Political Science 104 Wednesday, November 18. Agenda Office Hours Assignment #5 Discussion on the Final Readings/Lecture Recap Guide to the Paper Working our Way Through Assignment #6 Assigned Readings Tufte, Visual and Statistical Thinking Shively, Craft , Ch. 10
Wednesday, November 18
Professor John Woolley (American Presidency) is holding a talk for undergraduates tomorrow night (8pm) at the Faculty Club. The topic is “Does Barack Obama Prove Our Democracy Works?” There will be free food.
The UCDC Program is accepting applications for Spring 2008. The application is due on January 15, 2009.
This assignment was fantastic! Little to no complaints…
If you’re ever asked to come up with a hypothesis and how you would test that hypothesis on a test… you should be as detailed as possible. So, introduce the hypothesis (which you all did), extract concepts, then extract variables. Testing has to do with variables, so you need to go this far when you’re asked to test a hypothesis.
Tells the difference between expected and observed. Only with a 2x2 table.
Expected Value = Row x Column _
# of Observations
How much of our dependent variable is being explained by independent variable. Only with a 2x2 table.
Q = (B*C) – (A*D)
(B*C) + (A*D)
How much of our dependent variable is being explained by independent variable. Can use more than a 2x2 table.
SPSS can calculate this. You won’t be expected to calculate this by hand.
Tells the difference between expected and observed. Can use more than a 2x2 table. SPSS can calculate this. You won’t be expected to calculate this by hand.
3. State and discuss what you currently believe is the most likely explanation or interpretation of the relationship between your independent and dependent variables, and describe the model you need to estimate to test whether this interpretation is supported by the data. (This will involve specifying the appropriate null hypothesis for your research hypothesis.)
This section is asking you to come up with CONTROLS.
Here’s what a control variable is:
In this sense, you have to establish what your main hypothesis is: High levels of personal wealth means that people will participate more. The null hypothesis is the opposite of your main hypothesis: Levels of personal wealth have no effect on political participation.
IF the control variable makes this relationship spurious (meaning that the above diagram is true), then you accept your null.
How do you control??? In crosstabulations and regressions!
3. Sketch out at least three alternative hypotheses to your expected interpreation, and briefly discuss how you would go about testing these. These should involve primarily challenges to the interal or external validity of your causal explanation (for instance, potential sources of “spurious correlation,” or intervening or antecedent variables) – that is, they should be alternative theories that you could test with data.
This section asks you to think of three different relationships (You can change the dependent, independent OR BOTH variables on your original relationship). You’re testing variations of your original hypothesis. So use your creativity here. If you want to test the control variables effects on your dependent variable, you could actually do that, or you could also come up with completely new relationships. So just develop the hypotheses here: High levels of education effects political participation.
4. Discuss what statistical methods you will use to test your causal hypothesis formally.
Now that you’ve specified three new hypothesis, how would you test these. You don’t have to do this for the assignment, but you do have to do it for the paper. This is basically outlining what you would do for these new hypotheses with the same process in #1 and #2 of this assignment.
How far should I go? Should I write my whole paper?
I advise you to write as much of your paper as possible. At minimum, you’ll only be graded on these five questions and how well you carry them out. However, this is your last chance to get written comments on the paper. If you write a full draft of your paper, I can read it and return comments to you the following Tuesday. You then have a two-day turnaround to revise the comments, which is ample time if you’re turning in a draft.
If you’re just turning in bare bones, you might have a hard time integrating my comments into the work you might have done in that week.
Think about this strategically. You can only benefit from having me read as a complete of a draft as possible.
You don’t have to turn in a complete draft. I will look at graphs, outlines, and bullet-pointed statements. You can fill in details later.