User's GuideThis user's guide describes how to modify the parameters in the complete spreadsheet model and analyze their impact on the projections.
The spreadsheets use the same algorithm for making projections. For details of the model, check out the file, algorithm.doc, which explains each parameter in greater detail.
If you have questions or problems with the spreadsheet, please contact:
FairVote- The Center for Voting and Democracy
Part 1: Monopoly Politics 2004 Projections
You can do 4 things with this spreadsheet:
1. Modify parameters to examine their impact on projections
2. Enter 2004 results
3.Views summaries of projections by category and party
4. View summaries of the accuracy of the projections
1. Modifying parameters
The spreadsheet contains all of the data used to make projections: partisanship of the district, which seats are open seats, date of election of incumbents, and prior 2 election results.
It also contains the following parameters that the user can modify by typing a new number into the specified cell:
Parameter Cell Default value Range
Buffer for open seats A16 11% 0-25%
Increase for "underachieving" first-termer A15 33% 0-100%
Decrease for "overachievers" new incumbents A14 67% 0-100%
Absolute reduction for "underachieving" incumbent A11 1% 0-25%
Decrease if previous race uncontested A13 35% 0-100%
Decrease for "overachieving" incumbent A10 60% 0-100%
Adjustment to better 2nd election A12 33% 0-100%
Competitive J10 3% 0-30%
Tight J11 5% 0-30%
Comfortable J12 10% 0-30%
Landslide J13 >10% 0-30%
Dem 2-party share J15 50% 0-100%
If you enter a value outside of the range of a parameter, the projections may not be meaningful.
2. Enter 2004 results
The spreadsheet is set up so that you can enter the winning party (D, R or I) and percent in columns V and W, rows 21 to 455. Once you do that, summaries showing the accuracy of the projections will automatically appear.
Be sure to enter the winning percentage as a percent (ie, 43.7%, not 43.7, which would appear as 4370%). To enter a party, use a capital letter: D, R or I.
3. View summaries of projections
A variety of summary information about the projections appears in yellow-shaded cells in columns N-P and Q-T in rows 8 to 16 of the spreadsheet.
Columns N-P: Total number of projections by category
Landslide: The number of projected landslide victories.
Comfortable: The number of projected comfortable victories.
Tight: The number of projected close wins.
Vulnerable: The number of "vulnerable” incumbents.
Competitive: The number of districts in which we do not project a win.
Other: This is a check for errors. It is the number of districts that do not fall in the above categories. If it is not zero, there is a problem with the model or the parameters. Please contact the Center if this happens.
Columns Q-T: Projections by party
This table shows the projections by category for seats held by each party. These figures are based on the incumbent party, and since this is a redistricting year, some of the new seats don’t have incumbent parties. Also, one seat (VT-1) is held by an Independent. That’s why the total number of seats is less than 435.
4. View summaries of the accuracy of projection
Columns V-W: Totals of vulnerable, losing and open seats
Vulnerable: This lists the number of projections labeled as “Vulnerable” and, once data on the 2004 results is entered, the number of those incumbents who lost.
Losers (not open): After 2004 results are entered, this will show the number of losing incumbents.
Open seat party changes: The number of open seats whose party changed as a result of the 2004 election.
Open seats: The number of open seats this year.
Columns X-Y: Accuracy of projections by category
These columns remain zero until 2004 results are entered.
For each of the projection categories, these columns list the number of projections in which the party projection was accurate and the number in which the winning margin was correct.
Columns Z-AA: Results of races by prediction category
For each prediction category, these columns show the number of races that ended up being landslide, comfortable, wins or losses (from the perspective of the incumbent party).
The cells highlighted purple indicate races where the winning margin or party was wrong.
For example, with the default settings, there are 100 projected comfortable races. After entering the results (party and winning percent) from the 2004 races in column N and O, you will see in columns V-Y the results in the 100 projected comfortable races. Some of those races will be won by landslide, others by comfortable margins and others by less than comfortable margins. Any remaining races in this category would be losses for the incumbent.