A study of the limits of polling

On January 31, 2008, in General, by Neil Stevens

The new poll is out! Rasmussen reports a tight race in California: McCain 32, Romney 28, Giuliani 14, Huckabee 11, Paul 5. Wowee, that’s a tight race with lots of excitement, isn’t it? Too bad it gives us almost no information on who’s in position to win more delegates.

Let me just say as an aside that I have nothing against Rasmussen Reports. My understanding is that it’s one of the better political polling organizations, because they do a better of ensuring to get likely voters when likely voters are appropriate. That’s why I reference them here, and every time I look up polls.

However when we look at political polling to tell us who’s in a better position to win a race, we need to ensure that the polling matches the electoral process. And this poll fails to do that at all in California.

California’s Republican primary isn’t a simple first past the post, winner take all event for all the delegates. No, we instead pledge the delegates in the manner they are awarded to us: All of the 53 Congressional districts award three delegates to the candidates who wins them, and the statewide winner receives our state’s 11 at large delegates and bonus girly-man delegate.

In some states, this structure might not matter. Maine and Nebraska even award their Presidential Electors this way (Maine since 1972 and Nebraska since 1996), however this system has yet to cause either state to split its Electoral votes between two candidates. This is because these states are relatively small and homogeneous in population.

California though, is huge and diverse. California is of course our most populous state, with more people than the twenty least populous states, and is third in land area. We sent nineteen Republicans to the House of Representatives last election, including more fine conservatives than many Republican states. We aren’t New York, if ACU ratings are to be believed, sending Republicans with ACU ratings in Democratic territory. We have three in the 70s, with Mary Bono Mack bringing up the rear at 72 (and falling, sadly). And yet Arnold Schwarzenegger is our governor.

How on Earth, then, are we supposed to guess what the above poll numbers mean, in terms of delegates and actually winning the primary? Well, we know McCain and Romney are statistically tied for the at-large delegates. That’s it. All we learn is that Mike Huckabee is not in the running to take 11 of our 170 pledged delegates. That’s not helpful.

Can we possibly guess what the numbers mean? I’m not sure. The Secretary of State gives primary results by county, but we haven’t often had a top of the ticket primary that mattered. Schwarzenegger won as Governor without winning a primary in 2005. McCain has quit by the time we voted for the Presidential nomination in 2000. Lungren was uncontested on our side for Governor in 1998. Serious candidates for Senate are hard to find against Feinstein and Boxer.

So we have one race to consider, that’s it: March 2002, Governor, Businessman Bill Simon v. LA Mayor Richard Riordan v. Secretary of State Bill Jones. Statewide Simon won 50-31-17. Let’s consider the candidates.

In theory, it was Jones’s turn. He actually held statewide office, and arguably the highest office held by a Republican since Dan Lungren gave up the Attorney General’s office to lose to Gray Davis four years earlier. This guy was a successful conservative. He wrote our three strikes law. He sued ‘vote pairing’ organizers, who sought to trade votes for Gore and Nader like commodities. He also ran on anti-corruption and good government, classic Contract with America-style values, but he just didn’t catch on statewide. He did, however, win a few counties: Fresno, Kings, Madera, Mariposa, and Tulare, all around his old Assembly district in the heart of California. Taking majorities as a favorite son vote tell us nothing, unfortunately, about who this kind of candidate can reach.

So, we move on to Richard Riordan. What’s he like? Think of Rudy Giuliani (though ironically Giuliani himself supported Bill Simon in this race, having known Simon when Simon worked for him as an Assistant US Attorney). Riordan was the ‘moderate,’ pro-abortion Catholic candidate in the race, who ran on his record making LA safer, as well as a perceived ability to beat Gray Davis (Sound familiar?). Davis’s campaign feared him, too, so they actually ran ads statewide highlighting how out of step Riordan was with California Republicans, having been as ‘moderate’ as to donate money to Dianne Feinstein. And, it worked. Riordan won one county: Los Angeles, where he was mayor, by 7 points. The only other counties he cracked 40% in were San Francisco (-4 to Simon) and Ventura (-3 to Simon). He did the best in some of the most Democratic parts of the state, but only won a plurality (not even a majority) where he was the favorite son candidate.

The attacks on Riordan worked because there was a solid conservative in the race Republicans could back instead. Bill Simon, wealthy banker, pro-life Catholic, was able to speak forcefully on issues that excited Republicans statewide in a way that Riordan couldn’t and Jones chose not to. Combine that with out of state support and money, and Simon was able to run as a frontrunner from the start. With Davis attacking Riordan, Simon was free to attack Davis.

So what do we conclude about McCain and Romney? Well, first off, Romney probably has a Mormon advantage in southeastern California, and could win San Bernardino and Riverside counties, breaking ahead of McCain with the help of determined co-religionist turnout. LA and SF are probably up for grabs now that Rudy Giuliani is out. Beyond that? We’ll vote like the rest of the party. We vote Republican issues.

So I expect John McCain to take the majority of California’s delegates come Tuesday, because he’s the mainstream frontrunner. But I could have told you that without the Rasmussen poll, so I just don’t see the point in taking a statewide poll for a by-district election.


Comments are closed.

Nima Jooyandeh facts.