Skip to main content
ABC News
New Hampshire Is Tiny And Pretty Weird. That Could Help Maggie Hassan.

We remain in something of a summer doldrums for polling, and the overall outlook for November remains about the same as in recent weeks. But polls on the race for Congress have continued to inch slightly toward Democrats in what may reflect the impact of the Supreme Court’s decision to overturn Roe v. Wade.

In the Deluxe version of our midterm election forecast, Republicans have a 85 percent chance to win the House and a 51 percent chance to win the Senate, both largely unchanged from when we launched the model three weeks ago. Meanwhile, in the Classic version of the model, which sticks to purely quantitative factors and leaves out the expert race ratings published by the Cook Political Report and other such groups, Republicans are actually underdogs to win control of the Senate, with a 39 percent chance.

Part of the reason Republicans are not better positioned in the Senate is they don’t have a lot of easy pick-up opportunities, but one race where Republicans could gain traction is the New Hampshire Senate race. Democratic incumbent Sen. Maggie Hassan has led in all but one poll this year against a series of potential Republican opponents, but the GOP shouldn’t write the race off. Hassan’s lead is small in most of these polls, and it’s still early, with the primary not scheduled until Sept. 13.1

New Hampshire is also an unusual state in that it has some factors that should make it relatively fertile ground to defeat an incumbent, and others that should make it a challenging one. So let’s take a look under the hood.

Americans are pessimistic, and it’s changing our politics

The factor that could help Republicans most is that New Hampshire — famous for polling leads that seem to evaporate overnight, like Barack Obama’s against Hillary Clinton’s in the Democratic primary there in 2008 — is notoriously swingy. I don’t just mean that it’s a swing state, although it frequently is in general elections. I also mean that it has a lot of swing voters: moderate, independent voters who split their tickets. Consider that, in 2020, Democratic Sen. Jeanne Shaheen won reelection there by 16 percentage points, while Republican Gov. Chris Sununu was reelected by 32 points on the same ballot.2

FiveThirtyEight measures this with our elasticity index, which is derived from voter-level survey data3 to reflect how sensitive a state is to national trends. New Hampshire has an elasticity index of 1.21 — the fourth-highest in the country — which means that for every 1-point shift in the national environment, we’d expect New Hampshire to shift by about 1.2 points instead. So, for instance, a 5-point swing toward Republicans nationally would produce a 6-point swing there. That makes Hassan a bit more vulnerable.

What makes a state swingy? Generally speaking, the most reliable group of Democratic voters are Black voters, and the most reliable group of Republican voters are evangelical white Christians. New Hampshire is quite secular and quite white, so neither party’s base is very large there.

Likewise, most other high-elasticity states are relatively nonreligious and have relatively few Black voters, although states with large numbers of Latino voters can be pretty swingy, as Democrats are discovering to their dismay.

New Hampshire is a swingy state

FiveThirtyEight’s 2022 state elasticity scores

State Elasticity Score
Rhode Island 1.29
Hawaii 1.23
Alaska 1.22
New Hampshire 1.21
Vermont 1.15
Massachusetts 1.14
Maine 1.14
Utah 1.11
Connecticut 1.10
North Dakota 1.09
Colorado 1.09
Wyoming 1.08
New Mexico 1.08
Iowa 1.07
Arizona 1.07
Washington 1.06
Nevada 1.06
Wisconsin 1.05
West Virginia 1.04
Indiana 1.04
Michigan 1.03
Nebraska 1.03
Kansas 1.03
South Dakota 1.03
Ohio 1.02
Florida 1.02
Idaho 1.02
Arkansas 1.02
Texas 1.01
New Jersey 1.01
Minnesota 1.00
California 1.00
Pennsylvania 0.99
Oregon 0.98
New York 0.97
Illinois 0.96
Montana 0.96
South Carolina 0.96
Oklahoma 0.96
Kentucky 0.95
Missouri 0.94
Tennessee 0.94
Virginia 0.94
North Carolina 0.93
Louisiana 0.92
Georgia 0.90
Maryland 0.87
Delaware 0.86
Mississippi 0.85
Alabama 0.83

Elasticity isn’t the only factor the model considers, however. Another is a state’s population. And in smaller states, the incumbency advantage is larger.

If you’re a Senate history buff, think about states that have a handful of electoral votes, like Vermont, West Virginia, Hawaii or North Dakota. They tend to have fairly long-tenured senators. Only four senators (Bernie Sanders, Patrick Leahy, Jim Jeffords and Robert Stafford) have represented Vermont since 1975, for example. (Although Leahy is retiring and not seeking reelection this year.)

Why might the incumbency advantage be larger in small states? I can think of a few theories. One is that, in a smaller state, it’s easier for a senator to maintain a strong, personal relationship with influential constituents and to be visible in her community. A second theory is about competition: In a smaller state, there are fewer highly talented politicians with national aspirations, so an incumbent will be up against weaker opponents, on average.

A third hypothesis: It’s all about the pork. Since each state has the same number of senators, a senator from a small state is more likely to extract concessions in exchange for pork-barrel spending,4 and this can make them more popular across party lines with voters at home (think Sens. Lisa Murkowski of Alaska or Joe Manchin of West Virginia). Whatever the reason, the statistical tendency of small-state Senate incumbents to overperform is fairly robust, and this is accounted for by our model.

What else helps an incumbent? The weirder a state, the bigger the incumbency advantage. OK, I should probably clarify what I mean by “weird.” A more precise term would be “dissimilar” or “distinctive.”

As part of our forecast, we apply a process called CANTOR5 that determines which states and congressional districts are most similar to one another according to a variety of demographic, geographic and political variables — such as race, religion, income and education, population density, latitude and longitude, and voting behavior in recent presidential elections. A state like Hawaii qualifies as distinctive by this process, for instance. It has a far bigger Asian American and Native Hawaiian population than other states, it’s very Democratic, it’s very expensive and it’s truly out there in the middle of nowhere in the Pacific Ocean.

New Hampshire counts as somewhat distinctive, by contrast. As I mentioned, it’s quite white and quite nonreligious relative to the rest of the country. It also has above-average incomes and education levels but below-average population density (in the way that FiveThirtyEight measures it). On the other hand, it is toward the middle of the road politically. Overall, it ranks as the 25th-most distinctive state. Michigan is the state which most closely resembles the rest of the country, by contrast.

New England is distinctive

States that are least similar to other states, according to FiveThirtyEight’s CANTOR

Rank State
1 Hawaii
2 California
3 Utah
4 Alaska
5 Vermont
6 Massachusetts
7 Wyoming
8 New York
9 Maryland
10 West Virginia
11 New Jersey
12 Mississippi
13 North Dakota
14 Rhode Island
15 South Dakota
16 Idaho
17 Maine
18 Arkansas
19 New Mexico
20 Connecticut
21 Washington
22 Texas
23 Alabama
24 Oklahoma
25 New Hampshire
26 Nevada
27 Montana
28 Louisiana
29 Florida
30 Delaware
31 Oregon
32 Tennessee
33 Kentucky
34 Illinois
35 South Carolina
36 Colorado
37 Georgia
38 Arizona
39 Nebraska
40 Virginia
41 Iowa
42 Minnesota
43 North Carolina
44 Kansas
45 Missouri
46 Pennsylvania
47 Indiana
48 Wisconsin
49 Ohio
50 Michigan

Factors that CANTOR considers include race, religion, age, gender, household income, educational status, voting behavior in recent presidential and state legislative elections, percentage of residents born abroad, population density, geographic region, latitude and longitude.

How could distinctiveness provide an advantage to an incumbent? In some ways, the answers are similar to why small states have a leg up. If a state has a population with unique concerns — say, religious minorities like Mormon or Jewish voters, or a racial minority like Native Hawaiians — incumbents can exhibit competence by serving those populations in ways that somewhat cross party lines and are less subject to the whims of the national political environment.

Another related theory is that there could be a stronger bond between voters and politicians in distinctive states. The politician may be a member of the minority group, or they may have ties to an industry that is uniquely important in the state. In the same sense that you probably know a married couple who are “perfect for one another,”6 there may be stronger matching in these cases.

But a statistical measure of distinctiveness is only as good as its inputs, and there are some things CANTOR leaves out. Being home to an industry like gambling in Nevada or coal production in West Virginia is potentially politically important, for instance, but CANTOR doesn’t take this into consideration.

Overall, New England is a culturally distinctive region that is hard to pin down statistically, but is visible in everything from its cuisine to its intense loyalty to the Boston Red Sox. My subjective experience is that New England is considerably weirder — excuse me, “more distinctive” — than our metric describes it, and that means New England tends to be quite loyal to its incumbents. Hassan will hope that pattern holds.7


  1. What the hell, New Hampshire?

  2. The contrast to a state like New Hampshire is a state like Georgia, which is also a swing state, but not because it has a lot of swing voters. Instead, it’s mostly a swing state because it has some very Republican-leaning groups like evangelical white voters and some very Democratic-leaning groups like Black voters and young, college-educated professionals, who roughly cancel one another out. Races in Georgia, then, tend to be more about who can turn out their base. New Hampshire is more about persuasion.

  3. Specifically, we use the Cooperative Election Study from 2014 through 2020, which contains demographic information and presidential voting preferences for tens of thousands of voters. From this data, we can then determine how many voters in each state are likely to vote for each party and how many are plausibly up for grabs.

  4. Of course, any senator can ask for concessions. But a pork-barrel project that makes a big dent in California or Texas’s finances is going to be pretty expensive. A small state is much cheaper to buy off.

  5. In the tradition of PECOTA and CARMELO, we created a stupid backronym for it: Congressional Algorithm using Neighboring Typologies to Optimize Regression.

  6. Often, of course, a backhanded compliment which means they’d annoy the living hell out of anyone but one another.

  7. Although New Hampshire itself has had reasonably high turnover in the Senate.

Nate Silver is the founder and editor in chief of FiveThirtyEight.


Latest Interactives