TO: Students in 605

FROM: R. B. Taylor

DATE: 2/8/06

RE: Comments on homeworks/answering questions to date

Have gone through the stuff you all handed in and have a few comments. It was generally pretty good. These comments are not offered as a general indictment, but rather to help focus your writing more. They are presented in no particular order.

1. When you introduce a variable, tell the reader how it works. If you just say "race is correlated" the reader has no idea what is going on.

2. When you explain a relationship -- a correlation or a b weight for example - explain the direction of the connection. Which group or persons score higher? In other words, be specific about what the relationship is telling you about. It is important to always know what a high score means. Don't just say "Women score higher on the index" say "Women as compared to men are less likely to think the police act fairly," for example.

3. I know these were just data you were playing around with, but when you start talking about the data you might be working with, you want to introduce the data file to the reader, not just jump into the results. You can start developing this methods type section of the paper you are going to be working on and just plug it in routinely when you do a weekly assignment. Or, at the least, for now, give the reader 3 - 6 lines on the data source.

4. Explain what the levels are in the data. Don't just say "at level 1" unless you have previously told the reader that level 1 corresponds (e.g.,) to residents and level 2 corresponds to (e.g.,) cities.

5. When you are done, try to say something general about what you have found. What type of general inference is warranted after you have taken us through the findings? What is the "takeaway" lesson?

6. Some of you all even had hypotheses stated beforehand - I think this will happen because of bla bla woof woof. Great idea!

7. REMEMBER that when you change levels your variable means something else. Some of you were talking about the Level 2 race variable like it reflected individual race rather than neighborhood racial composition.