POLS 6386 MEASUREMENT THEORY
Ninth Assignment
Due 11 April 2001


  1. Continuing our Lieberman project, you need to download all the Senate roll call votes taken to date from:

    U.S. Congress Web Site (Thomas)

    As of 3 April 2001 there were 64 votes cast in the Senate. Repeat the steps you took in part (3) of Homework 7 to get the roll call voting matrix. As a check on your results, you can download my version of the first 30 roll calls to check your SBIG107.PRN:

    First 30 roll calls of SBIG107.PRN

    1. Run SEN107Y.ORD through PERFL in one and two dimensions. Use SPSS to make a graph of the two dimensional results. Label Senators Lieberman, Helms, McCain, Feingold, Kennedy, Lott, Hagel, Wellstone, and Stabenow in the plot.

    2. Compare the results with last week's configuration. Has a week's worth of voting changed the level of fit or the structure in one or two dimensions?

  2. Merge the 107th roll calls into the combined data matrix for the 105th and 106th Senates -- SEN0506.ORD. To do this, use the program LIEBER.EXE and its control card file LIEBER.DAT that you downloaded for homework #8. To create the new SEN0507.ORD file simply change the number in LIEBER.DAT to the current number.

    1. We are going to start by running SEN0507.ORD through W-NOMINATE (WNOM9707.EXE). Download W-NOMINATE from my website

      W-NOMINATE Program and Example Files

      and upzip it into your \RCHTMS107\ directory. The control card file for WNOM9707.EXE is NOMSTART.DAT. It is set up for a matrix that is formatted differently than SEN0507.ORD. You need to use EPSILON to edit NOMSTART.DAT so that it looks like this:
      
      SEN0507.ORD
      NOMINAL MULTIDIMENSIONAL UNFOLDING OF 105TH TO 107TH SENATES
       1348    1    1
          2   19
       15.0000  0.5000
        0.0250   25
      (3X,9A1,7X,4A1,2X,6A1,5X,3600I1)
      (1x,I4,19A1,1X,4i4,51f7.3)
      (1x,I4,19A1,1X,51f7.3)
      The "1348" is the number of roll calls as of 3 April 2001. There will be more by the time this homework is due. Set this number to the correct amount when you run WNOM9707.EXE.

      Read all the example output files included in WNOMINAT.ZIP! and then report the following fit statistics for one and two dimensions: Percent Classified, Aggregate PRE, and Geometric Mean Probability.

    2. Use Excel to make a histogram of the distribution of roll call vote margins. These are in the output file NOM23.DAT. Below is the first part of NOM23.DAT. With 1348 roll calls, 273 had minority margins of .025 or less and were not included in the scaling. The distribution of the remaining 1075 roll calls is given below.
      
      NOMINAL MULTIDIMENSIONAL UNFOLDING OF 105TH TO 107TH SENATES    
       ROLL-CALLS READ=1348  NUMBER REJECTED= 273  NUMBER ACCEPTED=1075  CUTOFF= 0.025
       LEGISLATORS READ= 123  NUMBER REJECTED=   0  NUMBER ACCEPTED= 123  CUTOFF=  25
        DISTRIBUTION OF SCALABLE ROLL CALLS
         1 50 - 55     310  0.288
         2 56 - 60     229  0.213
         3 61 - 65     125  0.116
         4 66 - 70      89  0.083
         5 71 - 75      68  0.063
         6 76 - 80      60  0.056
         7 81 - 85      54  0.050
         8 86 - 90      43  0.040
         9 91 - 95      53  0.049
        10 96 - 97.5    44  0.041
      Hence, category (11) would be:
      
        11 97.5 -100.0 273  
      Make the histogram bars percentages. For example, 273 is 20.3 percent of 1348.

    3. Use Excel to make a graph of the eigenvalues of the double-centered squared distance matrix computed from the agreement scores from the combined matrix. These are in the output file NOM23.DAT.

    4. Use SPSS to make a graph of the two dimensional results. Label Senators Lieberman, Helms, McCain, Feingold, Kennedy, Lott, Hagel, Wellstone, and Stabenow in the plot.

  3. Run PERFL in one and two dimensions the new SEN0507.ORD.

    1. Report the one and two dimensional correct classification percentages.

    2. Use SPSS to make a graph of the two dimensional results of the combined scaling. Label Senators Lieberman, Helms, McCain, Feingold, Kennedy, Lott, Hagel, Wellstone, and Stabenow in the plot.

  4. Use HOUSYM2 to get an agreement score matrix for SEN0507.ORD. Run the agreement score matrix through KYST. Report the one amd two dimensional STRESS values and use Excel to compute the correlations between the one and two dimensional coordinates from KYST with those from PERFL and WNOM9707.EXE. Report the correlations one dimension at a time for each dimensionality using the format below:
    
    
          One Dimension Scalings
              PERF 1d    KYST 1d         WNOM 1d
    ------------------------------------------------
    PERF 1d  1		
    KYST 1d	-0.92682	1	
    WNOM 1d	-0.95408	0.990908	1
    
    
          Two Dimensional Scalings
     	1 PERF 2d	1 KYST 2d	1 WNOM 2d
    -------------------------------------------------
    1 PERF 2d	1		
    1 KYST 2d	0.99034	1	
    1 WNOM 2d	1	0.99034 	1
    
     	2 PERF 2d	2 KYST 2d	2 WNOM 2d
    -------------------------------------------------
    2 PERF 2d	1		
    2 KYST 2d	0.592688	1	
    2 WNOM 2d	1	0.592688	1
    
                      etc.