The Primary Model, developed by Helmut Norpoth, is an extension of his electoral-cycle model. The model provides a long-term forecast of the election outcome based on electoral histories, plus the candidates’ performance in early primaries. In particular, it uses the vote of the two most recent elections as well as the candidates support in primaries as predictor variables in a linear multiple regression model, which has been estimated based on data from all elections since 1912. The model’s vote equation reads as:
V = A + b1
Vt-1 + b2
Vt-2 + b3
DPRIM + b4
|Table 1: Overview of variables used in the Primary model|
||Incumbent party’s popular two-party vote in the last election|
||Incumbent party’s popular two-party vote in the second to last election|
||Incumbent-party (Democratic) candidate primary support in New Hampshire and South Carolina|
||Opposition-party (Republican) candidate primary support in New Hampshire and South Carolina|
|V||Incumbent share of the two-party vote|
Assuming a hypothetical Clinton-Trump race, the model predicts Trump to gain 52.5% of the two-party vote, compared to 47.5% for Clinton.
The Primary Model has, with some modifications, been used to forecast each of the five elections since 1996. The following chart shows the model’s forecasts and the actual results in each election. On average, the model missed the final election outcome by 3.1 percentage points.
Norpoth, H. (2016). Primary model predicts Trump victory. PS: Political Science & Politics, 49 (4), 655–658.