Via Calblog, I found this neat county-by-county map of the recall results; there’s all sorts of cool tables available here. It’d be nice if our state could put together something similar for this year’s gubernatorial race too.
Rather than talk about the California recall results directly—a topic I can add little to the existing discussion on anyway—I’ll just point to my August post on how recalls are compatible with representative democracy. And if California Democrats want to try to kickstart a new recall election against Schwarzenegger immediately—something they can probably gather enough signatures for to qualify for the ballot, but will probably have major problems attracting support from the voters (and which probably will be an unneeded distraction in a year when the party will need to focus on getting out the state’s vote for a Democratic nominee)—more power to them.
Kevin Drum also thinks an Arnold recall would be a distraction from the Democrats’ 2004 campaign.
Just speaking for myself, the more whining I hear out of Democrats about “stolen elections” the more likely I am to vote for Bush just to spite them—bearing in mind that now, frankly, I could probably easily be convinced to vote for a sane Democrat like Joe Lieberman or even Serb-warlord-coddler Wes Clark. So, yes, California—and the maturity of Democrats’ response to having an ineffectual, embarassing governor of their own party who was a willing captive of Old Left interest groups quite deservedly tossed from office—will matter for me, and many other fence-straddlers, in 2004.
The California Recall has prompted a few questions about various polling techniques. As someone who’s put in his fair share of hours doing telephone survey research, and has heard a version of the “pitch” from Harris Interactive from one of their in-house statisticians*, I thought I’d try to clear up some confusion.
The “traditional” way of doing political polling these days is a system called “random digit dialing.” Basically, to get the number of respondents they need, professional pollsters call several thousand households from a list of residential numbers prepared by companies like Survey Sampling Inc.; if you’re feeling cheap, there are other alternatives that can be used (with a much higher non-response rate). (Before RDD, we did stuff like what Zogby did in Iraq recently; that sort of quasi-random “man on the street” interviewing is common in non-industrialized countries, and essentially the same as contemporary exit polling in the United States.)
RDD worked pretty well for polling until computers arrived on the scene in the mid-80s along with the hardcore telemarketing industry. In the past two decades, response rates have dropped off sharply, requiring more calls to get a valid sample for statistical inference. Coupled with answering machines and caller ID, the effectiveness of RDD for getting a truly random sample has been undermined.
The Internet allows a few new options. Internet survey delivery allows respondents to complete surveys at their own convenience, and also permits the delivery of non-verbal stimuli (like photographs, long blocks of text to be read, and drawings), which is useful for experimental designs. The drawback is that just sticking a survey on the Internet will result in a non-random sample, the most notorious instance of which is the abomination known as the “web poll.” Since respondents to web polls self-select, we have no idea how representative they are of the public at large.
Two groups in the U.S. have tried to tackle the non-random response issue from different directions. Knowledge Networks (KN) solves the representativeness problem by only offering the surveys to a randomly-selected sample of households. Rather than recruiting a new batch of respondents for each survey (like in a traditional phone survey), KN has a rolling panel of several thousand households that participate in studies. They are provided with free WebTV service for the duration of their panel membership, and in exchange must participate in a certain number of surveys. The surveys are delivered via WebTV to the household. (This approach is basically the same as that employed by the Neilsens for television ratings.) As in a traditional phone survey, some weighting is done to adjust the sample to account for stratification and clustering effects. KN’s co-founders are Stanford University professors Norman Nie and Douglas Rivers; Stanford apparently has an arrangement for reduced-cost surveys with KN due to this relationship (at least judging from the number of Stanford professors and graduate students I see at conferences using KN-based experimental and survey data).
The other approach, employed by Harris Interactive, is to do post-hoc adjustments through a technique called “propensity weighting.” Harris has a truly Internet-based panel with a larger membership than KN’s panel (some of the difference in membership size is due to Harris also doing survey work outside the United States; however, they also use bigger samples for each survey for other reasons which I’ll get to shortly). Surveys are administered via the user’s web browser in response to invitations, and participants receive points for participating in surveys and also get entries in regular drawings for cash prizes. Instead of ensuring that participants are representative of the population at large, Harris uses propensity weighting to reweigh respondents based on their demographic and behavioral characteristics and the frequency of those characteristics in the population at large (weighting schemes for other survey techniques are generally based on the design of the sampling procedure). It is important to emphasize that Harris’ technique is not based on random samples. However, propensity weighting is designed to make the sample behave “as if” it was selected randomly.
Which technique is better? All of them have flaws, particularly if trying to reach certain subpopulations like the homeless and indigent (Harris’s technique might find the occasional homeless guy who checks his email at the library; KN and RDD would never catch him). For voting research, however, all of the techniques would probably fare better. Generally speaking participation is correlated with the variables that would be associated with having a telephone, a stable household, and Internet access. To the extent that some population groups are less likely to be online, propensity weighting should adjust for that (in the case of Harris).
Earlier this year, Political Analysis had an article that compared all three techniques, which found that generally RDD, KN, and Harris provided estimates of population parameters within the reported margin of error, with a few notable exceptions. For inferential statistics (trying to figure out the relationships among variables), which is generally what political scientists are interested in, the sampling issues are relatively unimportant, but for the descriptive statistics (trying to figure out what the population-at-large is like) pollsters and the media care about, there may be more important issues that weren’t addressed in the PA piece.
But generally both KN and Harris appear to have credible techniques that have been backed up with actual election results, so their conclusions are as likely to be correct as those of traditional surveys like the Field Poll and L.A. Times.
* Disclaimer: I got a free set of noise-cancelling headphones for answering a few dozen Harris surveys. I have received nothing from the Leland Stanford Junior University except a rejection letter when I applied to them when I was in high school and a free campus tour from a total ditz named “Mei Mei.” The joys of a photographic memory.
Daniel Drezner has the scoop on a poll of Californians conducted by Knowledge Networks on behalf of Stanford University’s Hoover Institution from August 28 through September 8 that finds Arnold Schwarzenegger in a commanding lead and the recall with 62% support, contrary to many polls that show Bustamante in the lead and the recall question in a dead heat. One possible explanation for the difference:
The Stanford/Knowledge Networks survey is the first to ask voters to choose from the same list of 135 candidates that they will see on election day. Previous polls have restricted voters’ choices to the top candidates and have allowed respondents to select “undecided” or similar options.
If this methodological difference alone* makes that large a shift in the results—and there is fairly good reason to believe that it does—then there’s good reason to believe that the existing polling is flawed, since this methodology more accurately reflects the balloting environment.
Meanwhile, SacBee columnist Daniel Weintraub thinks a Schwarzennegger-McClintock detente may be in the offing.
Robert of Boomshock has some thoughts on the meaning of the poll as well. As for Knowledge Networks’ methodology, I recommend this page which explains how their panel works; it's pretty dissimilar from Harris Interactive’s approach. KN in general has some pretty smart people on board (as, for that matter, does HI) who’ve put a lot of thought in how to make Internet-based surveys representative.
* There are other possible explanations for the difference; for example, it could be simply the result of random chance (hence why pollsters attach 95% confidence intervals to these estimates) or it could be the result of different compositions of the WebTV-based Knowledge Networks panel and the set of Californians who respond to telephone surveys. On the latter point, the only independent comparison I’ve seen between Knowledge Networks’ methodology and telephone polling (RDD) shows the two to provide roughly equivalent estimates of population parameters (Harris Interactive’s propensity-weighting system for its opt-in Internet-based system, however, seemed to outperform both RDD and Knowledge Networks).
Dan Weintraub notes that the Los Angeles Times poll of California voters—the first to show a lead for Bustamante outside the margin of error—included a special sample of 125 Latino voters. Dan hasn’t get clarification yet as to how the Latinos were counted in the overall poll, which interviewed 1,351 (self-declared?) registered voters, 801 of whom were deemed “likely” voters.
The key question is whether the 125 Latinos were all “likely” voters or just registered. In terms of registration numbers, the count seems reasonable in terms of a sample of Californians; however, if all 125 were “likely” there was an oversampling of Latinos which should have been corrected. (* For more on this, follow the Read More link.)
So the big question is whether or not the oversampling was an issue in the main poll, and if so whether it was compensated for. If it wasn’t, the Times poll is giving us a very biased estimate of the population parameter (in this case, the percentage of likely voters who are planning to vote for Bustamante or leaning that way).
Another possible source of the high Bustamante number is that the Times poll included “leaners” in addition to voters who initially declared a preference for a particular candidate. (Generally in surveys on vote choice, if you say “I don’t know” to the first question, a followup question will ask if there’s a candidate you are leaning towards.) If other polls aren’t combining the two categories, this could explain a big part of the difference. It might also be of substantive interest; if Bustamante’s support includes a disproportionate share of leaners, they would be easier for other candidates to sway than voters who are committed to Bustamante.
* If the oversampling was part of the master sampling design, the Latino voters could have been assigned a lower individual weight to correct for the oversampling (the technical term for this sort of procedure is stratification). This is generally hard to do in telephone surveys. (You can generally fudge it using quotas—which raise theoretical problems as well.)
However, if the oversampling was done post hoc, the only valid correction (at least in terms of classic probability sampling) is to only use the Latino voters who were in the original sample for inference to the population at large. An alternative approach called propensity weighting—which would weight the Latinos less due to their overrepresentation relative to the other ethnic groups—is somewhat controversial (however, Harris Interactive, the online survey research company, has demonstrated to at least some researchers’ satisfaction that Internet surveys with propensity weighting can actually do a better job of estimating population parameters than standard telephone surveys or Knowledge Networks’ competing WebTV-based survey product based on probability sampling; see this article from Political Analysis for details). There’s a subtle difference between the two weighting schemes (propensity weighting versus stratification), but in terms of sampling theory it’s a very important difference.
Interestingly enough, if you don’t care about the actual population parameters, sampling bias is far less problematic. For example, if you wanted to estimate a statistical model explaining support for candidates in the recall, sampling bias really isn’t an issue provided your explanatory variables account for the cause of the sampling bias. So if Latinos are oversampled, and you include a variable representing Latinos in the model, that particular oversampling problem is no longer an issue for making inferences about individual behavior. (However, if you don’t include such a variable, you’ll end up with non-spherical errors which will have unknown effects on your significance tests. It’s not quite omitted variable bias, but it’s still an ugly problem. Hence to be safe it is best to use the sample weights if you have them available, particularly if you don’t know how your sample was stratified.)
At first glance, the nation’s first statewide recall election in modern history seems like a fairly bad testing ground for past theories of political behavior. Yet there are a few things worth considering from the body of knowledge we already have from over 50 years of behavioral research (starting with Berelson, Lazarsfeld, and McPhee’s famous Elmira studies and the National Election Studies conducted at the University of Michigan):
The one thing political science can’t do is forecast this election; there’s simply no precedent for it. The big question remaining is whether or not the “no” strategy on the part of the Democrats persists much into September; if it does, the election isn’t effectively Bustamante vs. Schwarzenegger; it becomes Gray vs. Arnold. My belief is that the former election is probably much more winnable for the Democrats than the latter.
James Joyner at OTB* is getting rather tired of Rush Limbaugh’s anti-Ahnold schtick. As James points out, the state is mostly left-of-center these days (certainly relative to the rest of the country); at best, all the Republicans can hope for is someone who combines some semblance of fiscal conservatism with moderate social views. Someone channeling Roy Moore isn’t going to fly. Hence James concludes:
So, the question for California Republicans (aside from whether the recall was a good idea to begin with) is which of two plausible alternatives they prefer: Bustamonte or Schwarzenegger.
California isn’t Alabama. For some odd reason, a number of people in the state don’t seem to be capable of recognizing that.
In more recall news, Bill Simon has quit the race, essentially turning the contest into a three-way race between Davis Lite (Cruz Bustamante), Schwarzenegger, and right-wing darling Tom McClintock, as the left-wing gadflies like Arianna Huffington and Larry Flynt have failed to make any dent in the polls.
* For some reason, every time I abbreviate Outside the Beltway I have images of perps and skells. I guess that’s my fault for watching too much NYPD Blue (where “OTB” stands for “off-track betting,” a typical hangout for lowlife suspects on the show, typically those involved in illegal gambling as well as wagering on the ponies).
A number of people, including a healthy chunk on the right (most notably political commentator and Washington Post columnist George Will), don’t particularly care for the California recall election, considering it (variously) anti-democratic, unfair, or inconsistent with the will of the Founders. (Matthew at A Fearful Symmetry has rescinded his previous opposition in this particular instance due to Gray Davis’ general pissiness.)
The last point is fairly easily dealt with; unlike in gubernatorial elections, the president is indirectly elected via the Electoral College. The Electoral College was originally designed as sort of a half-way house between parliamentary democracy and direct election: like in a parliamentary system, the executive (in the parliamentary case, usually the prime minister) would be indirectly elected by the electorate. However, the Electoral College only does two things—electing the president and vice president—and then they go home; in a parliamentary system, the same body remains in office to approve, amend, or reject legislation proposed by the executive, and possibly—eventually—to remove the executive from office if it no longer reflects the preferences of the legislature. This removal, common to all parliamentary systems, is known as a vote of no confidence; if it succeeds, the executive must resign and be replaced, or new elections for a new legislature (and thence a new executive) are called.
The vote of no confidence is one way in which proponents of parliamentary democracy believe it leads to more stable government (the mirror image of no confidence is the power of the executive to dissolve parliament and call for new elections). So, how would we bring this benefit into a presidential system* without undermining the separation of powers? Obviously, a traditional vote of no confidence is out, as it would allow the legislature to remove the executive at will, and allowing the executive to dissolve the legislature would have similar problems.
The obvious solution is to allow the people who elected the executive and the legislature a “vote of no confidence” of their own. And, essentially, this is what the recall is: it allows the electorate to remove an executive or member of the legislature who is no longer acting consistently with their preferences. Since there is no continuous assembly of the electorate, and we don’t schedule election days on a regular basis with no expectation of some election taking place, the recall petition procedure allows the electorate to schedule a recall election if one is needed. And, since presidential systems don’t work well when there is no executive, there is a simultaneous election of a replacement executive (in parliamentary terms, it is a constructive vote of no-confidence). This system allows the electorate to work around deadlock between the legislature and executive, while at the same time not hurting the formal separation of powers between the executive and legislature (which would be a problem if we gave similar powers to either branch).
Perhaps most importantly, though, the recall provision substantially mitigates the problem of “lame duck” politicians who are subject to term limits. While the empirical evidence of “shirking” is decidedly mixed, the threat of a recall election may motivate term-limited single-minded seekers of reelection to behave more consistently with the preferences of the people who elected them, which is surely an outcome favored by proponents of the “delegate” model of representative democracy (as opposed to the Burkean “trustee” model).
To be sure, there is some fiddling at the margins that may be worthwhile. Some have suggested that the signature requirement for both setting a recall election and qualification for the ballot is too low, although at least in the former case it seems like getting a million registered voters to actually sign a petition is a rather daunting task to begin with; few, if any, organized interests in the state can claim that many members. And it might be reasonable to require some sort of run-off if the plurality winner doesn’t have a clear margin above the second-placed challenger (majority runoff is one possibility, but a threshold of 45% has also been suggested in the political science literature, and Shugart and Carey suggest the use of what they describe as the “double-complement rule” in Presidents and Assemblies), or to use an alternate balloting system like approval or Condorcet voting. But generally speaking, the recall provision is sound and there is no good reason why it should not be adopted elsewhere—it’s one of the few “progressivist” reforms that actually is good for democracy.
* There is no good term for the system of governance employed in the 50 states and a number of territories (as well as the subnational governments of other presidential democracies such as Brazil), although with minor variations (most notably, in the composition of the executive, the length of the terms, and the relative power of the three branches of government) it is consistent with the presidential system employed at the federal level in the United States and other democracies. So, for the purposes of this discussion, a “presidential system” is one in which the legislature and executive are separately elected by the people for fixed terms of office and with formal separation of powers between the executive and legislature—regardless of the actual name given to the executive.
Dan Weintraub at California Insider has an interesting post breaking down the numbers in the California recall race. If the polling holds up, the plurality winner may get at least 40% of the vote—not bad for a ballot with 134 other names on it.
In somewhat related news that doesn’t justify its own post, in the library today I picked up some light reading: Sampling of Populations: Methods and Applications by Paul S. Levy and Stanley Lemeshow. Disturbingly, it has (almost) nothing to do with my dissertation. I think that just proves I’m a total geek.
Georgy Russell (Signifying Nothing’s preferred candidate for governor of California) asks today:
I continue to be misquoted, and to have my quotes taken out of context. What’s up with that?!
Amazingly enough, there is no record of Russell having been interviewed by Maureen Dowd.
In all seriousness, when I ran for Congress three years ago, one report described my beliefs as being pro-prostitution (despite nary a mention of prostitution by me). I guess that’s what I get for having a platform plank calling for the legalization of all forms of consensual sex among adults.
Meanwhile, James (whose site I use as a substitute for blogging about things myself, since he has a comparative advantage in such matters) has a post with info on a number of the non-famous candidates for governor.
I think I’m in love. In a platonic way, of course…
* Well, technically speaking it’s not my book, since my name’s not on the cover. But I did write some of it…