Close Close

State of the 2018 House Race

California, National, New Jersey, Ohio, Pennsylvania, Virginia, West Virginia

Dems do better in Dem precincts than GOP in GOP precincts; Race, gender and education divide core support from persuadable voters

West Long Branch, NJ – The Monmouth University Poll finds the race for control of the House of Representatives to be tight in a new analysis of polls conducted in eight competitive bellwether congressional districts over the summer. These districts – which include a mix of suburban, rural, and industrial areas of the country – show that the electorate can be divided into groups of base voters and persuadable voters according to key demographics. Moreover, a precinct level analysis of President Donald Trump’s performance in 2016 is an indicator of support in these 2018 midterm races. These results paint a picture of the underlying fundamentals of the battle for control of Congress before voters fully engage in the fall campaign.


Summary

These are key findings from 3,390 interviews conducted in CA48, PA01, PA17, NJ03, NJ11, OH12, VA10, WV03 from June to August. Among all potential voters, support is evenly divided between the Democratic (43%) and the Republican (42%) candidates for Congress across the districts polled. The Democratic advantage increases slightly (47% to 43%) among likely voters. It is important to note that Republicans routinely won these eight districts by double digit margins in recent election cycles.

  • Where voters live has an impact on the margin of support. GOP House candidates are underperforming in Republican precincts relative to the Democrats’ performance in their base precincts. The Republican lead is between 4 and 13 percentage points in precincts that Romney/Trump carried, with the range depending on the size of the GOP presidential ticket’s margin. The Democratic lead is much stronger at 15 to 28 points in districts carried by Obama/Clinton. In competitive districts – those where the average margin was less than 5 points for either party’s presidential ticket – Democratic House candidates have a slim lead of 4 points. Also, the Democratic House candidate does better overall in precincts where Trump did worse than Romney even after controlling for the precinct’s partisan lean.
  • Race, education and gender define key voting groups. Republicans’ core voting bloc is comprised of white men without a college degree, while Democrats can count on strong support from white female college graduates as a well as women of color regardless of educational attainment. White women without a degree and white male college graduates are more competitive groups, as are, to a lesser degree, men of color. White men without a degree who are registered Democrats and women of color who are registered Republicans are the most likely to cross party lines in their 2018 House vote.
  • Partisan differences in election interest. High interest is more prevalent among voters supporting the Democratic candidates (62%) than it is among those supporting the Republicans (54%) in these eight races. The highest levels of interest come from college educated white men (75%) and women (72%) who are supporting a Democratic House candidate.
  • Strongly held opinions of Trump lean negative. While voter opinion of Trump is evenly divided at 49% approve and 48% disapprove, there is a negative gap when looking only at strongly held opinions – 33% strongly approve and 40% strongly disapprove in these eight districts.

“One thing we saw in the special elections to fill U.S. House vacancies this year is that Democratic support grew the closer we got to Election Day. This does not necessarily mean the same will happen as we approach November, but Republicans in those prior races started out better positioned than we see in many competitive races for the midterm,” said Patrick Murray, director of the independent Monmouth University Polling Institute.

Full results

These eight House districts are particularly competitive because Donald Trump’s vote share was less than Mitt Romney’s in election precincts that encompass just under half of the combined electorate. Republican House candidates are doing worse in precincts where Trump underperformed even after controlling for the partisan lean of those precincts. Furthermore, Republican House candidates are not doing as well overall in Republican precincts as Democratic candidates are doing in Democratic precincts. This performance gap currently offsets the natural GOP lean of these congressional districts.

“Enthusiasm is contagious and this equation favors the Democrats right now. For example, we saw in Virginia’s gubernatorial contest last year that living in an area with large numbers of Democratic voters correlated with higher turnout levels. There are some indications from recent primaries that Republicans may be catching up on a national level, but they still lag behind Democrats in these key competitive races,” said Murray.

The analysis of Monmouth’s congressional polling also finds that sorting voters according to race, education and gender provides a useful way to identify base voters versus persuadable voters. Republicans can count on white men without a college degree, while Democrats have strong support from white women with a college degree as well as women of color regardless of educational attainment. Importantly, voter enthusiasm currently seems to be higher among core Democratic groups than it does among the GOP blocs.

“In many respects, control of the House will depend on turning out voter blocs that each party knows it can count on for support in these districts. Monmouth’s House polling to date supports the narrative that white men who haven’t graduated college will be a key constituency for Republicans while white women with a college degree will serve the same role for Democrats. And even though these competitive districts tend to have smaller minority populations, Democrats cannot neglect engaging women of color, who are solidly behind them,” said Murray.

White women without a college degree and white men with a college degree tend to lean Republican but these groups include a large number of voters who are not strongly supportive of Trump. These are voters who Republican House candidates need to shore up while the Democratic House candidates need to coax them over to their side. Men of color prefer Democrats but this group also includes a significant number of voters who are supporting a GOP House candidate.

“These findings do not predict that Democrats are going to sweep these eight House contests. But they do indicate that the Republican candidate in each of these districts is facing an unprecedented challenge to hold that seat,” said Murray.

Monmouth’s polling in the special elections for Pennsylvania’s old CD18 in March and Ohio’s CD12 in August showed that the Republican candidates started off in the lead a month or two before those elections (+5R points in PA18 and +9R points in OH12 using Monmouth’s current standard likely voter model). Those leads evaporated by the week prior to ballots being cast (+2D points in PA18 and +1R points in Ohio using our standard likely voter model).

House vote intent by precinct type

Monmouth’s summer polling of these eight House districts shows the Republican candidates ahead in precincts where the average presidential vote in 2012 and 2016 favored the Republicans (Romney and Trump), while the Democratic candidates are ahead in those areas won by the Democrats (Barack Obama and Hillary Clinton). This is certainly expected but the most important aspect of this finding is that the GOP House candidates are underperforming in Republican precincts relative to the Democrats’ performance in their base precincts.

Data from recent elections have shown that turnout can vary by the partisan nature of where voters live, with higher turnout being seen in Democratic areas. This pattern has not been consistent in all elections since 2016, but has been of particular note in the Virginia gubernatorial race and the Alabama special U.S. Senate election in 2017 as well as the recent special U.S. House election in Ohio’s 12th Congressional District. Monmouth’s House polling allows us to group voters according to the partisan lean of the precinct where they live, based on the presidential vote in 2012 and 2016. Comparing the current state of the race to past presidential vote gives a sense of how the current Republican nominees are doing relative to that benchmark.

2018 House Vote by Precinct Type

Precincts where the 2012-16 average
presidential vote margin was…

TOTAL

Dem +25 points

Dem 10-24 points

Dem 5-9 points

Comp-etitive

Rep 5-9 points

Rep 10-24 points

Rep 25+ points

POTENTIAL VOTERS        
(Share of electorate) (9%)(13%)(7%)(14%)(9%)(25%)(23%)
         
REP candidate42%29%35%32%41%43%46%48%
DEM candidate43%57%50%55%45%39%38%35%
Dem net

+1

+28

+15

+23

+4

-4

-8

-13

         
LIKELY VOTERS        
(Share of electorate) (10%)(14%)(7%)(14%)(10%)(26%)(20%)
         
REP candidate43%32%36%32%41%45%47%53%
DEM candidate47%57%55%62%50%44%42%37%
Dem net

+4

+25

+19

+30

+9

-1

-5

-16

         
Source: Monmouth University Poll, summer 2018 combined data (CA48, PA01, PA17, NJ03, NJ11, OH12, VA10, WV03)

Specifically, the GOP candidate lead among all potential voters in Monmouth’s polling is just 13 points in precincts that Romney/Trump carried by an average of 25 points or more. It’s an 8 point lead for the House Republican in precincts where the GOP presidential margin was 10 to 24 points, and a lead of just 4 points in precincts where the presidential margin was 5 to 9 points. On the opposite end of the spectrum, the Democratic House candidate lead is much stronger in districts carried by Obama/Clinton ranging from 15 to 28 points depending on the size of the presidential vote margin. In competitive districts – those where the average margin was less than 5 points for either party’s presidential ticket – Democratic House candidates have a slim lead of 4 points at this stage of the race. Despite the Democrats’ relative dominance in their base precincts, these races remain close because just 29% of all voters live in precincts where the Democratic presidential candidates won by 5 points or more on average. The majority of voters (57%) live in precincts where the presidential vote has averaged at least 5 points in favor of the Republicans.

Another way to look at precinct level vote is by examining how Trump performed in 2016 relative to Romney in 2012 as an indicator of any anti-Trump element in this year’s House vote. In general, the Democratic House candidate does better in precincts where Trump underperformed Romney after controlling for the precinct’s partisan lean. Specifically, the Democratic House candidate lead among potential voters in Democratic precincts is 22 points in those precincts where Trump underperformed and 18 points in those precincts where Trump did the same as or better than Romney. The Democratic House candidate advantage in competitive precincts where Trump underperformed Romney is 14 points compared to a much smaller 4 point lead in competitive precincts where Trump did better. And while the Democratic House candidate trails by 5 points in Republican leaning precincts where Trump underperformed Romney, that gap is smaller than the 12 point deficit facing Democratic candidates in Republican precincts where Trump did not underperform. Again, it is important to note that precincts where Trump underperformed Romney represent just under half (46%) of the entire electorate in the eight districts polled. On the other hand, Republican precincts where Trump did the same as or better than Romney actually encompass a sizable 37% of the total electorate.

2018 House Vote by Precinct Type & Trump Performance

Precincts where… 

Trump did worse
than Romney

Trump did same as or better
than Romney

and the 2012-16 average
presidential vote margin was…

TOTAL

Dem-ocratic

Comp-etitive

Rep-ublican

Dem-ocratic

Comp-etitive

Rep-ublican

POTENTIAL VOTERS       
(Share of electorate) (13%)(13%)(20%)(10%)(8%)(37%)
        
REP candidate42%32%37%45%33%40%47%
DEM candidate43%54%51%40%51%44%35%
Dem net

+1

+22

+14

-5

+18

+4

-12

        
LIKELY VOTERS       
(Share of electorate) (15%)(14%)(24%)(9%)(7%)(32%)
        
REP candidate43%35%37%47%33%40%50%
DEM candidate47%57%56%43%53%51%38%
Dem net

+4

+22

+19

-4

+20

+11

-12

        
Source: Monmouth University Poll, summer 2018 combined data (CA48, PA01, PA17, NJ03, NJ11, OH12, VA10, WV03)

House vote intent by race/education/gender

Race, education and gender are key variables in vote choice. The large aggregated sample allows us to examine the vote choices of these groups more closely. Overall, three groups emerge as driving the base of support for each party – none of which should come as a surprise to anyone who has been following American politics since 2016. The main Republican bloc is white men without a college degree, while Democrats draw disproportionate support from two groups – white women with a college degree and women of color regardless of educational attainment. White women without a degree and white men with a degree are more evenly divided in their House vote intent. Men of color also tend to be a somewhat competitive group, although more lean Democratic.

Among all potential voters, Republican candidates in these eight House races have a lead of 24 points among white men without a college degree while the Democratic candidates have a lead of 21 points among white women with a college degree. The Democrats have an overwhelming advantage of 54 points among women of color. Men of color prefer the Democrat by 10 points. Republican House candidates have a slight advantage among women without a college degree (5 points) and men with a college degree (6 points).

2018 House Vote by Race/Education/Gender

 

TOTAL

White no college degree male

White no college degree female

White college educated male

White college educated female

Hispanic, black, Asian, other male

Hispanic, black, Asian, other female

 
POTENTIAL VOTERS        
(Share of electorate) (23%)(25%)(16%)(19%)(9%)(8%) 
         
REP candidate42%54%44%48%34%35%14% 
DEM candidate43%30%39%42%55%45%68% 
Dem net

+1

-24

-5

-6

+21

+10

+54

 
         
LIKELY VOTERS        
(Share of electorate) (21%)(23%)(18%)(21%)(9%)(8%) 
         
REP candidate43%57%48%49%33%37%15% 
DEM candidate47%32%42%44%60%50%73% 
Dem net

+4

-25

-6

-5

+27

+13

+58

 
         
POTENTIAL VOTERS
Grouped by Party Registration*
     

Democrats (39%)

 (18%)(26%)(14%)(22%)(9%)(11%) 
REP candidate15%24%18%17%9%6%8% 
DEM candidate74%59%69%78%85%77%85% 
Dem net

+59

+35

+51

+61

+76

+71

+77

 
         

Republicans (38%)

 (25%)(26%)(19%)(18%)(8%)(4%) 
REP candidate72%78%75%74%70%66%42% 
DEM candidate15%12%9%16%17%21%41% 
Dem net

-57

-66

-66

-58

-53

-45

-1

 
         

Independents/other (22%)

(28%)(20%)(16%)(14%)(11%)(11%) 
REP candidate36%51%35%41%23%40%5% 
DEM candidate37%25%36%40%54%28%55% 
Dem net

+1

-26

+1

-1

+31

-12

+50

 
         
* In VA10, party primary vote history was used in lieu of registration

 

Source: Monmouth University Poll, summer 2018 combined data (CA48, PA01, PA17, NJ03, NJ11, OH12, VA10, WV03)

Partisan registration provides a stable, historical context for understanding how much change in voting patterns has actually occurred among these groups. Overall, registered Democrats prefer their party’s House candidate by a 59 point margin (74% to 15%) while registered Republicans prefer their party’s House candidate by a similar 57 point margin (72% to 15%). Independents are evenly split (37% for the Democrat and 36% for the Republican).

The highest degree of party splitting comes from white men without a college degree who are registered as Democrats – 24% of this group support the Republican House candidate. Likewise, 18% of white women without a degree and 17% of white men with a degree who are registered as Democrats are actually supporting the Republican. Among registered Republicans, women of color (41%) are the most likely to cross party lines and vote Democratic, although they make up a very small share of GOP-affiliated registrants.

Political interest and views by race/education/gender

Overall 54% of voters express a lot of interest in the upcoming congressional elections, but there is a significant disparity by voting intent. High interest is more prevalent among voters supporting the Democratic candidates (62%) than it is among those supporting the Republicans (54%) in these eight races. The highest levels of interest come from college educated white men (75%) and women (72%) who are supporting a Democratic House candidate.

“It is probable that overall voter interest will grow as we get into the thick of the fall campaign. The question is whether Republicans will catch up with Democrats on this score,” said Murray.

High Interest in House Election by Race/Education/Gender

 

TOTAL

White no college degree male

White no college degree female

White college educated male

White college educated female

Hispanic, black, Asian, other male

Hispanic, black, Asian, other female

 
POTENTIAL VOTERS        
TOTAL54%51%45%65%60%53%46% 
         
Among those supporting the
Republican House candidate
54%54%46%63%53%53%55% 
         
Among those supporting the
Democratic House candidate
62%56%51%75%72%59%51% 
         
Source: Monmouth University Poll, summer 2018 combined data (CA48, PA01, PA17, NJ03, NJ11, OH12, VA10, WV03)

Monmouth’s summer House district polling also found that voter opinion of Trump was evenly divided at 49% approve and 48% disapprove. However, there is a negative gap when looking only at strongly held opinions – 33% strongly approve and 40% strongly disapprove. As may be expected, the president has an overall net positive rating among white men without a college degree (+39 points). He also has a smaller net positive rating among white women without a college degree (+10) and white men with a college degree (+3). However, that positive advantage shrinks dramatically when looking at strongly held opinion. Trump’s net rating when looking just at the gap between strongly approve and strongly disapprove drops to +26 among white men without a degree and +2 among white women without a degree. It flips to a net negative (-7) among white men with a degree.

Trump Job Rating by Race/Education/Gender

 

TOTAL

White no college degree male

White no college degree female

White college educated male

White college educated female

Hispanic, black, Asian, other male

Hispanic, black, Asian, other female

 
POTENTIAL VOTERS        
Approve49%67%53%51%33%42%18% 
Disapprove48%28%43%48%64%53%74% 
Net approve-disapprove

+1

+39

+10

+3

-31

-11

-56

 
         
Strongly approve33%49%36%33%21%26%7% 
Strongly disapprove40%23%34%40%56%42%63% 
Net strongly approve – strongly disapprove

-7

+26

+2

-7

-35

-16

-56

 
         
Source: Monmouth University Poll, summer 2018 combined data (CA48, PA01, PA17, NJ03, NJ11, OH12, VA10, WV03)

Trump gets an overall net negative rating from white women with a college degree (-31) and women of color (-56). He also has a net negative rating among men of color (-11). These margins are not much different when looking only at those who have a strongly held opinion of the president.

“This is another measure of enthusiasm about who is more likely to turn out to vote in November. Currently, the anti-Trump vote is more galvanized than the pro-Trump vote,” said Murray.

Looking at opinion of the candidates running for the House in these eight districts, Democrats earn a slightly better personal rating (+19, 31% favorable and 12% unfavorable) than Republican candidates (+11, 32% favorable and 21% unfavorable). However, many voters had no opinion of the major party candidates in their congressional districts when these polls were taken over the summer. Of note, Democratic candidates tend to be slightly less known in these districts because half of the districts Monmouth polled feature Republican incumbents running for re-election.

Democratic candidates tend to get net positive ratings from all demographic groups – ranging from a +7 rating among white men without a college degree to a +35 rating among white women with a college degree. Views of the GOP field are more mixed – ranging from a net negative -8 among women of color to a positive +25 among white men without a college degree. The biggest gaps between the ratings for the two parties’ candidates in these races is among women of color (swinging from a +29 rating for the Democrat to a -8 rating for the Republican) and white women with a college degree (swinging from a +35 rating for the Democrat to a -1 rating for the Republican), and to a lesser degree among white men without a degree (swinging from a +25 rating for the Republican to a +7 rating for the Democrat). Men of color tend to give better ratings to the Democratic (+16) than the Republican (+3) candidates. White women without a college degree give similar ratings to both candidates (+15 for the Democrat and +17 for the Republican) and white men with a degree do the same (+18 for the Democrat and +14 for the Republican)

2018 House Candidate Rating and Congress Control by Race/Education/Gender

 

TOTAL

White no college degree male

White no college degree female

White college educated male

White college educated female

Hispanic, black, Asian, other male

Hispanic, black, Asian, other female

 
POTENTIAL VOTERS        
Republican candidate rating       
Favorable32%40%32%40%31%23%13% 
Unfavorable21%15%15%26%32%20%21% 
Net favorable

+11

+25

+17

+14

-1

+3

-8

 
         
Democratic candidate rating       
Favorable31%24%26%35%43%27%34% 
Unfavorable12%17%11%17%8%11%5% 
Net favorable

+19

+7

+15

+18

+35

+16

+29

 
         
         
Congress control party preference*      
Republican39%53%44%44%32%34%10% 
Democrat39%22%34%41%52%43%58% 
D net

0

-31

-10

-3

+20

+9

+48

 
         
* Party control preference was not asked in WV03

 

Source: Monmouth University Poll, summer 2018 combined data (CA48, PA01, PA17, NJ03, NJ11, OH12, VA10, WV03)

“Republican candidates cannot run away from Trump or they will lose their base support. At the same time, they need to get persuadable voters to look at them separately from the president. Right now, GOP candidates in these competitive districts are coming up on the negative side of the ledger when compared with their opponents,” said Murray.

Finally, voters in these eight districts are split when asked whether they would like to see the Republican Party (39%) or the Democratic Party (39%) in control of Congress. Women of color (+48 D) and white women with a college degree (+20 D) prefer Democratic control while white men without a degree prefer Republican control (+31 R). Other groups are more divided including men of color (+9 D), white women without a degree (+10 R), and white men with a degree (+3 R).

“We are seeing some groups who tend to lean Republican on the fundamentals, such as white men with a college degree, being open to voting Democratic based on the quality of the candidates. It will be interesting to see how the campaigns try to appeal to these voters,” said Murray.

METHODOLOGY

The Monmouth University Poll was sponsored and conducted by the Monmouth University Polling Institute. The results described in this release are drawn from 3,390 live telephone interviews with voters in eight competitive U.S. House districts conducted from June through August 2018. The individual districts include California-48, Pennsylvania-1, Pennsylvania-17, New Jersey-3, New Jersey-11, Ohio-12, Virginia-10, and West Virginia-3. Information about the methodology for each of these individual polls can be found on Monmouth’s website. The pool of potential voters represents 85% of all registered voters in these districts (limiting the pool to those who have either voted in at least one election since 2010 or have registered to vote since 2016).

DEMOGRAPHICS (weighted) 
Potential VotersLikely Voters

Party Registration

Party Registration

38% Republican43% Republican
39% Democrat42% Democrat
22% Neither15% Neither
  

Self-Reported Party ID

Self-Reported Party ID

33% Republican35% Republican
36% Independent33% Independent
30% Democrat32% Democrat
  
49% Male49% Male
51% Female51% Female
  
17% 18-3412% 18-34
23% 35-4920% 35-49
33% 50-6435% 50-64
28% 65+33% 65+
  

83% White, non-Hispanic

84% White, non-Hispanic

6% Hispanic

6% Hispanic

5% Black

5% Black

5% Asain

3% Asain

2% Other

2% Other

  

57% No college degree

52% No college degree

43% 4-year college degree

48% 4-year college degree

  

Click on pdf file link below for full methodology and results by key demographic groups.