Math 217—Guidelines for Consulting-Project Paper

• Read carefully through the “Communicating the Results of Scientific Investigations” handout (taken from Grinnell), and note my specific comments. Heed all their advice and use their template format (title; abstract; introduction; methods/data collection; results; discussion; acknowledgements, if needed; appendix, if needed; and references).

• Some additional thoughts on sections of the paper:

o The introduction motivates the research questions. You learned some background from Mary, but you need to bolster this with a few additional references, which means you must do additional reading (and these should be included in your references list).

o Your knowledge of the methods/data collection comes directly from Mary. Be sure you understand exactly what she did and why she did it (email her if you have additional questions).

o Recall Mary wants you to answer her “big” research questions, but she’d also like you to data “snoop” and find any other interesting findings (that is, answering her big questions is a minimum requirement, but your analysis will be more helpful and thorough if you also look more widely at the data and give Mary some additional insights).

o So your Results section and Discussion section can be broken down into subsections (e.g., “Main Research Questions” and “Additional Findings”).

o In your Results section, include only the most pertinent information (don’t skimp, but don’t overload the reader). Within the section, include only tables and graphs that are integral—additional information can be included in an Appendix (and referenced from the Results section). But don’t use the Appendix as a big “data dump.” Only include things that are referenced (even briefly) in the Results section.

o For her big research questions, Mary would like inference made about the households of all three counties (that is, of the whole population of interest). This means 1) you must adjust for multiple comparisons (if there are too darn many tests, you might want to ask Mary if she can boil down her big questions more), and 2) you must try to verify that the sample (with only a 31% response rate) is representative of the whole population—if it isn’t or if you can’t verify this, then you must place a caveat on your inferences.

• Thorough data analysis takes a lot of time and effort. You will go down many paths—some fruitful, some not. You might be tempted to tell your reader the whole story of how hard you worked and how much time you spend. Don’t do this. Everyone knows that you spend a lot of time with your analysis. The main point of the paper is to distill all this hard work into the most pertinent and important findings. You will need to spend a lot of time thinking about how to convey your complex analysis into concise findings in an organized way.

• Please talk with me about your analyses, your paper writing (e.g., what to include, what not to include), and any other questions you have.

• Again, this is your opportunity to do “real” statistics. You are serving as consultants on what appears to be a cutting-edge project. It may feel uncomfortable to work “without a net,” but this is ultimately the best learning opportunity you can get.