Off the Kuff Rotating Header Image

ballot order

Ballot order

Kevin Drum finds this paper, entitled “The Ballot Order Effect is Huge: Evidence from Texas”, by a professor at Sam Houston State, and notes that it confirms what we have all long believed, that being first on the ballot in a non-partisan race like a primary or a municipal election is an advantage. From the paper:

Across all twenty-four contests, the effect is invariably positive and, with two exceptions in runoff elections, statistically significant. The smallest effects are found in high-profile, high information races: the Republican primary for U.S. Senator, which featured the incumbent, John Cornyn; the governor’s race, which featured long-time Attorney General Greg Abbott; and Land Commissioner, which featured well-known political newcomer George P. Bush. In these races the ballot order effect is only one or two percentage points.

Larger estimates obtain for most “medium-profile, medium-information” races such as Comptroller, Railroad Commissioner, or the Democratic nominee for U.S. Senator. Most of these fall in a fairly tight band that ranges from three to five percentage points. Estimates are even larger in the low-profile, low-information judicial elections, generally ranging from seven to ten percentage points. Overall, the ballot order effect tends to be larger in contests that receive less attention and in which voters are likely to know less about the candidates on the ballot.

[…]

In an ironic twist of fate, we were recently able to [test our hypothesis] with the March, 2016 Texas Republican primary, held just after the first draft of this paper was completed. Featuring a highly visible Presidential race, it drew twice as many voters as in 2014—and had contests for three Supreme Court positions, one of which was between Paul Green and Rick Green, two men with common first names and identical last names. It was The Perfect Storm, and our logic implies that this should lead to large ballot order effects. This is immediately evident in the histogram of county vote shares presented in Figure 2(a), without even looking at ballot order: in a race won with 52.1% of the statewide vote, virtually no county’s vote was nearly evenly split. Instead Paul Green’s vote shares are bifurcated into two clusters, one around 40%, and another around 60%, suggesting a ballot order effect approaching twenty percentage points. The regression results in Figure 2(c) confirm this: the coefficient estimate is 19.4 percentage points. We have never seen a ballot order effect this large, and may never again.

Drum concludes that randomizing ballot order for each voter, which is something that is certainly feasible with electronic voting machines, is the best answer to this. I’ve been on that hobby horse for a long time, so it’s nice to have some empirical evidence in my corner, but in the absence of a new law from the Lege, nothing will change. But we persist in highlighting the problem, in the hope that some day our cries will be heard.

I should note that while the first-on-the-ballot effect is largest in low-information races like judicial primaries and executive offices like Railroad Commissioner, some races defy that effect. I will always cite the three-way Democratic primary for RRC in 2008, between gentlemen with basic, simple names, as Exhibit A for counterexamples. Mark Thompson, who nearly won the race on the first go, basically carried every county regardless of where he was on the ballot. Here’s Harris County:


Dale Henry       85,153  32.00%
Art Hall         69,377  26.07%
Mark Thompson   111,598  41.93%

Travis County:


Art Hall         37,444  30.87%
Mark Thompson    57,909  47.74%
Dale Henry       25,959  21.40%

Dallas County:


Art Hall         45,670  24.84%
Dale Henry       57,234  31.13%
Mark Thompson    80,980  44.04%

Three different orders, Mark Thompson was second or third on all three, and yet he easily led in all three counties, despite being a first time candidate with no money. Henry had been the Democratic nominee for Railroad Commissioner in 2006, and Hall had been a City Council member in San Antonio (Hall did carry Bexar County, though Thompson came in second), yet Thompson overcame it all and ran away with the nomination. Till the day I die, I will never understand that result.

Ballot order drawn

vote-button

Here is the official ballot order for City of Houston candidates this November, via Chron reporter Mike Morris on Twitter. You’re all familiar with my rant about ballot order by now – we have electronic voting machines, they should simply randomize the ballot order for each voter – so I’ll just skip it and move on. Whether anyone’s ballot position ultimately makes a difference or not – I sure hope it doesn’t, but I wouldn’t bet on it – we’ll have to wait and see. All I know is that in any field with more than four candidates, I’d rather be first or last than anywhere in between.

This would be a short entry if this were all I had to say, so in the interest of filling out a proper length, here are two announcements about candidate forums. On Monday, Mental Health America of Greater Houston is hosting a Mayoral forum on behavioral health, a topic I’m willing to bet you haven’t heard much about in this election. The Houston Police Department has one of the only Mental Health Divisions in the entire country, so this is an issue that needs some public discussion. MHA of Greater Houston, NAMI of Greater Houston, the Council on Recovery, and the Houston Recovery Initiative are partnered for this event. That’s this Monday, August 31, at 6:30 PM at the University of St. Thomas, Jones Hall, 3910 Yoakum – see here for details.

Want a forum for candidates other than Mayoral candidates? On Thursday, September 3, you can attend a forum on environmental issues for At Large Council candidates, brought to you by the Citizens’ Environmental Coalition, League of Women Voters of Houston, and over 20 cosponsors representing environmental organizations in the Houston region, including Hermann Park Conservancy. The event is at 6 PM at the Cherie Flores Pavilion in Hermann Park, and it will be moderated by yours truly. It’s free and open to the public – see here for details. Don’t leave me hanging, come on out and hear what the candidates have to say.

Meet your Constitutional amendments

A pretty uninspiring bunch, if you ask me.

vote-button

Now that the dust has settled on the 84th Texas Legislature, voters are getting the first official look at which constitutional amendments they will be voting on come November.

Texas Secretary of State Carlos Cascos on Wednesday took the last step to place seven propositions on this fall’s general election ballot, all of which were approved by two-thirds of all state lawmakers during the just-ended session. Per state law, they are chosen randomly in a drawing to assign the order in which each proposition will appear on the Nov. 3 ballot.

All told, they run the gamut of state issues, from the serious to the mundane, and they create a narrative of the session that is not at all inconsistent with what really happened under the Pink Dome.

Here are the amendments, in the order they will appear on your ballot.

Proposition 1 (SJR 1)

“The constitutional amendment increasing the amount of the residence homestead exemption from ad valorem taxation for public school purposes from $15,000 to $25,000, providing for a reduction of the limitation on the total amount of ad valorem taxes that may be imposed for those purposes on the homestead of an elderly or disabled person to reflect the increased exemption amount, authorizing the legislature to prohibit a political subdivision that has adopted an optional residence homestead exemption from ad valorem taxation from reducing the amount of or repealing the exemption, and prohibiting the enactment of a law that imposes a transfer tax on a transaction that conveys fee simple title to real property.”

Proposition 2 (HJR 75)

“The constitutional amendment authorizing the legislature to provide for an exemption from ad valorem taxation of all or part of the market value of the residence homestead of the surviving spouse of a 100 percent or totally disabled veteran who died before the law authorizing a residence homestead exemption for such a veteran took effect.”

Proposition 3 (SJR 52)

“The constitutional amendment repealing the requirement that state officers elected by voters statewide reside in the state capital.”

Proposition 4 (HJR 73)

“The constitutional amendment authorizing the legislature to permit professional sports team charitable foundations to conduct charitable raffles.”

Proposition 5 (SJR 17)

“The constitutional amendment to authorize counties with a population of 7,500 or less to perform private road construction and maintenance.”

Proposition 6 (SJR 22)

“The constitutional amendment recognizing the right of the people to hunt, fish, and harvest wildlife subject to laws that promote wildlife conservation.”

Proposition 7 (SJR 5)

“The constitutional amendment dedicating certain sales and use tax revenue and motor vehicle sales, use, and rental tax revenue to the state highway fund to provide funding for nontolled roads and the reduction of certain transportation-related debt.”

I will be voting No on #s 3 and 7 and probably on 1, Yes on #2, and I have no idea yet on the others. What about you?

How many candidates are too many?

The Rivard Report brings up a point I hadn’t considered before.

Candidates or their representatives arrived at City Council chambers Monday morning to draw lots to determine the order of name placement on the ballot. As candidates waited in the audience, the room seemed to be filled with equal parts anticipation and dread. It doesn’t matter much if you are first, second or even third in a three-person race. Three our four names fit easily enough on a single screen of a voting machine.

But there are 14 people running for mayor, and in an informal street poll I conducted downtown Monday, I was unable to find a single person who could name six candidates. Quite a few people named three, several named four, a few named five and none could name six. Four of the candidates are running visible campaigns with yard signs, frequent public appearances, organized block walking events and participating in public forums.

But what about voters who won’t recognize the names of Ivy R. Taylor, Mike Villarreal, Leticia Van de Putte or Tommy Adkisson? The four frontrunners are seasoned officeholders who have run multiple campaigns and appeared on multiple ballots. But they face 10 other candidates, some of whom have filed for office before but none of whom have much name recognition or a record of holding elective office. I’m talking about Paul Martinez, Douglas Emmett, Michael “Commander” Idrogo, Raymond Zavala, Rhett Rosenquest Smith, Julie Iris “MamaBexar” Oldham, Cynthia Cavazos, Gerard Ponce, Pogo Mochello Reese, and Cynthia Brehm.

The voting machines are going to have as hard a time as the voters with the mayor’s race. There is simply no way to list all 14 names on a single computer screen, and I wonder if even two screens will prove sufficient. It’s even more of a challenge when two of the candidates feature “Commander” and “MamaBexar,” nicknames that have to be listed.

If you are a candidate listed on the second screen, you have to wonder: How many people will think the contest is only between the candidates listed on the first screen and cast their vote before they get to the next screen? The computer allows a voter to reverse a decision and also prompts a voter to review his or her choices before pressing “VOTE,” but that’s small comfort to a second page candidate.

Here’s the Bexar County Elections webapge on their voting system. The video didn’t load for me, and the ES&S Flash Demonstration links are broken, but the picture at the bottom gives some idea of what they use. Here in Houston, we’ve not had a 14-candidate race in recent years that I can recall – there were 19 candidates in the January 1995 special election for Council At Large #4 – but we did have ten for At Large #2 in 2011 and twelve for District D in 2013. I’m pretty sure that Harris County’s eSlate machines were able to list everyone on a single page. At least, I don’t recall hearing anything about the candidate list spanning multiple pages. If San Antonio is like Houston, then Mayor will be the first race on the ballot. If the voting machines in Bexar County really can’t fit 14 names onto one page, then that seems like a serious flaw with them. Is this a real concern? I’m having a hard time wrapping my mind around it.

This is also an opportunity for me to bring up one of my favorite hobbyhorses, which is that the draw for ballot position is ridiculous. I still can’t understand why an electronic voting machine system can’t be programmed to randomize ballot order for each race with multiple candidates and each voter. I’m sure it would take a change to state law to allow that – or better yet, require it – and I know that there would still need to be a draw for candidate order on mail ballots, but still. This seems like such a simple fix to a problem that vexes people in every single non-partisan election. Can we please do something about it?

Ballot order and the HCC lineup

The ballot order has been determined for the November city elections. You can click over and see them, I’ll just use this opportunity to once again say how ridiculous it is that in the year 2013 we are still drawing names out of a hat for ballot order. There’s no technical reason why our electronic voting machines cannot be made to randomize ballot order in non-partisan and primary races for each voter. Whatever advantage there may be to appearing first (or last) on the ballot, we should not let that have any effect on our elections. A technical fix would be simple, but first we’d need the Legislature to mandate it. Maybe if they spent a little less time chasing the vote fraud phantom and spent a little more time thinking about how to make elections better we could have this.

Meanwhile, HCC has announced its lineup for the fall election:

The following candidates filed an Application for a Place on the November 5, 2013 General Election Ballot:

 

District

Candidates

(Listed alphabetical last name)

Term Expiration

I

Zeph Capo

Yolanda Navarro Flores

Kevin J. Hoffman

December 31, 2013

(Expired Term)

II

Bruce A. Austin

Dave Wilson

December 31, 2013

(Expired Term)

III

Dane D. Cook

Herlinda Garcia

Adriana Tamez

December 31, 2015

(Unexpired Term)

V

Roy A. Cormier

Robert Glaser

Phil Kunetka

December 31, 2017

(Unexpired Term)

VII

Neeta Sane

Ann Williams

December 31, 2013

(Expired Term)

The big news here is that District V Trustee Leila Feldman, who had been appointed to replace Richard Schechter when he resigned, is not running for a full term. I had said that she was, based on not having heard otherwise. Of the three who are running in V, Glaser is a previous City Council candidate, and I know nothing about the other two. Neeta Sane and Bruce Austin both picked up opponents on deadline day; I presume that’s the same tiresome Dave Wilson that has inflicted himself in recent city of Houston and Democratic primary elections, but I don’t know for sure. Anyway, I’ve updated the 2013 Election page as best I can with what I can find. As always, if I’m missing something that you know about, please leave a comment or drop me a note. Thanks.

Two suggestions for better elections

Ed Kilgore writes that Tennessee Democrats shot themselves in the foot with the winner of their low-information, low-profile Senate primary.

[F]acing incumbent Sen. Bob Corker will be some obscure dude named Mark Clayton, who won a plurality of the vote in a large field of unknowns via the inestimable advantage of appearing at the top of the ballot thanks to his alphabetically superior surname. Turns out Clayton is an enthusiast for homophobia along with various classic conservative extremist memes, including the “NAFTA Super-Highway” and “FEMA Concentration Camps For Patriots.” The Tennessee Democratic Party quickly disowned Clayton, but the damage to the state ticket is already done.

I’d say the situation provides some empirical evidence relevant to two issues of how states conduct elections. Tennessee is one of the relatively few southern states without a threshold requirement of the percentage of votes needed to secure a party nomination. Requiring runoffs can have pernicious effects, but on the other hand, it’s a good way to avoid deeply embarrassing accidental nominations, as Texas Democrats showed earlier this week by nominating former state senator Paul Sadler for the Senate instead of perennial candidate (sometimes as a Republican) Grady Yarbrough, whose first place finish in the primary seems to have been primarily a matter of voters confusing him with the late liberal Sen. Ralph Yarborough.

As for the alphabetical ballot listing issue, it’s long past time for every state to list non-incumbent candidates randomly. Otherwise Tennessee primary ballots may regularly feature crazy-person candidates with names like Aaron Aardvark, and Democrats may fondly remember the days when they worried Bob Corker’s last opponent, Harold Ford, Jr., was not sufficiently progressive.

We do have runoffs here, and as was the case in 2006 with Barbara Radnofsky versus Gene Kelly, I think that helped the voters figure out who the candidates were and why one was such an obviously better choice than the other. In our case at least, Paul Sadler would have been the nominee anyway under Tennessee’s system, as he finished first among the four candidates in May. One possible reason why Sadler did so much better in the runoff may be the email sent by the TDP highlighting the differences between his record and that of his runoff opponent’s. Some people disliked that action as you can see from the comments on that post, but I’m not one of them. I do think it’s appropriate and increasingly necessary for the party to play at least an informational role in primary elections. I don’t want to see smoke-filled-room shenanigans of the old days, but I do want to see the party – state or county as appropriate – put together and disseminate basic factual data about candidates in its primaries for the benefit of the voters. Stuff like primary voting history, political contributions, previous candidacies, that sort of thing. Sure, that would normally be the province of a candidate’s campaign, but I see this as being as much in the interest of the party whose banner is going to be carried by the winners of these races. This isn’t foolproof, as there are people who change their minds about which party best represents them and we want to be welcoming to them, but on the whole I think it will do a lot of good. If nothing else, if it forces parties to do a better job of maintaining contact information for its members, it will be a win.

As for randomizing ballot order, that’s a longtime hobbyhorse of mine. I do not understand why in this age of electronic voting machines that isn’t standard practice by now. As far as I’m concerned, every election in Texas that requires a majority of the vote to win – primaries, special elections, local and city elections, pretty much everything except the even-year November generals – should always have randomized ballot ordering. No one should be at an advantage or a disadvantage because of the luck of a ballot draw. I can’t even begin to think of an argument against this.

Another overview of the candidates

The Chron has another overview of the lineup for City Council and Mayor. This time, as far as I can tell, the slate they provide matches exactly the candidates listed on the City Secretary’s page. It also has a bit more information about some of the candidates, noting who has held or run for office before, which in turn leads to one of the weirdest quotes I’ve seen in a story like this:

Michael “Griff” Griffin boasts the longest record of futility in what is now his 10th run at City Council. He also seeks the At-Large 2 seat, telling people, “This is the last chance you get to vote for Griff.”

So does that mean he won’t make an 11th attempt to run for office in 2013? Why stop now after all this time?

In related news, the ballot has been determined by the random drawings of a couple of local media folks. I’ve said this before and I’ll say it again: Given that ballot order can be a determining factor in low profile races, it is utterly ludicrous that we still allow a random drawing to have that kind of effect in our elections. Nowadays we vote on computer terminals that are (or at least should be) perfectly capable of randomizing the order for each individual voter. We should insist on that being the norm in our elections and get rid of this cheesy, outdated ritual once and for all.

Anyway. I will be running the last of the interviews with Council candidates over the next three weeks – At Large #2 this week, challengers in At Large #1 and #4 next week, and At Large #5 the week after that. In October I’ll have the Mayor, the City Controller, and candidates from the HISD and HCC Trustee races. Then I get to catch my breath before the 2012 primary season begins. Enjoy the ride and get ready to vote.

Ballot position and the Republican races

Inspired by a comment JJMB left on the previous post about the effect of ballot position on the judicial races, I went and looked at the Republican results to see what I could see. Here’s what I found.

– Though there were the same number of races on each ballot, the Republicans had far fewer contested judicial primaries than the Democrats had, a total of sixteen if you include the Supreme Court and Appeals Courts, all of which were unopposed on the Democratic side. By comparison, there were 30 contested Democratic judicial primaries. The reason for fewer multi-candidate GOP races is simple: the vast majority of them featured incumbents, who generally went unchallenged. On the one hand, having fewer contested races probably saves on brainpower, since you had fewer things to pay attention to and have fewer decisions to make. On the other hand, you still have to slog through all of those uncontested races, and having those fewer decisions to make may well be more boring and glazed-eye-inducing. I have no way to evaluate that, so I’ll leave it to you to decide if that makes voters more or less likely to find a shortcut.

– Skipping the two Supreme Court races, there were nine contested primaries for Appeals Court or District Court. The candidate listed first won twice, lost six times, and made it to a runoff once. I should note that of the six candidates who were listed first but lost, three of them were taking on incumbents – Evelyn Keyes, Mike Massengale, and Lynn Bradshaw Hull. I would not consider those races to be relevant to this conversation, as incumbents will have a higher level of name recognition, and presumably access to the needed resources to run a real campaign. I don’t get Republican mail so I can’t verify that, but I sure did see plenty of Mike Massengale signs in people’s yards. On the other hand, one of the top-spot candidates who won was a District Court incumbent (Sharon McCally) who took on a sitting Appeals Court incumbent (Leslie Brock Yates) in what turned out to be a rather nasty race. Again, I don’t think this is a relevant example.

– There were five contested races for County or Probate Courts. Four of the five candidates listed first lost; the exception was Don Smyth. One of those losing candidates, Charles Coussons, had withdrawn from the race between the filing deadline and the election, but too late to have his name removed.

– Finally, we come to three non-judicial races, for District Clerk, County Clerk, and Tax Assessor. Paul Dwight, Stan Stanart, and Don Sumners, respectively, were listed first for each, with Dwight losing and the other two winning. Given the track record of candidates listed first up to this point, I can’t claim that it helped Stanart or Sumners. At best, I’d say any effect was inconclusive, since their results were atypical. I’d probably make a stronger case for Stanart being helped than Sumners since I think the Tax Assessor race was higher profile, and featured an incumbent to boot. But again, I’m generally not on the business end of Republican campaigning, so I can’t accurately assess that. All I can tell you is what I’ve laid out here, and you can make of it what you will.

– Finally, reader Gwen sent me a spreadsheet, which I have made available as a Google doc, which contains a basic model that predicts the winner in the Democratic judicial primaries based on ballot position and the Chron endorsement. You are welcome to play around with it and see if you can build on what it does.

The effect of ballot order on the judicial races

I’m going to have something tomorrow about how the different endorsing groups did in the various Democratic judicial primaries, but before we get too deep into the weeds, let’s pause to consider the effect of ballot position. Two candidates filed a lawsuit after the HCDP mistakenly listed them second, so it is certainly perceived to have an effect on races. Did being listed first on the ballot in a judicial race correlate with winning?

The results can be seen here, and the short answer is Yes and No, and the determining factor appears to be what kind of bench you were pursuing. In the District Court races – Criminal, Civil, Family, and Juvenile – it was actually a distinct disadvantage to go first. Of the 18 such races, by my count the first candidate won four times, lost twelve, and made it to a runoff twice. Frankly, that result shocked me at first, but I think what it points out is that in these races, other factors were more important. For one thing, the District Court races appeared fairly early in the ballot, presumably before most people’s attention spans started to wane. For another, this was where the money was. I don’t know about you, but I got a blizzard of mail from District Court candidates and from endorsing groups on their behalf. I even saw a TV ad towards the end for one candidate, Shawn Thierry. As such, I think these races were sufficiently high profile that people were mostly choosing names with which they had some familiarity, and because of that the ballot position factor was minimized.

The reverse was true in the County Court and Probate Court races. Of those twelve contests, the winner was the first candidate nine times. Further, of those nine winners, six exceeded 60% of the vote, with three topping 70%; two of the remaining three scored over 59%, while the last one was winning a three-candidate race. None of the three winning candidates who were listed second got as much as 52% of the vote. Among them was Dennis Slate, who was one of the plaintiffs against HCDP. These races were farther down on the ballot – after all of the District Court races, in particular – and had much less money in them. I think I may have gotten a mailer or two from these candidates, but it certainly wasn’t much; being listed in an endorsing organization’s slate was probably the most exposure a lot of them got. If ballot order can have an effect, these are the races where you would expect it to happen, and based on these results, it’s hard to argue with the idea. I’ll say it again: We really need the next generation of eSlate machines to be able to randomize ballot order. It’s just wrong that such a silly thing could affect the outcome of an election, but it sure seems like it did.

Speaking of such things, the race immediately after all of these judicial contests was the County Clerk race. Sure enough, Ann Harris Bennett was listed first, and sure enough, she cruised to an easy win. I’ve heard it suggested that Bennett, who is African-American, won on the strength of turnout in the African-American State Rep districts. Having analyzed the draft canvass data, I can tell you that she did in fact do very well in those districts, but that wasn’t determinative. If you simply remove HDs 131, 139, 141, 142, 146, and 147, Bennett still defeated Sue Schechter by about 5,000 votes, and that’s with Schechter piling up a 2,800-vote margin in her old stomping grounds of HD134. If people thought Schechter’s name was enough for her to win, they were wrong. Besides HD134, the only districts she won were HDs 136, 138, and 148 – basically, all Inner Loop areas, where she was best known. Let this be a lesson for us all.

Ballot order

Here’s the ballot order for the November city elections. I don’t know how much being listed first really means – whatever it means, it likely varies a great deal from election to election, depending on the turnout and the level of interest in the races – but all things considered I’d rather be listed first than in the middle of the pack.

Be that as it may, I have to ask: Why don’t we take advantage of the fact that we use fancy computers for our voting interface, and randomize ballot order for each race and each voter? I mean, whatever the effect of being first is, why should there be a benefit from such a meaningless thing? In partisan elections, I’d just randomize which party appears first, with each race for a given voter having the same party appear first. But for primaries and non-partisan elections, why not take the dumb luck of the ballot position lottery out of the picture? We have the technology. Why not use it?

UPDATE: Muse has more.