Fall, 2006 (Roos, Soc. 501)

Assignment 4: Analytic literature review (due Wednesday, November 29th)

Now's the time to get thee to a library (physically, or electronically) to orient your final project sociologically. Look for articles or books that address the substantive issues you have raised. What sociological literature(s) is(are) relevant to help you frame your statistical investigations? I don't expect that you'll have a fully developed literature review for this assignment, but I do expect that by your final project you'll address at least 10 relevant articles/books. The further you go for this assignment, the further along you'll be toward your final project.

This assignment requires that you write an analytic literature review. Include no more than one first paragraph that reminds me what your larger paper is about. You have several weeks to accomplish all this, so make substantial progress toward your final project. If you do it right, you'll be able to weave your literature review into your developing paper without too much angst.

To write an analytic review, don't just review what X, Y, and Z say about your topic. Don't summarize the entire article or book. There is nothing more boring. Rather use previous literature to develop your own argument. Evaluate (don't just describe) how others' work helps you to organize your own ideas. How do you build on previous research (how do you move the field forward)? How does previous research lead you to expectations about the relationships among your variables? Ultimately, you'll use your review to structure your introduction (and organize your research questions), talk about methods in the methods section, and frame your results in the analysis section. So this step in the process is a very important one.

Continue on in your spare time running data that will be relevant to your final project (but don't include in this writeup). Try other trivariate relationships. Are there any relevant 4-variable crosstabs you should try? Watch out for small n's in columns, as an indicator of when to collapse, or when to stop adding variables. In general, however, you don't need to recode a dependent variable to fewer than 7 to 8 categories (check with me if you have any questions). Independent variables should be from 3 to 5 categories, depending on the n's in columns. You'll need to justify substantively why you collapse the way you do.

Additional help: Ch. 15 and Appendexes B-D in Schutt can give you some hints on how to research and write a research report. Also Miller can be invaluable (esp. Chs. 11-12). Check out some recent journal articles (e.g., American Sociological Review, American Journal of Sociology, Work and Occupations). Finally, check out the last section of the syllabus for additional resources: "Research, Thinking, and Writing" (especially Clarke, on writing and criticism).

Assignments must be typed.