The Campaign Spot

One More House Race Poll That’s Unbelievable (UPDATED)

This poll result is fascinating, but I’m taking a bit of salt with it . . .

Bennie Thompson’s congressional seat in the 2nd District has been considered “safe” so no polling has been done on the race until now. A recent poll conducted September 11-17 by a Baton Rouge, Louisiana firm, JMC Enterprises, indicates that the race with Bill Marcy is now considered a “toss-up.”

The poll of 441 people revealed that 81% consider themselves “pro-life” with 5% saying they were “pro-choice” and 14% undecided.

Thirty-two (32%) percent of those polled said they were satisfied with the job their current congressman is doing while 38% said they were not satisfied and 30% were undecided.

Thirty-five (35%) percent of the respondents said if the election were held today, they would vote for Bennie Thompson while 34% said they would vote for Bill Marcy, with 31% undecided.

They say “441 people,” which I’m interpreting as just “adults,” not registered voters or likely voters. In most polls, the more you tighten your screen for likely voters, the better your results appear for Republicans. But in this case, I can’t help but suspect that the screen would filter out a lot of undecided Mississippians who aren’t going to vote this year.

This is a D+12 district that includes most of the Mississippi River Delta and is 65 percent African-American. Thompson has represented this district since 1993. His smallest share of the vote was 54 percent in 1994; he won 69 percent in 2008.

Republican challenger Bill Marcy’s web site is here.

UPDATE: The great Sean Trende writes in, “Sample is 50-50 white/black. Turnout won’t be that way in November. On the other hand, ol’ Bennie had some close-ish races back in the day, so I’m not 100 percent dismissive of it.”

ANOTHER UPDATE: JMC Enterprises writes in and shares their memo, which adds more detail:

Registered Voters – Randomly picking voters from the voter lists and calling them. While this is the least labor intensive, this is also, in our opinion, the least useful – in the 2008 Presidential election, for example, 32% of registered voters in Mississippi did not vote. It doesn’t make sense to include these voters in any poll sample!

…Our philosophy about which population to use depends on the election, but we are generally comfortable with a “likely voter” model (as opposed to a “registered voter” model), particularly after we analyzed the 2nd Congressional district voter turnout in the 2002/2006 midterm elections and the 2007 statewide elections. For this poll, we chose a sample of 7,782 households (441 respondents) consisting both of “registered voters” and “likely voters” who were “live called.” Though this would seem to contradict our stated preference above, we used this broader universe so that we could best analyze where this race stands right now.This survey was conducted September 11-17, 2010.

We also wish to emphasize that the racial breakdown of the “registered and likely voter” poll sample was 54-46% black. “Likely voters” are 50-50% white/black. The racial breakdown of those who responded to this poll, however, was 63-37% white. Therefore, where appropriate, we have weighted the survey answers to reflect the racial breakdown of those we believe who will vote in the general election – in other words, we believe that a 50-50% racial breakdown will be applicable to the election this fall.

It sounds like there are a lot of traditional Thompson supporters who aren’t likely to vote this fall…


The Latest