[EL] early voting drop off

[Smith, Brad]

I would not disagree with any of your points, and that's why I'd think that indications of roll-off would be only very tentative - but one has to begin somewhere, and I can't think of anything else around which to begin to construct a model.

Your first question is certainly a good one that needs exploration. We do generally know that more campaigning does increase information, though, so I would be inclined to think it does so in down ballot races - indeed, especially so, since such races typically get little unpaid media coverage.

I believe that issues two and three that you raise below are not too relevant to the type of race I am considering, which are local races, typically with small electorates. That is, I would think hard partisans are less likely to stop voting in down ballot for Senate, or House, or state legislator; but they may well be dropping off - by the time they hit school board, village council, township trustee, township cemetary board, party precinct representative, and such. For these races, almost all campaigning, such as it is, comes in the final two to four weeks before the election. So it might show up there (or it might not- the fact that early voters are more engaged may overcome the lack of campaigning.) I would see the former as interesting and perhaps pointing in the direction of my personal observation; the latter possibility would be to me of little meaning until, as you point out, you could learn more about voter characteristics, which may be impossible. (Digression: does fact that early voters are more highly engaged and decided really defeat concerns about voter information? No. It does mitigate them. I doubt, though, that the mitigation has much effect in these local races of the type that interest me.)

As for having lots of time, lots of time to go through no information is no more valuable than a little bit of time to go through no information. Again, there might be some mitigation to the lack of campaigning, in that voters might take time to try to learn what they won't three weeks later at the booth on election day - I suppose with the modern web, there is some information available. But most of these campaigns do nothing outside of a month, usually less. My working hypothesis, therefore, would be that the added info would more than offset the "time at the table" effect. (And of course, for live early voters - I don't limit this to voting by mail - I would expect that the "more time" argument pretty much flows down the drain). And point 4, I would suggest, is almost irrelevant for these types of races, in which $1000 in total expenditures is a hugely "expensive" campaign, and there is virtually no media coverage, and there is certainly nothing we would think of as "voter mobilization" by the campaign itself (although local parties may sometimes endorse and promote candidates-usually via slate cards-in these largely non-partisan races). The campaigns could respond by spending more, but I see that as relatively unlikely. And if more spending in these small, local races would be the necessary result - as point 4 suggests it would - that too raises issues about the costs and benefits of early voting. At this point, I doubt that such responsive behavior has yet taken place.

So that's the question. I agree, the best would be to actually try to study what voters know when they vote in down ballot races. But I'm quite sure there is no good info on that as it pertains to early voters vs. election day voters, and I'm not sure there is a good way to start looking at this information, to get the breakdowns needed, and least absent some pretty pricey survey research. My thinking was that if drop off increases for early voters - despite the points you raise below - that would better indicate that I may be on to something, or that if true, the dimensions are relatively large - large enough to overcome other factors). That would encourage me to look more. So I'm looking for some way to start looking at this, not necessarily hoping to find fire, but figuring if there is any smoke that it makes it a more promising thesis than no smoke. That smoke may not lead back to a fire, and on the flip side lack of smoke doesn't mean there's no fire there. But it does seem like a place to at least start.  Is there any research out there that would suggest more or less promising lines of inquiry?


Bradley A. Smith

Josiah H. Blackmore II/Shirley M. Nault

   Professor of Law

Capital University Law School

303 E. Broad St.

Columbus, OH 43215

614.236.6317

http://law.capital.edu/faculty/bios/bsmith.aspx

________________________________
From: Paul Gronke [paul.gronke at gmail.com]
Sent: Sunday, February 16, 2014 2:29 PM
To: Smith, Brad
Cc: Election Law
Subject: Re: [EL] early voting drop off

Brad

I think your logic works at first blush, but when you reason out the way that early voting has played out, I think the logic is a lot harder to sustain.

The assumption is that roll off is a function of lack of information.  That seems reasonable.

But the next assumption is that the difference between early voting roll off and election day roll off is that voters on election day have more information about down ballot races than do early voters. That assumption is a lot harder to sustain for a variety of reasons:

1) Down ballot races in general suffer from low salience and information.  Do we know that the marginal improvement in information in the last two or three weeks of an election makes a significant difference in down ballot races?

2) Early voting (by mail or in person) is a choice made by the voter, and if we assume (and we have empirical evidence of this) that early voters are more informed voters and "decided" voters, then it's actually possible that roll off will be *lower* among early voters.  What we'd really need to do is control somehow for a set of characteristics of voter--age, education, income, race, etc--then within these subpopulations, compare roll-off rates.   That's really hard to do, as you point out in your email.

3) By mail voters can vote at a leisurely pace, and officials in Oregon and Washington have long argued this as a point in favor of VBM, particularly given our long ballots with complicated referenda and initiatives.  I'm not sure this is true, but it's not obviously not true.

4) And then the political scientist / institutionalist reminds us that campaigns and other institutions are not static, but are responding to changing behavior by voters.  When campaigns begin to mobilize during the early voting period, election officials engage in outreach efforts, and media outlets change the timing of their coverage all to correspond to early voting, then logic becomes way more complicated.

---
Paul Gronke Ph:   503-517-7393
                        Fax: 503-661-0601

Professor, Reed College
Director, Early Voting Information Center
3203 SE Woodstock Blvd
Portland OR 97202

EVIC: http://earlyvoting.net






On Feb 16, 2014, at 11:11 AM, Smith, Brad <BSmith at law.capital.edu<mailto:BSmith at law.capital.edu>> wrote:

Paul,

Yup, that's what I mean, and I understand those problems, which is why I wondered if they is much data on it. I've not been aware of much. I'm also interested in all early voting, not simply voting by mail.

Thanks, this was helpful, if only to confirm what I suspected, and I was not aware of the Hanmer paper.

I am curious because my own anecdotal perception is that early voting is bad for voter knowledge on down ballot races. I think there is a logical reason to suppose this could be generally true, but I have no idea if in fact it actually is true. I don't know of anything that attempts to compare voter knowledge on down ballot races between early voters and election day voters, but looking at residual vote rates could be a place to start - lack of knowledge about down ballot races might start to show up in drop off there.


Bradley A. Smith

Josiah H. Blackmore II/Shirley M. Nault

   Professor of Law

Capital University Law School

303 E. Broad St.

Columbus, OH 43215

614.236.6317

http://law.capital.edu/faculty/bios/bsmith.aspx

________________________________
From: Paul Gronke [paul.gronke at gmail.com<mailto:paul.gronke at gmail.com>]
Sent: Sunday, February 16, 2014 1:10 PM
To: Smith, Brad
Cc: Election Law
Subject: Re: [EL] early voting drop off

Brad

Can you clarify if by "drop off" you mean the same thing as "roll off", or a voter not completing the full ballot?

Keep in mind that except for a very few studies (all, I think, involving Michael Hanmer and Michael Traugott through a relationship with Oregon election officials early in the 00's), we almost never have access to the actual ballot images, but have to infer indirectly by the residual voting rate.

And even if we did have ballot images, unless we can attach this to characteristics of the voter, we can't infer that roll-off is due to technology or to some characteristic of the voter that is correlated with technology.  For example, in the study that Stewart and I conducted in Florida (available at supportthevoter.gov<http://supportthevoter.gov/>), we show that the highest residual voting rate in Florida in 2008 was in two precincts that were wholly contained within senior citizen care facilities.

Senior citizens use no-excuse absentee voting in Florida at higher rates than other groups in the population.  But there may be other reasons that older voters roll off more frequently in Florida than younger voters.  Is it vote by mail?  Is it the age of the voter?  Or, as Charles and I suspect, is it both--age + the makeup and format of the absentee ballot were interacting in such a way to make it harder for elderly voters to complete.

Ideally, we'd want to experimentally treat voters to different voting technologies but within the same or same set of contests because roll-off is affected by so many other things besides technology.    Obviously, we can't do that.  Some have tried to estimate by looking at very similar precincts where one precinct uses VBM and the other does not (these are all California studies).  Or we could examine the residual vote rate across different modes of balloting.

Finally, the most important reason that election day voting has much lower roll off in general is that voters, when they feed a ballot into an OCR or complete a ballot on a DRE, may be prompted if they cast an undervote.  Obviously, this does not happen with a by-mail ballot.

What I can dredge up from memory and my bibliography:


Hanmer, Michael J, and Michael W Traugott. 2004. “The Impact of Voting By Mail on Voter Behavior.” American Politics Research 32: 375–405. http://apr.sagepub.com/cgi/content/abstract/32/4/375.

Look at residual vote rates (and straight ticket balloting) on down ballot races during Oregon's VBM transition.  There was no noticeable increase or decrease in either.

Alvarez, R. M., D. Beckett, and C. Stewart. 2012. “Voting Technology, Vote-by-Mail, and Residual Votes in California, 1990-2010.” Political Research Quarterly 66(3): 658–70. http://prq.sagepub.com/cgi/doi/10.1177/1065912912467085 (February 16, 2014).

Compare residual vote rates in Presidential, gubernatorial, senatorial, and proposition contests across voting technology and across twenty years in California.  Residual vote rates overall are higher with VBM, are smallest in presidential contests, and then in order: propositions, governor, then senate.

---
Paul Gronke Ph:   503-517-7393
                        Fax: 503-661-0601

Professor, Reed College
Director, Early Voting Information Center
3203 SE Woodstock Blvd
Portland OR 97202

EVIC: http://earlyvoting.net<http://earlyvoting.net/>






On Feb 16, 2014, at 8:49 AM, Smith, Brad <BSmith at law.capital.edu<mailto:BSmith at law.capital.edu>> wrote:

Does anyone have (or can you point me to studies) with info on drop off for early balloting vs. election day balloting. I'm curious if one or the other has more drop off in voting for down ballot offices.


Bradley A. Smith

Josiah H. Blackmore II/Shirley M. Nault

   Professor of Law

Capital University Law School

303 E. Broad St.

Columbus, OH 43215

614.236.6317

http://law.capital.edu/faculty/bios/bsmith.aspx

_______________________________________________
Law-election mailing list
Law-election at department-lists.uci.edu<mailto:Law-election at department-lists.uci.edu>
http://department-lists.uci.edu/mailman/listinfo/law-election


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20140216/a70a5577/attachment.html>