. Back up your assertions with references to your Results, graphs, tables, etc.
6. Appendices: There can be one or more appendices.

Criteria used to judge performancewill include the following factors:

• the validity of the analyses and models which underlie them, and the technical execution of the analysis, as evidenced be selection and application of software,
• the thoughtfulness and simplicity of the approaches,
• the quality of the written report, The report should be carefully organized and should be clear and accurate.Simple, telegraphic (even graceless) prose is preferred to sentences and paragraphs which are convoluted of otherwise confusing.

Oral Exams

After the report is graded, the examining committee meets each student in an oral wrap-up. This oral questioning provides an opportunity to explore various points in greater depth and to clear up questions such as those relating to choice of models, priorities and analyses not reported. Questioning is not necessarily limited to just those topics covered by the student in the written portion of the exam. The examining committee will provide the student with a written evaluation of the students analysis.

Taking the Exam

Picking up a copy of the exam constitutes "taking the exam." There will be a due date for the written reports and failure to turn in the reports will constitute failure of the exam. Errata sheets will be accepted by the committee at the time of the oral exam. Obviously, such errata should be kept to a minimum.

References

Cochran, W.G. (1983). Planning and Analysis of Observational Studies. New York: Wiley.

Cox, D.R. and Snell, E.J. (1981) Applied Statistics: Principles and Examples. London: Chapman and Hall.

Miller R.C. Jr., Efron B., Brown B.W. Jr. and Moses, L.E. (1980). Biostatistics Casebook. New York: Wiley.