A new University of Delaware Center for Political Communication survey finds that 51 percent of likely voters in Delaware would vote for Democratic candidate Hillary Clinton and 30 percent would vote for Republican candidate Donald Trump, if the presidential election were being held today.

Libertarian candidate Gary Johnson was the choice of 7 percent, and 2 percent said they would vote for Jill Stein of the Green Party.

The representative telephone survey, conducted on September 16-28, 2016, interviewed 762 likely Delaware voters by landline and cell phone.

Other findings from the survey:

• Clinton leads among both women and men, but there is a large gender gap. Clinton leads Trump by 33 points among women (59 percent to 26 percent) versus 7 points among men (42 percent to 35 percent).

• Clinton and Trump are virtually tied among white respondents (40 percent Clinton, 39 percent Trump), but she leads him 90 percent to 1 percent among African American respondents and 68 percent to 14 percent among Hispanic respondents.

• Clinton has a 37-point lead over Trump in New Castle County (60 percent to 23 percent) and an 11-point lead in Kent County (46 percent to 35 percent), but Trump leads in Sussex County (44 percent to Clinton’s 34percent).

About the study

The National Agenda Opinion Project research was funded by the University of Delaware’s Center for Political Communication (CPC) and the William P. Frank Foundation. The study was supervised by the CPC’s research director, Paul Brewer, a professor in the departments of Communication and Political Science and International Relations.

Results are based on telephone interviews with a representative sample of 900 registered voters. Telephone interviews were conducted by landline (450) and cell phone (450, including 187 without a landline). The survey was conducted from Sept. 16-28, 2016, by Princeton Survey Research Associates International. Statistical results are weighted to correct known demographic discrepancies. The margin of sampling error for the complete set of weighted data is ± 3.8 percentage points. Results for likely voters are based on a sample size of 762, with a margin of sampling error of ± 4.1 percentage points.

Readers should be aware that in addition to sampling error, question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of public opinion polls.

Contact Paul Brewer at 302-831-7771 for more details about the survey’s methodology.