There’s been quite a bit of reaction to my suggestion last night that Senator Lisa Murkowski of Alaska is viable as a write-in candidate against Republican Joe Miller and Democrat Scott McAdams (Ms. Murkowski confirmed last night that she will indeed run a write-in candidacy.) Eric Ostermeier of the University of Minnesota, for instance, notes that there is a fairly rich history of write-in candidacies in Alaska – but none of those candidates have won their races.
It’s probably best to take a two-step approach when addressing the question of Ms. Murkowski’s chances – since there are really two separate questions we need to answer:
1. What would Ms. Murkowski’s chances be if she were listed on the ballot as an independent against Mr. Miller and Mr. McAdams?
2. How might those chances be impacted by the fact that she is not listed on the ballot, but instead is a write-in candidate?
Murkowski’s Chances as a Named Independent Candidate
The default assumption should probably be that Ms. Murkowski would have a decent chance if she were named on the ballot alongside Mr. Miller and Mr. McAdams, simply because polls which have tested this scenario show a close race.
Public Policy Polling, for instance, had Ms. Murkowski trailing Mr. Miller by just four points in a poll that hypothesized she was named on the ballot as a Libertarian candidate. And a poll by an Alaskan pollster, Dave Dittman, reportedly showed Ms. Murkowski ahead of each of her opponents, with 37 percent of the vote to 32 percent for Mr. Miller and 19 percent for Mr. McAdams.
You can certainly make a case that there is more downside than upside in Ms. Murkowski’s numbers. Trying to split the uprights between Mr. Miller and Mr. McAdams will not be easy, and relatively few candidates from outside the two major parties have been elected to major offices like governor or United States Senator in recent times (although, Alaska provides for one of the exceptions: it elected Walter Hickel, then of the Alaska Independence Party, as its governor in 1990).
A critic might also point to the case of Gov. Charlie Crist in Florida. Mr. Crist — who pre-emptively left the Republican Party once it became apparent that he was going to lose the primary to Marco Rubio -– has now fallen behind Mr. Rubio in general election polling after leading him initially, and his independent Senate bid now has only has about a one-in-five chance of succeeding, according to FiveThirtyEight’s forecast.
But Ms. Murkowski has some advantages that Mr. Crist does not. Political identity is quite fluid in Alaska, and 43 percent of voters classify themselves as independent there, as compared to 29 percent in Florida. Being disconnected from the major party brands might be as much an advantage there as a handicap.
Another problem that Mr. Crist has run into, meanwhile, is that the Democratic opponent in his race, Kendrick B. Meek, is African-American, as are about 13 percent of Florida’s voters. Since Mr. Meek runs very strongly among African-Americans, these are votes that Mr. Crist would have a difficult time picking up. But in Alaska, a state in which there are three white candidates and an overwhelmingly white (and Native American) electorate, there are fewer such boundaries for Ms. Murkowski. Mr. Meek’s campaign is also considerably better financed than Mr. McAdams, who remains unknown to many voters outside of the town of Sitka, where he is the mayor.
Ms. Murkowski -– although she will surely run toward the center to differentiate herself from Mr. Miller –- might also have more access than Mr. Crist does to voters on her right. This is because, according to the Public Policy Polling survey, 32 percent of likely Republican voters have an unfavorable impression of Mr. Miller, as do 54 percent of independents. In contrast, only 11 percent of Republicans in Florida have a negative view of Mr. Rubio, according to last month’s Quinnipiac poll.
It is simply not that hard to envision a scenario like the one illustrated below, in which Ms. Murkowski got the votes of about 50 percent of independents, 25 percent of Republicans (mostly those who disliked Mr. Miller), and 35 percent of Democrats (mostly those who thought Mr. McAdams’ candidacy was not viable and were voting tactically). That would get her to about 37 or 38 percent of the vote overall, which could potentially be enough to win depending on how Mr. Miller and Mr. McAdams split the remaining votes.

The Impact of a Write-In Candidacy
Suppose you accept the argument that, were Ms. Murkowski running as a conventional independent candidate, with her name listed on the ballot, she would stand a decent chance of winning a three-way race. How much of a penalty might she suffer from the fact that she is not in fact named on the ballot, and that she must instead count on voters to write her name in?
One thing that isn’t as helpful as it seems is to cite the performance of past write-in candidacies in Alaska. For example, in 1968, Senator Ernest Gruening -– in a race with some parallels to Ms. Murkowski’s -– got 17 percent of the vote as a write-in after having lost the Democratic primary to Mike Gravel. What we don’t know, however, is how many votes Mr. Gruening would have received were he not a write-in. Would polls of the race have shown Mr. Gruening polling in a rough tie with Mr. Gravel -– as polls now show Ms. Murkowski doing with Mr. Miller? Or did Mr. Gruening’s relatively poor finish have nothing to do with the fact that he was a write-in? It’s difficult to know without combing microfiche copies of the Anchorage Daily News. But just because someone ran as a write-in and failed to win is not all that telling: if I, Nate Silver, were to move to Alaska, and run as a write-in, and I secured only 12 votes, it would not tell us much of anything about Ms. Murkowski’s viability.
This is why, in the previous article, I focused on the case of Shelly Sekula-Gibbs, who ran as a write-in in Texas’ 22nd Congressional District after Tom Delay resigned from Congress in 2006. A poll that tested Ms. Sekula-Gibbs as though she were a conventional candidate overestimated her standing, placing her at 52 or 53 percent of the vote when she in fact received 43 percent. The difference between the result Ms. Sekula-Gibbs was projected for by the poll, and the one she actually received on Election Day, could be a proxy for the magnitude of the penalty that Ms. Murkowski is liable to suffer for being a write-in.
Ms. Murkowski has some advantages that Ms. Sekula-Gibbs does not -– in particular, money and name recognition, and a state that seems prepared to take a fairly liberal interpretation of voter intent (although ballots that just list her name as “Lisa” probably will not count.) It seems unlikely that very many Alaskans will go to the polls on Nov. 2 unaware of Ms. Murkowski’s candidacy.
Another advantage that Ms. Murkowski has is that the Senate race will headline the ticket (Alaska also has a gubernatorial race on the ballot this year, but it is not looking very competitive.) Texas, by contrast, had a high-profile gubernatorial race in 2006; voters who were primarily motivated to vote because of that contest might not have paid much attention to the U.S. House race, and may not have been aware of Ms. Sekula-Gibbs’ write-in bid. Thus -– although Ms. Murkowski will almost certainly lose some votes because she is write-in –- she might not suffer as substantial a penalty as Ms. Sekula-Gibbs endured.
A Caveat or Two
So, my conclusion, is that Ms. Murkowski indeed stands a chance –- and more than a trivial chance -– in her write-in bid. The race would probably be competitive if Ms. Murkowski were running a conventional, independent candidacy. And there is not sufficient evidence to conclude that being a write-in will harm her to the point that she is not viable in a race where she otherwise would have been.
Still, being a write-in does require more effort -– quite literally -– on the part of one’s voters, and so I wonder how well Ms. Murkowski will fare if she chooses to run a negative campaign (her rhetoric last night, in a rally announcing her write-in bid, was fairly combative). Unlike in the case of Ms. Sekula-Gibbs, voters will have the choice of both a named Republican and a named Democrat on the ballot. Voters who were persuaded by Ms. Murkowski that Mr. Miller was a poor choice, but who liked Ms. Murkowski none the better for it, might find it more natural to vote for Mr. McAdams, or the Libertarian or Alaska Independence Party candidates, instead.
And regardless of the tenor of her campaign, any win by Ms. Murkowski is liable to be ugly in a procedural sense. If the result between Ms. Murkowski and her two opponents is at all close, litigation is likely to result. Although minor misspellings will count for Ms. Murkowski, how about a ballot cast for “Liza Murklusky”? Or one where the voter spells her name properly, but does not fill out the oval to indicate they have made a write-in choice? Perhaps the state’s disposition will be to throw “Lisa M.” ballots out, but if Ms. Murkowski loses by a few hundred votes, her lawyers are sure to have something to say about that.
Even election junkies like me do not enjoy protracted legal battles, as occurred following the presidential election of 2000, and the Senate race in Minnesota in 2008. But that is one possible result of Ms. Murkowski’s write-in candidacy. So is her eventual victory.
This post has been revised to reflect the following correction:
Correction: September 19, 2010
An earlier version of this post misspelled the name of Walter Hickel.