---

The Datafile:

• parole.dsk : Data File.

Open the data file. You are interested in these two variables:

• race or ethn : A = Native American, B = Black, M = Mexican, W = White
• parole? : D = Denied Parole, R = Received Parole

Examine:

You want to examine the association between race/ethnicity and parole decisions. Select race as Y and parole? as X , then under Calc choose Contingency Tables. Specify the information you want shown in the table by going to the Calculation Options - Table Options checklist. You will at least want to see Percent of Row Total and the Chi-Square value.

The Contingency Table:

Examine the information presented in the contingency table ‡. Interpret the results in terms understandable to the judge hearing the case.

• Compare the percentagees of parole denials for the various racial/ethnic groups.
• What is the null hypothesis being tested?
• Interpret the P-value.
• What is your conclusion?

The Lawsuit:

Now compare the parole decisions for the groups who brought the lawsuit to the decisions for the others.

• First create a new Bernoulli variable having a value of 0 if the inmate was Black or White or 1 if the inmate was native American or Mexican-American:
  • Data → New Derived Variable.
  • Name the variable A.or.M then define it with the

    Formula:

    (textof('race or ethn')="M") OR (textof('race or ethn')="A")

• Select A.or.M as Y and parole? as X, and create a new contingency table and Chi-Square analysis ‡.
• Interpret these results and write a conclusion for the judge.

The Judgement:

The judge in this case decided that there was no evidence that anyone was treated unfairly. Does that agree with your finding? (Should courses in statistics be required in law schools?)