1. Fill out survey (5).
2. Bring 2 formatted 3.5" disks to class.
3. SPSS, ch. 3, Statistical Concept Questions 1, 2. Ch. 4, Statistical concept questions 1, 2, 3.
4. Computer Analysis. Using the nom434 dataset
run a frequencies on jparty. Print the output and discuss your results.
Run descriptives on qual and ideol. Print the output and discuss your results.
1. Enter survey responses into spss. Save on floppy (a:) with
filename yourlastname.sav where you use your last name (up to 8
characters) for yourlastname. Using felt-tipped pen, write your last name
2. SPSS ch. 7, stat concepts 1, 2.
3. Choose two variables from our survey (survey1.sav) that you think are related to one another.
a. what is your dependent variable?
b. what is your independent variable? Why do you think it will have an effect on your dependent variable.
c. run a crosstabs between your variables, including the c2 statistic. Present the table.
d. discuss your results.
1. SPSS ch. 8. Stat concepts 1.
2. Using NOM434.SAV, run and print a scatterplot of novotes and interest group opposition. Include the linear regression line and the r-squared statistic.
a. what are the independent and dependent variables?
b. why might you expect the independent variable to affect the dependent variable?
c. discuss your results.
3. Using our survey data, create a clustered bar chart comparing groups of cases. The clusters will represent the mean of sctjob, the category axis will be ideology, and define your clusters by sex. Print your chart and discuss your results.
For all assignments, either print up your results (preferred), save them to a floppy, or write them down on paper.
1. Using the class survey, recode party into a 3 point scale. Write down what categories you collapsed. Choose and run a crosstab using the new party ID variable. Discuss your results.
2. Select THOMAS votes from conf334.sav. Recode ADA scores (0=most liberal; 100=most conservative) into high, medium, and low categories. State how you collapsed the categories. Run a crosstabs using vote by ada. Discuss your results. Note: for people working at home, the conf334 dataset is too large for the SPSS student software. One option is to use a university computer to open the conf334 dataset and select the Thomas cases using the “delete” option. Then save the Thomas cases to your floppy. Alternatively, and less preferably (because you don’t get to practice selecting cases), you can copy to floppy the Thomas.sav file from the class directory, or download the the Thomas.por file from the course web site.
Using the Vin334 database, choose a Justice from which to
examine the relationship between his preliminary vote (e.g., cert vote) and his
report vote (i.e., his vote on the merits).
Note: the variable name for each justice's preliminary begins with the letter p followed by the justice's abbreviated name. For example, Douglas's preliminary vote uses the variable name pdoug. Vinson's preliminary vote uses the variable name pvin. The report votes use the letter r followed by the justice's abbreviated name (e.g., rdoug, rvin). The Pvotes are coded G=grant, 0=deny. The Rvotes are coded R=reverse, A=affirm.
The assumption for the analysis is that the justice's desire to reverse or affirm at the merits (as measured by whether he actually reverses or affirms at the merits) influences whether he votes to grant or deny at the cert stage. (This assumption will make substantially more sense after you have read chapter 5 in Segal and Spaeth.)
Run the appropriate analysis and discuss your results.
1. Using the class survey, RECODE k1...k10, so that you can correctly COMPUTE a new variable "know" that will be the total number of correct responses. Run DESCRIPTIVES on the new variable "know". Then using the independent samples t-test, determine whether knowledge levels are significantly different for males and females in our survey. Print, save, or copy your output and discuss your results.
2. Using the nominee data set, determine whether voting in civil
liberties cases (CIVL) differs for Republican and Democratic justices.
Repeat the analysis for economics (ECON) cases.
Discuss your results. Print, save or copy your output and discuss your results.
1. Using the nominee data set:
a. run a scatterplot using ideology and votes in civil liberties cases. Discuss your results.
b. run a regression using the same variables. Discuss your results.
c. What is the relationship between the plot in (a) and the statistical output in (b)?
As always, present your output in one form or another.
1. Run a regression predicting the number of no votes each nominee
received from his/her qualifications and his/her ideological extremity.
Note: We do not have an ideological extremity variable. To do this regression we need to create one from the variable IDEOL. The new variable must have the characteristic that the more extreme the nominee is (i.e., the closer the nominee is to either +1 or -1), the higher his/her ideological extremity score will be.
The problem with using IDEOL as is is that those justices who are ideologically extreme are at the high (close to +1) and low (close to -1) ends of the scale. We need to transform (using COMPUTE) IDEOL into a new variable such that low scores on IDEOL (i.e., those close to -1) become high scores on the new variable, while high scores on IDEOL also retain high scores on the new variable. In other words, we need a mathematical function that turns relatively low negative numbers into relatively high positive numbers.
There are at least two common mathematical functions that will do that, and it's part of your job to figure out at least one of them. I will simply explain how they can be run (once you figure out what they are).
One of them can be accomplished simply by writing out the mathematical expression in COMPUTE VARIABLE dialog box; the other option requires choosing the appropriate function from the FUNCTIONS box.
Two mathematically incorrect examples follow. For instance, if you wanted to create a new variable EXTREME that simply doubled the value of ideology, you would put EXTREME in the Target Variable box and either Ideol + Ideol, or Ideol * 2 in the Numeric Expression box.
Or if you wanted to create a new variable EXTREME that was the square root of ideol (which you obviously can't do since you can't take the square root of a variable with negative numbers [unless you're dealing with imaginary numbers]), you would put EXTREME in the Target Variable box, place the cursor in the Numeric Expression box, click on SQRT in the in the Function box, and enter Ideol inside the parentheses.
Many of the abbreviated functions in the Function box should be obvious. But if you want to be sure on any of them, click on the Help box at the bottom of the COMPUTE VARIABLE box, and then click on the highlighted Function box term. You will get a menu that describes each of the commands. Finally you may want to review pp. 550-554 of the SPSS book.
Now that you've created the new variable, discuss your results and present your output.
I have put a new dataset up:SOLGEN.SAV. The dataset contains data on each case in which the Solicitor General filed an amicus curiae brief on behalf of the petitioner or respondent in all Supreme Court cases between the 1953 and 1981 terms of the Court. The variables in the dataset are the citation of each case, the side (petitioner or respondent) supported by the Solicitor General (SGSIDE), whether the Solicitor General won or lost (SGWIN), and the number of justices on the Supreme Court appointed by the incumbent president’s party (SUPPORT).
1. Run a logistic regression predicting whether the solicitor general won or lost each case (sgwin), using the side supported by the solicitor general (sgside) and the number of justices on the Supreme Court appointed by the incumbent president’s party (support) as independent variables.
2. Do you expect the SG to be more likely to win when he supports the petitioner, or when he supports the respondent? Explain. (Hint: see chapter 5 of Segal and Spaeth)
3. Do you expect the SG to be more likely to win when when there are more justices from the incumbent president’s party, or when there are fewer? Explain. (Hint: see chapter 8 of Segal and Spaeth.)
4. Discuss your results.
The Spaeth database has two variables needed for this homework assignment: MOW (majority opinion writer) and MOA (majority opinion assigner).
1. Select cases from the 1975 through 1980 terms (e.g., term ge 1975 and term le 1980).
2. Run a frequencies on MOA. Discuss your results (and don’t just give the numbers. Explain why we get the results we observe.)
3. Run a crosstabs using MOW and MOA.
a. Which is the independent and which is the dependent variable? Explain.
b. Discuss your results. Do the results suggest an attitudinal assignment pattern by Burger? By Brennan? Explain. (Hint: to answer this question you need to know how close the justices are ideologically to one another. One general source of that information the ideology scores in the nominee dataset. Another source is the voting scores in the same data.