October 7, 2019

2019 Virginia Legislative Election Survey

"Voters strongly favor Democrats in Assembly election; Trump trails any 2020 Democrat, Wason poll shows."

Summary of Key Findings

1. Democrats have a significant advantage over Republicans on voter enthusiasm, 62% to 49%. Further, 84% of Democrats say they will "definitely vote," compared with 74% of Republicans and 75% of Independents.
2. Democrats lead Republicans by 13 points on the generic ballot test among likely voters, 49% to 36%. These advantages remain under our most stringent model.
3. Voters prefer that Democrats control the General Assembly after the election, 53% to 37%, powered by a 17-point advantage among Independents.
4. Governor Ralph Northam's approval is at 51% and voters are fairly optimistic about the state of the Commonwealth. But voters are pessimistic about the state of the country, with 62% of voters reporting the country is on the "wrong track."
5. President Trump's approval is at 37%. In our 2020 ballot test, he trails a generic Democrat, 51% to 36%. (This data largely predates the Ukraine disclosures.)
6. Questions on gun control, health care, minimum wage, abortion and other topics show voters focused on both national and state issues, and significantly more likely to vote for candidates who support Democratic Party positions.
7. Just 6% of voters say they read the Mueller Report.

Analysis

Turnout and Enthusiasm. In the first of three surveys the Wason Center will conduct on the upcoming 2019 Virginia state legislative elections, we find an electoral environment that suggests Democrats maintain a significant Trump Era enthusiasm advantage. Since the 2016 presidential election, Virginia Democrats and left-leaning Independents have been more politically engaged than during the Obama administration. This has allowed them to translate their growing demographic advantages in Northern Virginia, Hampton Roads, and the Richmond metro area into political gains, picking up 15 seats in the House of Delegates in the 2017 cycle along with winning the governor’s mansion in a landslide, and flipping 3 Republican-held seats in the U.S. House of Representatives in the 2018 midterms. Both elections were marked by increases in voter turnout that benefitted Democrats. Although the turnout increase in 2017 was modest, rising from 43% in 2013 to 47.6% in 2017, 2018 turnout increased astronomically, increasing nearly 18 points to 59.5% over 2014 when it was 41.6%

As such, Virginia election watchers have been attempting to reconcile two equally powerful trends: the strong turnout of Democrats in the Trump Era and the long-established low turnout of Virginia’s off-off year cycle, in which only state legislative races are contested and there is no statewide contest to attract voters. Generally, the off-off year low turnout has favored Republican candidates, as it did in 2015, prior to the election of Donald Trump.

The data presented here suggest that even if turnout is still fairly low on Nov. 5, this year’s electorate will have more in common with the electorates of 2017 and 2018 in terms of its demographic and -- more importantly -- partisan composition, than with the electorate of 2015. Combined with the facts that state Senate Republicans are facing this newly energized, Democrat-friendly electorate, and that a court-ordered redistricting to eradicate racial gerrymandering yielded more Democrat-leaning House of Delegates districts, this suggests that Democrats are well positioned to pick up the seats they need to take control of both chambers of the General Assembly. The party has not controlled the state Senate since the beginning of this decade and has not controlled the House of Delegates since 2000. Picking up one of the chambers would be a significant political victory. Picking up both would give Democrats a governing trifecta and control of the state’s 2021 reapportionment and redistricting process.

Due to the idiosyncratic nature of low-turnout elections, which are decided as much by the voters who don’t show up as by those who do, predicting the outcome of the Virginia state legislative elections is no easy feat. To shed light on these elections, we are releasing this initial statewide survey, which presents data using two different “likely voter” models, and later in the cycle releasing two “horserace” surveys. Polling individual House or Senate races is inhibited by the relatively small geographic areas of state legislative districts. Instead, we will combine several competitive races into one House survey and one Senate survey to produce reliable estimates. We will not poll any individual races.

This statewide survey finds Democrats carrying significant interest and enthusiasm advantages heading into the final stages of the campaign. Overall, 55% of voters interviewed report being “very enthusiastic” about voting in this election. Democrats have a 13-point advantage, with 62% “very enthusiastic,” compared with 49% of Republicans and 49% of Independents. We also find a 10-point advantage for Democrats among voters who report they will “definitely” vote. Overall, 78% of voters said they will definitely vote, with 84% of Democrats definitely voting compared to 74% of Republicans.

This survey also finds Democrats holding a 13-point advantage on the “generic ballot” test, which asks voters if they will cast their state legislative ballots for the Republican or Democratic Party’s candidates. This gap persists under all versions of our likely voter model and is further supported by the strong preference voters express that the Democratic Party control the General Assembly after this election: 53% of voters want the Democrats to control the Assembly, compared with 37% who want Republican control, a 16-point gap.

“Given the significant interest and enthusiasm gaps measured in this survey, we expect some version of the Trump Bump to manifest in the 2019 Virginia state legislative elections,” said Dr. Rachel Bitecofer, the study’s author and the Wason Center’s elections analyst.

Issues and impact. A series of questions measuring the impact of candidate positions on voter support for state legislative candidates reveals an electorate focused on both state and national issues, and aligned more closely with positions advocated by Democratic Party candidates. This may partly explain the significant advantages Democrats hold on the generic ballot and control questions. “It’s clear that national politics are on the minds of Virginia voters this fall,” said Quentin Kidd, the Wason Center’s director. “Like or not, there’s no way for state legislative candidates to run in a vacuum -- their national party brands influence their fortunes.”

Voters report being more likely to vote for candidates advocating gun control than those who do not. Overall, 83% of likely voters support expanding background checks, 67% of those “strongly.” In addition, 67% of likely voters indicate they would be more likely to support a candidate who supports banning assault weapons -- a sizable majority. While 51% said supporting an assault weapons ban would make them much more likely to support a candidate, it’s notable that just 19% said that would make them much less likely to give their support. Several other issues were extremely popular with voters: 66% of voters would be more likely to support a candidate who supports a $15 federal minimum wage; 73% a candidate who advocates more spending on transportation infrastructure; 68% a candidate who supports Medicaid expansion; and 76% a candidate who supports ratifying the Equal Rights Amendment.

Less attractive to voters would be candidates who advocate limiting legal immigration into the country. Although 36% of likely voters said this position would make them more likely to support a candidate, 55% report that it would make them less likely, and 36% much less likely.

Also notable is that an issue getting a lot of attention in national Democratic politics, so-called “Medicare for All,” polls poorly among Virginia voters, even among voters who express robust support for other liberal policies. Just 36% of likely voters report that position would make them more likely to vote for a candidate, while 50% say it would make them less likely. Likely voters are a bit more receptive to a candidate running in support of free public college tuition, as 51% report that would make them more likely, compared with 43% who say it would make them less likely. Again, this is fairly modest support given the robust support in the survey for proposals such as gun control. Just 25% of likely voters said they would be more likely to support a candidate who wanted to ban abortions except when the mother’s health is in danger, a law similar to those passed recently in Alabama, Ohio, and Georgia.

Trump and Impeachment. This survey also asks voters if a state legislative candidate’s support for President Trump or support for impeachment of President Trump makes them more or less likely to vote for that candidate. Overall, 59% of voters said they were less likely to vote for a state legislative candidate who supports President Trump, while 37% said they were more likely to vote for such a candidate. On impeachment, 49% of voters said they were more likely to vote for a candidate who supports impeachment, while 44% said they were less likely. It should be noted, however, that most of this survey was fielded before the emergence of the Ukraine issue, and the subsequent announcement of an impeachment inquiry by Speaker of the House Nancy Pelosi. The result was the collection of 488 surveys of registered voters before the release of the controversial phone transcript and 238 surveys after, plus 373 “likely voter” surveys before the release of the transcript and 194 after the release. Although we see a significant change in preferences following the release of the transcript on both the “supports Trump” and “supports impeaching Trump” questions, we will not be reporting these changes because we cannot be sure they are not a product of statistical noise. However, both questions are included in our next two surveys. We take pains to highlight this in this report because the data we do present for these questions are predominately from the period before the Ukraine disclosures and may not provide an up-to-date reflection of public sentiment.

In our first test of Virginia’s 2020 presidential ballot, President Trump is trailing a “generic” Democratic Party opponent by 15 points. Again, most of this data precedes the Ukraine disclosures. Overall, 51% of voters said they would vote for the unnamed Democrat and 36% said they would vote for President Trump. Another 6% indicated they would vote for a third party candidate and 6% said they didn’t know who they would vote for. Even more problematic for Republicans is that only 28% of voters say they will “definitely” vote for Trump, while 42% say they will “definitely” vote for the Democrat. These data, along with statewide election results in recent years, suggest that Virginia will not retain its “swing state” status in the 2020 cycle.

Mueller Report. The survey asks Virginia voters to indicate the extent of their exposure to the Mueller Report on Russian interference in the 2016 presidential election. Just 6% of voters claim to have read the entire report, while 27% say they have read excerpts of the report. Another 43% say they have digested media coverage about the reports and 21% say they have not read the report or heard much about it. The direct exposure numbers do not change much when the sample is limited to those voters who say they are highly likely to vote in the Nov. 5 state legislative elections. However, under our most stringent likely voter model, the percentage who say they have read the report doubles to nearly 13%. We want to stress a central point here. This is a survey not only of voters, but of the most reliable, civically engaged voter pool of the electorate: those who participate in the state legislative off-off year election cycle. This pool of voters is not representative of all Virginia adults, or even all Virginia voters. It is more educated and -- perhaps more importantly -- much more civically inclined than the overall pool of voters and even more again than the overall adult population of the Commonwealth. Taking 6% as a hard number here suggests the overall percent of the population that has read the Mueller report is likely much lower.

How the survey was conducted:

The results of this poll are based on 726 interviews of active, registered voters in Virginia including 238 on landline and 489 on cell phone, conducted September 4-September 30, 2019. Percentages may not equal 100 due to rounding. The margin of error for the whole survey (Full Sample Model, n=726) is +/- 4.1 % at the 95% level of confidence. This means that if 50% of respondents indicate a topline view on an issue, we can be 95% confident that the population’s view on that issue is somewhere between 45.9% and 54.1%. The margin of error for the more restrictive model (Likely Voter Model #1, n=566) is +/- 4.6 at the 95% level of confidence. The margin of error for the more restrictive model (Likely Voter Model #2, n=300) is +/- 6.4 at the 95% level of confidence. Please note: it is not generally the policy of the Wason Center to release data at n=300, however, Likely Voter model #2 is an attempt to simulate the expected low turnout of registered voters in this cycle. Margins of error are higher for subgroups within cross-tabs. A reasonable estimation of crosstab margin of error is to double the topline margin of error. All error margins have been adjusted to account for the survey’s design effect, which is 1.26 in this survey. The design effect is a factor representing the survey’s deviation from a simple random sample, and takes into account decreases in precision due to sample design and weighting procedures. Sub samples have a higher margin of error. In addition to sampling error, the other potential sources of error include non-response, question wording, and interviewer error. The response rate (AAPOR RRI Standard Definition) for the survey was 13%. Five callbacks were employed in the fielding process. Live calling was conducted by trained interviewers at the Wason Center for Public Policy Survey Research Lab at Christopher Newport University. The data reported here are weighted using an iterative weighting process on region, sex, age, and race to reflect as closely as possible the population of likely voters in the Commonwealth’s state legislative cycle, an election cycle that typically produces an average voter turnout rate of 30%.