POLI 300                                             PROBLEM SET #12B                         no later than 05/21/09

(replacing PS #14C)

SPSS CROSSTABULATIONS

1.         Public opinion researchers generally believe that individual presidential approval responses are highly “colored” by party identification — that is, Republican party identifiers tend to “approve of” Republican Presidents but not Democratic ones and conversely. More formally, researchers expect to find a strong association between PARTY ID and PRESIDENTIAL APPROVAL, the direction of which depends on the party affiliation of the incumbent President.

You have previously tested this hypothesis using the Student Survey data (Problem Set #10). Now test it using 1972-2004 ANES/SETUPS data.

            (a)       Does a simple crosstabulation of V29 (PRESIDENTIAL JOB APPROVAL) by V09 (PARTY IDENTIFICATION) provide an adequate test of this hypothesis?

            (b)       Explain why the answer to the previous question is “No” and then explain how you must refine the analysis in order to test the hypothesis appropriately.

            (c)       Carry out this more refined analysis by running one more appropriate crosstabulations.

            (d)       Write a brief paragraph or so summarizing your findings.

2.         Pose a research question that can be addressed by constructing crosstabulation(s) of variables in the 1972-2004 ANES/SETUPS data. Typically, such a research question might pertain to the influence of some demographic or other (independent) variable on (certain) political attitudes or behaviors (dependent variables). It may appropriate consider data for a single election year (e.g., 2004) only. Run the appropriate crosstabulation(s) and then write up your findings in a brief report of several paragraphs.

Note: For both questions, show the actual crosstabulations. (You may do this by attaching copies of appropriately formatted and labeled SPSS output.)