The political website FiveThirtyEight provides an innovative look at political forecasting and also supplies useful information about candidates and political districts.
The site FiveThirtyEight.com was active during the 2008 campaign season. Now it is a feature of the New York Times and can be accessed at fivethirtyeight.blogs.nytimes.com. The name comes from from the number of electors in the United States electoral college.
FiveThirtyEight uses a variety of methods to arrive at its results, including polls, where polls are weighted by several factors including recency, sample size, and the polling firm’s track record. Some polls are considered so unreliable that they are not included. The weighted polls results are adjusted by several factors, including a trendline adjustment and likely voter adjustment.
The data is further adjusted by factors such as the state’s Partisan Voting Index, individual monetary contributions received, and “a variable representing stature, based on the highest elected office that the candidate has held.”
There are additional steps in the analysis. Finally, the FiveThirtyEight procedures uses simulation, where various factors are considered randomly over a large number of trials.
When FiveThirtyEight reports its results, it also calculates the probability that a candidate will win the election. It might forecast, for example, that a candidate will finish with 55 percent of the vote, with the probability of winning at 85 percent. Winning, of course, means that the candidate gets at least one more vote than the closest opponent — no margin of victory is implied in the probability.
The site is also a useful repository of information such as voting record in selected issues, campaign finance, district demographics, and previous election results.
The FiveThirtyEight site doesn’t say this, but we can easily surmise that the lead that some candidates currently enjoy is the result of not only the policy positions of the candidate and the political landscape of the district, but importantly the product of the campaign the candidates have waged so far. Candidates with leads need to realize this and keep up their efforts.
FiveThirtyEight forecasts for Kansas
In Kansas, here are the results FiveThirtyEight forecasts:
For Kansas Governor: Republican Sam Brownback 60.5 percent; Democrat Tom Holland 37.6 percent. Probability of a Brownback victory is 99.9 percent. The vote difference has been narrowing very slightly, but the probability of a Brownback win is still overwhelming.
For U.S. Congress, District 3: Democrat Stephene Moore 42.5 percent; Republican Kevin Yoder 55.0 percent. Probability of a Yoder victory is 92.7 percent. This is the only Kansas Congressional district that is remotely competitive, described as “leaning Republican.” Yoder’s margin has been increasing very slightly.
For U.S. Congress, District 4: Democrat Raj Goyle 36.5 percent; Republican Mike Pompeo 61.0 percent. Probability of a Pompeo victory is 99.9 percent. Pompeo’s lead over Goyle has been growing since the September 17th version of the model for this contest. These results don’t include the SurveyUSA poll of just a few days ago, which showed Pompeo’s lead over Goyle widening.