The Portland Press Herald is exploring the shift in vote among Republican state representatives on LD 6, the bill that would allow municipalities to decide whether or not they want to require their superintendent to live within the school district. This bill passed with a strong supermajority of votes in April – over 75% support in both chambers – but was just vetoed by Governor “School Choice” LePage.

In theory, this shouldn’t have been a problem for the bill because a gubernatorial veto can be overridden by a 2/3 majority of legislators. This bill had far more support than that. However, here’s where politics steps in on our theories about consistency. 41 representatives changed their vote between April and May. 35 of those 41 were Republicans, who moved from supporting to opposing the bill. Since Governor LePage is the highest-profile Republican in the state, it is hard for Republican legislators to cross him without having reasonable cause to believe that negative consequences will follow.

This bill wasn’t even a strongly partisan bill. Support and opposition initially seemed to have more to do with local preferences than party – southern Mainers of both parties tended to support it, while representatives from other regions varied more. Initially there were 8 Democrats and 12 Republicans opposed to it. Two of the four Independent (“Unenrolled”) representatives supported it and other two opposed it. Nonetheless, even on something as relatively non-partisan as the decision whether to allow municipalities to determine the conditions of their superintendents’ employment, the governor’s veto drives some Republican legislative vote-changing.

Interestingly, as it turns out, the governor’s veto also drove some Democratic – and even Independent – legislative vote-changing! 5 of the 41 vote changes were among Democratic legislators, and 1 of the 2 Independent legislators changed his vote. All of those changes were moves in support of the bill. I hadn’t really thought about that kind of effect before looking, but I can imagine two causes for it. One is the governor’s unpopularity among Democratic voters. It’s possible that a legislator’s history of voting to override gubernatorial vetoes will be attractive to Democratic voters during the next election regardless of what the bill in question was actually about.  Another possibility relates to vote-whipping or lobbying by particularly effective bill sponsors. If that was the case, we should see different patterns of Democratic vote-switching depending on the issue or legislators involved.

At any rate, as we get into a period of increasing chamber votes and the Governor warms up his vetoin’ hand, it will be useful to have some ways to visually represent the impact of gubernatorial veto on legislative decision-making. I am excited that the Press Herald is thinking about ways to represent this visually, but I think they haven’t quite got it right yet – right now they’ve spread the visualization across two different pages, which makes it difficult to compare them. I can completely empathize; it’s really challenging to figure out how to convey information in a succinct, compelling visual which also contains a lot of information.

Just trying to think with them, though, about how to convey vote-changing visually. We’re going to be doing more of this, so we might as well explore our options! The relevant data source here is the set of roll call votes for this bill. Since there was only a veto-override vote in the House chamber, we really only need to look at the House roll-calls for LD 6 in April and May. A vetoed bill must must obtain 66% support in both legislative chambers to override the governor’s veto. (Incidentally, to see how your own legislator votes on any bill of interest to you, you can find roll-calls linked on the left-hand side of the bill summary pages.)

Here’s a summary of the vote totals:

April: 115 yes (74D, 39R, 2U), 22 no (8D, 12R, 2U) = 84% support among those voting, 76% of total chamber (including absent)

May: 86 yes (81D, 4R, 3U), 56 no (2D, 53R, 1U) = 60% support among those voting, 57% of total chamber (including absent)

The votes don’t add up to the same number because different numbers of representatives were absent on each voting day.

Clearly, the interesting story here is the vote-changing. Here’s one simple way to visually explore vote-changing by party. It’s easy to understand, but there’s not a lot of detailed information in it.

I converted the number of yes and no votes to percentages of total party members in order to compare trends by party. There are more Democrats than Republicans in the House, so if you don’t do this it makes it harder to compare party trends.

However, this visual misses some of the relevant information. Especially if you’re the Press Herald, you want to be sure to get legislator’s names in there somehow, so people can know how their own legislators voted and whether they changed, or didn’t change, their vote.

Here, and at the top of the page, my attempt at conveying that information. Click on the chart to be able to read the names.

Maybe a little closer? Not sure.

What ideas do you have for communicating visually about vote-switching?