Previous: Format:

Up: FREQ

Next: Output:

Purpose:

FREQ counts the proportion of seats expected to exhibit different specified forms of voting behavior, under numerous specific conditions. A few examples of the command's uses include:

1. Calculating the total number of seats expected to go Democratic. To do this, you would type FREQ 0.5 1.0;--that is, the frequency of districts with a Democratic two-party vote proportion above 0.5.

2. Calculating the number of expected ``marginal" districts, defined as all districts in which the victor gets less than 55 percent of the vote. To do this you would type FREQ 0.45 0.55;, because any district with a Democratic proportion of the two-party vote lower than 0.45 would represent a Republican victory above the cutoff for marginal status; any with a vote higher than that obviously represents a Democratic victory above the cutoff.

3. Calculating the number of incumbents expected to win reelection. To do this you would type READ INC!var $<$ dataset;, where INC!var is a vector containing 1 when the district has a Democratic incumbent, $-$1 when the district has a Republican incumbent, and 0 when it has either an open seat or two incumbents paired against each other.

4. Calculating the proportion of black incumbents predicted to win. You would type READ black $<$ dataset;, where black contains a 1 if the Democrat is a black incumbent, a $-1$ if the Republican is a black incumbent and a 0 if neither candidate is a black incumbent.

These examples do not exhaust the possible uses for this command. As always, the command may be used in cases of evaluation, prediction and counterfactual evaluation, as determined by your use of the read data commands. Note that in cases of evaluation, the expected proportion of seats won in general will differ from the actual proportion of seats won.