I am still a bit confused. NEAS says that these are just ideas for our student project. Are we allowed to do the project on any topic we wish or does the project HAVE to be on 1 of the 3 topics given (i.e. Interest rates, Sports or Weather). When looking through past projects, it looks as if people did this project on any random topic. Can I for example do a project on Health Insurance Claims trends for the Senior Citizen polulation (i.e. the Medicare population)?
[NEAS: The project can be on any topic you want. Over 450 past student projects on various topics are posted on the discussion forum to give you ideas for your own project.]
Jacob: Which student projects are posted on the discussion forum? Are these the best projects?
Rachel: We post projects that give other candidates ideas for their own projects. We don’t post projects on baseball won-loss records or three month Treasury bill rates because the NEAS project templates show how to set up the data, test hypotheses, and form ARIMA processes or regression models. We choose student projects with new ideas, getting data from other web sites.
Jacob: Do the student projects on the discussion forum show the correct modeling?
Rachel: Generally yes, with several qualifications. Much statistics is choosing the best model. Sometimes two or three models might be chosen; none is necessarily a correct or incorrect model. As long as you justify your choice, you get credit for the student project. An AR(2) process may be a somewhat better fit but an AR(1) process is more parsimonious. Two competent statisticians might choose different models. Say why you chose one model over the other.
Jacob: If a project mis-interprets the data, does it still get credit?
Rachel: Statisticians and economists who deal with a topic may know details that a candidate doing a student project does not know. The candidate may not be aware of an exogenous influence on the time series or the response variable. The student project does not test if you know the details of the topic; it tests if you can apply the statistical techniques to actual data. We don’t subtract points for not knowing details.
Illustration: Suppose one looks at a time series of the number of babies named Elvis. One will see spikes: when Elvis Presley first burst on the rock-n-roll scene and again when he died. If one knows these facts, one should investigate them; but you don’t have to know these facts to do the student project. Many candidates list these external influences, and they better interpret the spikes and outliers in their data. But you get credit even if you know nothing about these influences, as long as you show the statistical work.
Jacob: The discussion forum shows the Word document of the PDF file, not the data or the Excel workbook. Don’t we have to submit both?
Rachel: Yes, both are required. The statistical work you must do yourself. You can pattern the workbook on the textbooks or the NEAS project templates, but you must set up the workbook (or other statistical file). For the write-up, you can read what others have done to get ideas for your own comments.
dem