Sarah Anderson Methods Seminar May 21, 2004 The paper from the beginning of the quarter laid out the ideas I have for testing the pivotal politics model. With appropriations, the reversion point (or status quo) can be considered zero. Then appropriations for a given policy area should track with the preferences of the median voter, as measured by interest group scores. This test can also examine whether policy outcomes are influenced by the committees and by the parties. One of the main questions that arose in the seminar was simply descriptive. How do these appropriations change over time? Do they appear to be related to interest group scores? How much do the median House member and the median committee member differ? Below are several graphs depicting these relationships for two policy areas, agricultural subsidies and environmental spending). These areas were chosen, not necessarily because they are representative, but simply due to my particular interest in these areas. The first graph illustrates one of the main statistical issues to be dealt with: serial correlation. Clearly, spending on agricultural subsidies (and basically everything else) is increasing over time. In general, the budget of the federal government has been growing. There are several ways to deal with this problem. In my field paper, I included a lagged dependent variable, which eliminated the serial autocorrelation. I have also considered using spending as a percentage of the total budget, which has less of a time trend. The second graph shows agricultural spending as a percentage of the total budget. Another alternative is using first differences.The third and fourth graphs illustrate the relationship between spending (or percent of budget spent in the fourth graph) and median interest group scores. These graphs suggest that 1995 and 1996 may be outliers. Nonetheless, there does seem to be some relationship between the spending and ideology.Finally, the fifth and sixth graphs demonstrate the changes in the interest group scores of the median member of the House and median member of the relevant committee (either authorizing or appropriating). The solid line represents the median of the House and the dotted line the median of the committee. The two medians often move generally together, but there are some periods where this is not the case. Moreover, the committee medians are generally and understandable more volatile.The rest of the graphs illustrate these same features for environmental