Who is a Better Filter, Caucus Attendees or Primary Voters?

Every system requires trade-offs between competing values.  But we ought not to pretend that one set of filters has virtues that are simply not present. 

In a previous post, I examined the relationship between attitudes in the public and support for the Count My Vote initiative.  That post showed that support for the reform was strong across all groups except for one: those who are the most conservative and consider themselves active supporters of the Tea Party.  Even half of all strong Republicans support the reform.  Opposition is confined to a specific group within the Republican Party.

That kind of argument does not matter to some.  They base their argument not on what large segments of the public want but on ideas about how the political system functions.  That is fair.  Arguments about the proper political system do not start and end with public opinion.  Indeed, many of the arguments, such as the one promoted by Paul Mero, rest on the method by which public opinion gets filtered.

As he so eloquently states, “we have to find nonintrusive ways to filter the negative impact of irresponsible citizenship – irresponsible meaning single-issue voters, special-interest parasites, and uninformed citizens who think and vote based on selfish emotions. Utah’s caucus and convention system is that filter. And it works.”

Mero’s political system looks like Model 1.  In this case delegates selected at the neighborhood caucuses provide the filter between public opinion and the nomination.

Model 1

Model 2 is the system proposed by the Count My Vote initiative.  In this case, those who participate in the primary system provide the filter between public opinion and the nomination.

Model 2
The soundness of Mero’s argument rests on the quality of the filter.  According to him, primary voters are “single-issue voters, special interest parasites, and uninformed citizens who think and vote on selfish emotions.”

These are empirical questions.   One easily tested proposition is that those who are most likely to attend a caucus or convention are more informed than those who are less likely to attend.  In the previous post, we showed that those who were most supportive of the neighborhood caucus system were strong conservatives who identified with the Republican party.

Just after the caucuses in 2012, my colleagues and I asked participants in the Utah Voter Poll whether or not they attended their neighborhood caucus meetings.  This question allows us to examine whether self-reported caucus participants are better educated or more knowledgeable than non-participants.  (The Utah Voter Poll is a sample of voters only, so those who do not participate in Utah politics in any way are excluded.)

With respect to education, in our sample both caucus attenders and non-attending voters are highly educated – on average, they have had some college-level schooling.  To illustrate this point, below is a figure showing the average education levels of our two groups on a 6-point education scale ranging from some high school or less all the way to post-graduate education.  There is essentially no difference between the two groups.  The average education level for self-reported caucus attenders is 4.14 (which is the equivalent of having some college), and the average education for those who are less likely to attend caucuses is 3.92.  The difference between those averages is not statistically significant in our sample.

Average Education Levels of Caucus and Non-Caucus Attenders
But what about levels of knowledge about Utah or national politics? Do those who actually attend the caucuses know more about the political system?  To answer that question, I took advantage of the fact that the Utah Voter Poll is a panel of voters, many of whom have answered our survey questions at several different points in time. In March of 2013, my colleagues and I asked respondents to our Utah Voter Poll a series of simple factual questions about state and national politics.  (See my colleague Adam Brown’s discussion of the questions and the results here). We matched the survey responses from 2012 about caucus attendance with those same individuals’ 2013 responses to the knowledge questions.  In all, we have over 200 Utah voters who answered both the caucus attendance question and the knowledge questions.

As the figure below shows, caucus attenders appear to be slightly more knowledgeable than non-attenders, but the differences are small and not statistically significant.  (In other words, we cannot be sure that the difference between the two groups is greater than zero.)  The figure below compares the average number of knowledge questions self-reported caucus attenders got right with the average for those who did not attend.  The brackets show the 95% confidence interval for each average.  Individuals who attend the caucus do not have all that much more political knowledge than individuals who do not attend.

Average Knowledge Scores for Caucus and Non-Caucus Attenders
The other thing to remember is that the primary electorate includes the caucus attenders, and when we account for that fact, the already small gap shown above narrows even more.  In other words, it does not appear that one group is highly knowledgeable, while the other group is ignorant of Utah or national politics.  In fact, the far more noticeable trend is that knowledge of Utah politics is low among both groups.  Of the four questions about Utah politics we asked, both caucus attenders and non-attenders got about half right and half wrong, on average.  Both groups did much better in answering the six questions about national politics, where the average was 70% to 80% correct.

Thus, it does not appear that education or political knowledge are the defining factors that separate caucus attenders from non-attenders.  The difference between these two groups is not levels of information, but rather ideology:  research from the Center for the Study of Elections and Democracy shows that caucus attenders and delegates are much more ideological, even after controlling for knowledge, gender, income, etc.  This result is true in Utah and nationally.

The profound differences in ideology can be seen in the figure below, which shows the proportion of Republican caucus attenders and non-attenders in Utah who identify as strong conservatives.  (This analysis also focuses on the UVP respondents who answered both the caucus and the knowledge questions, but this time, we zero in on Republicans because Democrats have a different ideological mix.)  Fully 50 percent of Republican caucus attenders describe themselves as  “strongly conservative,” compared to about 30 percent of non-attenders, who tend to be more moderate.  And unlike the relatively small knowledge differences, this 20-point difference in ideology is statistically significant.  (In other words, we are very confident that the difference is not zero.)


Some advocates of the caucus system would say that this ideological difference is a good thing: perhaps the strongest conservatives should choose the Republican nominee.  But that is a very different argument than the one Paul Mero articulates above.  And having a nomination system that privileges the most intense ideologues also has a downside: it essentially means that those who choose the candidates in a caucus system tend to wear thick ideological lenses as they view and assess the candidates.  Such lenses often lead to what social scientists call “motivated reasoning,” a phenomenon where individuals’ choices are dictated primarily by their ideological dispositions.

I grant that no nomination system – no filter between public opinion and nominations – is perfect.  Every system requires trade-offs between competing values.  But we ought not to pretend that one set of filters has virtues that are simply not present.  As Utahns make their choices about the process, they should be fully cognizant of what they specifically gain and lose when they make the nomination process dependent on this particular group of individuals.

Possibly related posts:

Possibly related posts (automatically generated)

This entry was posted in Everything and tagged , , . Bookmark the permalink.

10 Responses to Who is a Better Filter, Caucus Attendees or Primary Voters?

  1. david smith says:

    So you didn’t ask the delegates their knowledge whose responsibility is to be in the know. Just those who elect them on caucus night? Plus did you ask them ideological questions concerning utah politics? Such the BLM issues, Utopia, gay marriage amendment, attorney general issues etc. If i were to ask lds people to name all 12 apostles I bet they would score lower than if I asked them basic doctrinal issues. Knowing issues is far more important than knowing who your minority leader is in the state legislature.

    • Adam Brown says:

      You ask good questions. This isn’t my post, but a few quick comments.

      (1) Delegates vs attendees. I agree that delegates would be an even more relevant comparison group, although looking at those who choose the delegates (i.e. attendees) is also relevant.

      (2) For ideological questions (general questions, not the specific ones you mention), take a look at the author’s previous post. This post is about knowledge levels, not ideology, but the previous post was about ideology (broadly construed).

      (3) You’re correct that knowing political names is generally less important than knowing a bit about policy, and your analogy to knowing the apostles vs doctrine is apt. But given that we’re looking at people within an electoral context who are choosing which specific people will hold office, it would seem that the names are also important. Especially when the names in question aren’t the minority leader, but rather the Speaker of the House, the Senate President, and the Lieutenant Governor.

      My two cents.

  2. Joseph T says:


    Help me understand. You had roughly 200 respondents self-select your survey, and you are taking a snapshot with that small number? Demographically, did they match the average Utah primary voter? Caucus attendee? What would the margin of error for this be “survey” be, 12-15%? If one of your students presented a finding with this margin of error would you not challenge them to find a much more reliable source of data? Of the 200 how many attended caucuses, how many did not?

    And in the end even this unreliable sample size demonstrates that the caucus attendee is more knowledgeable, though it is not statistically significant for many reasons. Of course methodology being a candidate.

    Please, replicate this theory with viable numbers and share those with us. I don’t want to be too cross but this is something the SL Tribune would run with, not a reliable source.

    • Adam Brown says:

      Though this isn’t my post, here are a few quick thoughts.

      (1) Respondents didn’t self-select into the survey. They were recruited in.

      (2) This margin of error calculator suggests that the margin of error is around 7 percentage points (assuming a 50-50 split on a yes/no question; as the split approaches 0-100, the margin declines).

      (3) A survey’s reliability is driven by its sampling method, not its size. The method is a random sample of Utahns who voted in recent elections. You’re correct that comparing the demographics would be a nice check on whether the random selection “worked” or not.

      (4) Because the findings are statistically insignificant (not for “many reasons,” just for one — the difference is within the margin of error), it is inappropriate to conclude that this analysis demonstrates anything.

      I agree that replication is always a good thing, and if I had funds to commission a survey on every question that interested me, respondents would get sick of answering questions long before I got sick of asking them.

  3. Lorie Fowlke says:

    As an attorney I can make a good argument for both sides but we should note that what the survey does not address is that the delegates selected by the caucus do become more knowledgeable about the candidates and issues between the time of the caucus meetings and the convention.

  4. Josh Daniels says:

    It would be interesting to see this analysis/study extended to add specific study of the delegates themselves (you may have already done something similar as I know you study delegates from time to time). I realize that the delegates are a subset of the caucus attendees; but, I would be interested to see if there is any significant difference between the delegates, the caucus attendees, and the primary voters in general.

    As I see it, the real filter here is the delegates themselves and not the caucus attendees. So I would be interested to see some study of that specific filter. The caucus attendees may not have much more political knowledge or awareness than the average primary voter as this study shows, but perhaps they vote for people (delegates) who do (or perhaps not, but it would be interesting to see that data).

    Great work, thanks for providing this sort of relevant and timely study and data on our state political institutions.

    • Adam Brown says:

      Here are surveys of the delegates (both parties) from 2012 (click here then find the links to download the PDFs). The questions deal with priorities other than knowledge/education, but they’re still interesting.

  5. Chris Karpowitz says:

    Josh and Laurie,

    As Adam points out, this is not my post, but I’m familiar with these data, so I’ll weigh in a bit here, too. You’ve asked some great questions. I wish we had more information about the political knowledge of the delegates who emerge from the neighborhood caucus process. Perhaps that will be a subject of study in the future. It’s clear we need more questions, though as Adam rightly makes clear, people get tired of answering our questions before we get tired of asking them!

    In terms of the issue of representativeness, however, what we know is that delegates are even less representative (in terms of ideology, issue positions, gender, and several other attributes) than the caucus attenders. That is, the neighborhood caucus meetings are not particularly representative of the party generally, but delegates are even less so. Delegates have stronger opinions and are more “extreme” ideologically (I don’t mean that in a pejorative sense, but rather as a description of those who are more committed to a certain ideology). Some would argue that’s a good thing: maybe the most conservative or liberal partisans should choose the nominee, but not all strong partisans are committed idealogues. In fact, successful parties generally want to be “big tents,” including a mix of interests and ideologies. If the nomination process is left to only those with a certain set of ideological commitments, then the full spectrum of opinion in the party is unlikely to be represented.

    What’s more, the delegates do not tend to see their purpose as representing the views of their home neighborhood or of the party generally. In the jargon of the political science literature, they see themselves as “trustees,” not “delegates.” In other words, they take seriously the idea that they should make their own decisions based on their interactions with the candidates, and they don’t feel their purpose is to follow the preferences of their neighbors back home (or of the party more broadly). Again, whether that’s a good thing or a bad thing is probably in the eye of the beholder, but the important thing is to be clear about the tendencies in each system. Any system has virtues and vices (a primary system comes with its own set of problems, too), and the challenge for Utahns right now is to decide which balance of virtues and vices they prefer.

    • Bill A says:

      IOW, delegates tend to see themselves like any other leader, such as congressmen – they have a duty to “lead,” perhaps within the confines of the voters’ wishes, perhaps not. Whatever the disadvantages of selecting ideologues, they at least are more likely to understand that certain desires (e.g., more government but less taxes) are simply irreconcilable with one another.

      It seems rather late to be getting to the question of the superiority of a convention vs. primary, however. There are dozens of states with primaries substantially like SB 54/CMV. How do they compare? Given the voters’ overwhelmingly Republican tilt, Utah pols don’t seem to be particularly extreme or corrupt.

      There is a reason Count My Vote was backed by many of the richest people in the state, and that reason had nothing to do with the best interests of the people.

  6. Bill A says:

    “The other thing to remember is that the primary electorate includes the caucus attenders, and when we account for that fact, the already small gap shown above narrows even more. ”


    When you extract caucus attendees from the primary voters, the knowledge gap widens. It’s the “Bill Gates walks into a coffee shop and the average patron is a billionaire” effect. The gap between delegates – often well-educated, well-regarded members of their neighborhoods – and primary voters is even larger. In terms of knowledge: delegates > caucus-goers > primary voters. And that’s before the delegates begin to meet with the candidates.

    The ideal political system CMV pretended to be selling does not exist. The real question is whether the filter will be convention delegates or whether it will be money – money provided by special interests who pour millions of dollars into campaigns like Bennett’s in exchange for favors. You can criticize convention delegates all you want, but the fact that they tossed out Bennett, who as a member of the banking committee was responsible for regulating the very industry that collapsed our economy, is a strong argument in the delegates’ favor. They were not responsible for the quality of the alternatives to Bennett. They were responsible for throwing out an incompetent incumbent, which they did.

    The real world of politics is not “may the best man win.” It is about very well-funded incumbents, competent or not, versus their practically anonymous challengers. It is about holding incumbents accountable each cycle. The system put in place by S.B. 54, thanks to Count My Vote, will result in a greater advantage to incumbents, who can now mobilize their campaign war chests and high name ID to win in primaries with no runoffs. Jason Chaffetz beat Chris Cannon quite handily in 2008 because he was the unpopular Cannon’s sole opponent. Had Cannon faced multiple challengers, unpopular as he was, he still may have won. That does not speak well for Count My Vote.

Comments are closed.