Mon Feb 26 12:19:43 2007 36-726 Spr 2007 Instructions for rough draft of stat practice papers: ---------------------------------------------------- -1. What to turn in. Email me a single pdf file, *before* the end of the day Mar 9. The pdf file should be readable on any current machine. If you use LaTeX, no sweat. If you use MSWORD, sometimes the pdf file depends on fonts that are only on your computer, so please check it out on a few other computers (e.g. a linux box) before turning it in. 0. Content. I am expecting you to A. blend together and improve the individual project reports that you gave me earlier this semester, into a single report for each project group B. report on additional progress you have made since you first turned in your individual reports. 1. Organization. Your paper should be organized in a way that makes sense for your project. * If you choose one of the organizational strategies in the handouts, articles, and examples we have discussed in class, I will probably be fine with that. These strategies should work for all but one or two of the projects we have this semester. * You are welcome to choose or invent another organizational strategy if you wish. If you choose a different organizational strategy, you need to have a good reason for doing so. If it "works" for me, great. If not, you will at least have to justify your structure and you may have to restructure the paper. 2. Completeness. Some sections will necessarily be incomplete. Following a typical organization I would expect: A. Introduction should be complete and highly readable. Should contain appropriate background, lit review if appropriate, questions, conclusions, outline of remainder of paper, etc., to prepare reader for remainder of report. Likely the paragraphs(s) summarizing the main results of the paper will necessarily be tentative, incomplete or just a promise that 'we will write this paragraph when we know the answer!'. That's ok. B. Body. - Data. Most of you have data, so I would expect a fairly complete description of the data, how it was collected, and what portion of it you have. [even if you do not yet have (all) your data, you should be able to write coherently about what you know at this point] - Analyses. Necessarily incomplete at this point: Report on what you have done, what you are doing, and what you will do. Tie these together, so that there is something like a story: "since the EDA reported above (see especially Figure 3, as we discussed above) suggests there are two groups of neurons (and this is consistent with the literature on this part of the brain; see for example Smedley and Smith, 1892), we are currently developing an E-M algorithm to classify neurons into the two groups and fit different firing time distributions to each group. If this is successful we will use our classification to test Dr. Jones' hypothesis that only one of the two groups of neurons are involved in XYZ..." C. Results and/or Discussion Sections. Also necessarily incomplete. So right now use this section/sections to - review what you have learned so far, - discuss other research questions or analytic difficulties that remain, - and place the work you have done so far in a bigger context: + what is the larger scientific enterprise in which your work resides, + how far do you expect to get by the end of the semester, + and if you do get that far how does it relate to the bigger scientific picture? D. Appendices. Start using them to put the stuff that is important but not germaine to the main story you are trying to tell. Don't forget to organize appendices in a way that makes it possible for readers to find just the parts they want to look at, and don't forget to add some narrative (english!) to help readers understand what is in the appendices. For most of you the appendices will consist of details of data analyses, glossaries of terms, summaries of the science that don't belong in the main paper, etc. Some of you may find that more detailed lit reviews or mathematical derivations will fit well in appendices. 3. Grading and feedback. A. I will read the papers myself over break and comment on them. B. I will ask each group to read and comment on one other group's report, in the week after break. C. Clearly, I will be looking for the things I listed above. In addition I will be considering issues like the following: - Most important are issues like: + organization: can different audiences easily find what they need? + invisibility: mechanics, voice, language, style, putting the topic out front, and and not putting the writer or the writing out front + good writing is good teaching: what have you done to help the reader (both me, who knows little of the area of your project, as well as the investigator you are working with, who knows little statistics) deeply understand what you are doing? - Other issues, especially those discussed in the Ally online materials, and in the Bem and the Gopen & Swan papers, come next. - Does it sound like you are excited and intellectually engaged in the work (how much intellectual effort the reader will put into your paper depends on this!) - Least important is whether you have covered every statistical base. I will encourage suggestions about other possible analyses than the ones you are reporting on or proposing to do, but at this early date I will not downgrade you for not having thought of them. (in the final draft I will look for more completeness on the statistics/analysis side...) ====================================================================