Step 1: Selecting a topic
Choose your group and select a topic. You may design your own project or select one for which data is already available on the Internet. You can start with the sites below. But feel free to do your own web search.
http://www.statsci.org/datasets.html
http://lib.stat.cmu.edu/DASL/ (Click on data subjects)
http://www.bls.gov/data/home.htm
http://catalog.data.gov/dataset
http://www.census.gov/compendia/statab/cats/health_nutrition.html
Step 2: The proposal:
You will need to submit a project proposal that includes a clear description of what your group wishes to accomplish. What questions are you trying to answer? What are the relevant variables? What are their measurements (think variation)? If you intend to collect your own data, include a description of your statistical design and sampling plan. If it's a survey, include the questionnaire. If you are using available data, indicate the website you will be downloading the data from and include a brief description of the data set. How was it collected? Say something about the source (governmental data, private source, etc....).
Step 3: Data collection and data exploration:
After you collect your data you will probably want to enter it in EXCEL. Make sure you save the data set. Explore your data! Submit a report that glues together the following.
- An expanded version of the proposal from step 2 in which you will describe the real life problem of interest as well as the specific question (or questions) you are trying to answer. You will also provide a description of the method you used for collecting the data. Did you use an experiment? Why and what type of experiment is it? A survey? Include a copy of the questionnaire. If you rely on available data, be sure to describe how the data were collected and the reliability of their source.
- The appropriate graphical displays (histograms, boxplots, timeplots, scatterplots, qq plots, ...) as well as the relevant numerical descriptions of the data.
- A coherent narrative that describes your initial findings.
Step 4: Statistical inference:
You will very likely have to use an inferential method to answer specific questions. You will need to select the appropriate method. Do not use a method simply because it was taught in class. This is the formal stage of the analysis. You will have to do such things as stating hypotheses, conducting statistical tests, estimating parameters, checking assumptions, etc....
Step 5: The final report:
At this stage, you will essentially be adding two things to your report from Step 3:
- The results of the inferential stage: make sure you describe the procedure and you check the underlying assumptions (normality, for example). State your conclusions.
- Write an overall conclusion for your study. Self-critique your analysis. Mention shortcomings and biases. In general, in writing your conclusions, comments, and critiques, use language a non-statistician can understand (avoid mathematical symbolism as much as possible).
The final project report should be 4 to 6 pages long.