Sabato's Crystal Ball

The Political Science Election Forecasts of the 2016 Presidential and Congressional Elections, Part 3

UVA Center for Politics August 25th, 2016

facebooktwittergoogle_plusredditpinterestmail

Dear Readers: This is the latest in a series of political science forecasts for the 2016 races for the White House and Congress. We’ll be featuring forecasts from nine different individuals and/or teams this year, which James E. Campbell is assembling as part of a project for PS: Political Science and Politics that we are also featuring in the Crystal Ball. These models are based on factors such as the state of the economy, polling, whether an incumbent president is running for reelection, and other indicators. They can often be a better predictor of the eventual results than polls alone, and many are finalized months before the election.

We are pleased to feature the work of the many top political scientists who have built these models, both in an attempt to predict the outcome of the election and, more importantly, to identify the factors that actually affect presidential and congressional races. Below, Campbell lays out the details and outlook of his congressional forecasting model. Additionally, we have updated our running tally of presidential and congressional forecasts to include a new version of the presidential forecast by Robert Erikson and Christopher Wlezien, as well as the forecasts for Alan Abramowitz’s congressional model that are discussed separately in this week’s Crystal Ball.

— The Editors

The Seats-in-Trouble House and Senate Election Forecasts

By James E. Campbell of the University at Buffalo, SUNY

Description:

The twin Seats-in-Trouble forecasting models of congressional elections predict the net two-party aggregate seat change for the Democratic Party in the House and Senate, respectively. These are essentially hybrid models combining the virtues of intensive district-by-district and state-by-state in-depth handicapping of congressional races with the value-added rigor of the careful construction of an aggregate index and the systematic statistical evaluation of its historical relationship to seat changes in past elections.

The models use the late-August competitiveness ratings of congressional races from the Cook Political Report. Based on an evaluation of the correspondence of these individual race ratings to their election outcomes in the past, indices of each party’s number of vulnerable seats were calculated. The aggregate difference between the parties was used as an index and then included in a regression analysis to estimate its relationship to actual net seat changes. The seats-in-trouble index might be regarded as an extension of Oppenheimer, Stimson, and Waterman’s exposure model (1986). Rather than the number of seats nominally in jeopardy, the seats-in-trouble index counts those seriously threatened.

Both models have proven quite accurate in real-time application. The House model has been used since the 2010 election (though simplified in 2014). It was quite accurate in predicting the 2010 Republican wave and off by only six seats in 2012 and three in 2014. First used in 2014, the Senate model’s forecast missed the 2014 Republican nine-seat gain by a single seat.

Predictors:

For House elections, the seats-in-trouble index is the difference between the number of current Democratic seats that are rated as only leaning Democratic or worse and the number of current Republican seats in similar jeopardy.

For Senate elections, the index is slightly more stringent. It is the net number of a party’s seats (Democrats minus Republicans) rated as toss-ups or worse. The indices use ratings available in late August of the election year.

The Forecasts:

Based on seven House Democratic seats being rated as only leaning to the Democrats, toss-ups, or tipped toward going to the Republicans, and 33 House Republican seats being rated as only leaning to the Republicans, toss-ups, or tipped toward going to the Democrats — a net of 26 more Republican than Democratic seats-in-trouble — the model predicts that Democrats will gain 32 House seats in November. This would bring the number of House Democrats up to 220 members, two seats more than required for a bare majority. The forecast was made on Aug. 18, 2016.

Based on one Senate Democratic seat being rated as a toss-up or tipped toward the Republicans and eight Senate Republican seats being rated as toss-ups or tipped toward the Democrats — a net of seven more Republican than Democratic seats-in-trouble — the model predicts that Democrats will gain seven Senate seats. This would bring the number of Senate Democrats (including two Independents who caucus with Democrats) up to 53 seats, a majority. The forecast was made on Aug. 19, 2016.

Uncertainty:

Based on the distribution of out-of-sample errors, the probability of a Democratic House majority (the model erring in favor of the Republicans by more than 2 seats) is 61%. The probability of a Democratic Senate majority is 88% with the election of a Democratic president (a 50/50 Senate split) and 73% with the election of a Republican president (a 51D/49R Senate split).


Table 1: Forecasts of the 2016 two-party presidential vote

Table 2: Forecasts of the 2016 U.S. House election

Table 3: Forecasts of the 2016 U.S. Senate election