Sunday, February 21, 2010

Shooting the Messenger: See Next Posting: Results from a Student Poll of Republicans

For years I've had my students poll voters' preferences. They tend to be pretty good at it, unbiased, sometimes catching unexpected trends and sometimes angering candidates who don't like the results. In 1998, for example, my students found gubernatorial candidate Marty Chaves 10 or 11 points behind Governor Gary Johnson. He was not happy with this finding, and pointed out that a previous poll taken about 3 weeks earlier, had the race too close to call. I stood by my students and told him we had caught a new statewide trend against him which would show up in the subsequent polls. Not satisfied, he prepared a letter to send to the Las Cruces Sun News suggesting that, "with all due respect, Dr. Garcia should leave polling to the professionals." A mutual friend convinced him my previous polls had been pretty accurate; he would look silly if we were right, and in any case there was no percentage in insulting the integrity of 35 students at NMSU, much less a respected professor. So he did not mail the letter. He lost the race to Johnson by 12 points.

This year my students polled all 33 counties to see what Republicans were thinking about the governor's race. Once again candidates who didn't like the results were shooting the messenger. In a story written by Diana Alba of the Las Cruces Sun News Allen Weh suggested I was a "staunch supporter of Susana Martinez," (I am a Democrat) when he apparently did not like our findings that Martinez was ahead of him among those expressing a preference. This seemed to me to be a strange explanation of the results, which also showed Domenici to be way ahead of Martinez, and with Domenici having twice the number of Hispanics supporting Martinez. I would have thought it more likely that Martinez would criticize these results as being implausible. Someone from Janice Arnold Jones' campaign suggested the numbers in their "internal polling" differed from ours, which may be the case, but "internal" polling is notoriously biased in favor of the candidate doing the polling, and my students or I have no reason to be anything but as accurate as possible in reporting the results. I happen to have a soft spot in my heart for Arnold-Jones, for her courage and leadership last year in turning on a webcam in spite of House leadership opposition, in a committee meeting. But this soft spot did not stop me from reporting the results as accurately as I know how. I also have a healthy respect for Susana Martinez's vote-getting ability, since she was able to beat Greg Valdez, a former District Attorney, with a vote of 61%, when she last ran with opposition in a county (population 209,000) with a 2-1 majority of Democrats and in which Republicans rarely win any county-wide offices at all. But this didn't prevent me from reporting that she is much softer than you would expect among Hispanic Republicans, or that Allen Weh's name recognition is higher than hers. Just as I did when Marty Chavez sneered at my students' work, in this case I will firmly stand by the results of the poll my students took last week.

Having said this, it is also important to remain humble. There is always a chance of error. Even the most prestigious and experienced pollsters sometimes get it wrong. In almost all polls there is a 5% chance reality is actually outside the margin of error, no matter how well done. We try to instill in our students the importance of doing research as carefully and honestly as possible: it is easy to screw up.

By the time the next poll is taken the numbers will have changed as candidates get their names better known and their messages out. My bet is that the numbers will be compatible with what this poll shows about the horse race after the first lap. And I'll bet that between now and then--even if they deny it--those candidates not happy with the results will be using the information to make adjustments to their campaigns!

No comments: