STAT 538A Final Project
Due: noon, Wednesday, December 19
This final project requires each student to find a data set and to independently conduct a comprehensive data analysis and write a statistical report. Having access to an unpublished dataset, such as a dataset from your projects in other courses, would be excellent, but use of published data from other sources is also permissible. There are a number of on-line repositories for data sets, many of which are described at rdir-com: Free Datasets. Once you find a dataset, you need to analyze the dataset using models and methods learned in this course and report your results and conclusions. Your written report should be less than 10 pages including figures and tables (feel free to use R markdown with html output, but check your document length by exporting to a paginated pdf file)
Your report should contain the following materials:
- objectives of the study
- data description
- exploratory analysis: this section may include summary statistics, graphical displays of the data, and any preliminary conclusions. No models are assumed in this section.
- formal (confirmatory) analysis: this section should include regression models (GLM's or other models considered this term), model selection, model diagnostics, justifications of models and methods used, and interpretation of results.
- it would be desirable to consider two or more different approaches/models to see if the results agree and explain why or why not. Report any insights from different models and methods. Discuss advantages and disadvantages of each model/method.
- final conclusions, as well as discussion about validity of the conclusions.
- Do not use the same dataset as someone else in class.
- Tables and figures should be selective (i.e., do not include too many tables or figures).
- Computer outputs should be summarized (i.e., do not include raw computer code/output in your report) and should be interpreted.
Evaluation criteria
- The report should be concise, while retain important materials such as motivation and justification of the models and methods, interpretation of results, discussion of the validity of your results and conclusions, etc.
- English writing is important. Please make sure you write clearly and logically and make sure that readers can understand your report easily. It's wise to revise your report several times before submitting it.
- Make sure that your statistical analyses and conclusions are convincing and are understandable. It would be desirable to explain why you choose certain models/methods and whether these models/methods are valid in your particular case, and make sure that your conclusions make sense.