Gary King Homepage Previous: FREQ Up: FREQ Next: Purpose:

Format:

FREQ [SAVE num] [{DELTA:val1 VBAR:val2}] [EXPECTED] {[READ var1 [var2] $ <$ dataset] [val3 val4]};. When you use the val3 val4 version of this command, JudgeIt will report the proportion of seats in the electoral system with a predicted vote proportion between val3 and val4. The READ var1 [var2] $ <$ dataset version of this command allows you to indicate directly the districts in which you are interested with var1, which should be a variable containing zeros for all districts that should be excluded and ones or negative ones for those to be included. You should use a 1 if you are interested whether the Democrat will win and a $ -1$ if you are interested whether the Republican can win. If you include the optional var2, districts with var2=1 are deleted (this is especially useful if you use DELIF, since you can name the same variable in both commands). The reported statistic is the proportion of those districts in which the candidate for which you indicated interest is expected to be successful. DELTA:val1 specifies the amount that the actual or predicted average district is to be changed to simulate the consequences of a statewide swing to a party (valid possibilities range from $ -1$ to 1, but reasonable values are usually much smaller, such as $ -0.05$ to simulate a 5% statewide swing to the Republicans). The other way to simulate the consequences of a statewide partisan swing is using VBAR:val2, where val2 is the value of the average district vote you wish to specify. The standard errors printed are those based on the total variation; use the optional EXPECTED command to get standard errors based on estimation variability in the expected votes. See the discussion under SVCURVE for a discussion of the different types of standard errors.

The optional SAVE num command enables you to save the results of each run; If you have a FREQ WRITE; command at the end of your file, Judgeit will all the print the results from each FREQ command in a convenient table with averages sorted by num and within that by year.



Gary King 2006-01-07