July 26, 2010

Heinrich Maintains Solid Lead in New Mexico’s

First Congressional District
To: Interested Parties
From: Greenberg Quinlan Rosner Research

In the race for New Mexico’s 1st Congressional District, Martin Heinrich continues to be in a
strong position for victory over his challenger, Jon Barela. In a survey conducted earlier this
month among likely voters in the November election, Heinrich leads Barela by double digits, 53
– 41 percent. Heinrich has led in this race throughout the year.

Heinrich’s support is especially broad; he leads Barela among Anglo and Hispanic voters,
among both men and women, as well as with younger and older voters.

In addition, 52 percent of all likely voters in the 1st District approve of the job Heinrich is doing as
their representative. This includes a majority of both Anglo and Hispanic voters.

Table 1: “If the election for U.S. Congress were held today and the candidates were
Democrat Martin Heinrich and Republican Jon Barela -- for whom would you vote,
Democrat Martin Heinrich or Republican Jon Barela?
(Percent responding)

Heinrich – Barela
Total 53 – 41

Anglos 49 – 45
Hispanics 66 – 30
Men 52 – 41
Women 55 – 40
Voters under 50 58 – 39
Voters over 50 51 – 41

Heinrich’s advantage in this race continues a trend we have seen throughout the year. In
February, PPP released a survey showing Heinrich with a 9-point, 45 – 36 percent advantage.
And in April, our GQR survey had Heinrich ahead 55 – 38 percent.

These findings are from a survey of 505 likely general election voters in New Mexico’s 1st
Congressional District by Greenberg Quinlan Rosner Research. The survey was conducted by
professional dialers on July 8-13, 2010 and carries a margin of error of +/- 4.4 percent at the 95
percent confidence interval.

Washington, DC Seattle London Buenos Aires

www.greenbergresearch.com
Heinrich Maintains Lead

Greenberg Quinlan Rosner determines likely voters based on vote history from the voter file
when available, self-reported vote history when vote history is not available and stated intention
to vote in 2010. Likelihood of voting for each combination of these variables is estimated based
on an analysis of data from the last like election in 2006 and only those voters with a high
probability of voting are included in the sample.

 2010 Greenberg Quinlan Rosner, All Rights Reserved. July, 2010