Posted by: sweeneyblog | December 15, 2012

Examining Social Issues in Whatcom County

Update 12/18 – I’ve pulled apart this data so that you can examine these issues in isolation in response to the comments received. Check it out here.

In 2012, we had a really unique opportunity for political data nerds such as myself. We had two radioactive social issues on the ballot, marijuana legalization and marriage equality, that were being decided by the same group of people. This gives us a rare chance to examine where people disagree on these two seemingly disconnected social issues. As you can imagine, these two initiatives usually had similar results (places that voted for marriage equality also voted for marijuana), so my excellent graphics team (aka Sweeney’s Portraits, hire them today!) put together a map that shows where one initiative did better than the other. As usual, you can click the map to make it larger.



The County

The County

Map Legend

Map Legend

Explanation of the maps: I took the percentage of vote supporting marriage and subtracted the percentage of votes supporting marijuana. The result gives us the difference between those two measures. For instance, if your precinct voted 68% yes on Marriage, and 70% yes on marijuana, you would have a difference of – 2% and be colored white on the map. If you voted 68% on marriage, but only 52% yes on marijuana, then you would have a difference of  16% and be colored the medium purple. Lavender I guess? Colors are not my strong suit. But you get the idea.

So what have we got here? At first, it almost looks just like a partisan map, with marijuana doing much better than marriage in the County, and downtown Bellingham much more comfortable with homosexuality than blunts. However, there is a little more going on here. Take a look at Lynden compared with raspberry country. Lynden, with its strong conservative religious tradition, did not take kindly to either marijuana or marriage equality, but the farmlands showed hefty support for marijuana. Which means that while these farmers might still object to marriage equality on religious or cultural grounds, they see the potential revenue or positive effects of legalization. Not all areas of the County are culturally homogeneous.

On the other side of the coin, you have WWU’s campus. If I had to place money, I would have thought that these precincts would show an equal amount of support for LGBT and weed, but lo and behold the campus showed one of the biggest gaps between the two measures. I believe part of that discrepancy is due to the sheer number of young families that live in those precincts. Although they were the target of the campaign, young mothers remain one of the demographics most opposed to legalization.

Those are some of my thoughts, what do you see?  One of the challenges of political analysis is you can examine how people voted, but rarely can you find out why. What trends to you see between these two bills?



  1. Sorry, I’m not graph savvy at all. Those percentages you have in the key mean nothing to me. The white means they were both rejected? The Purples mean they voted more for marriage equality than pot? The greens mean they voted more for pot than marriage? All I gotta say is that young mothers really need a better education about pot.

    • The white means that the vote totals were effectively equal. So whether they rejected equality and marijuana or they embraced it, the votes were the same. You got the rest correct. Green means that marijuana did better than marriage, and purple means marriage did better than marijuana.

      • So is there a map that shows which precincts rejected both and approved both?

      • I could build that, but it ends up looking basically like a partisan map of Whatcom County. Lynden voted no on both, and Fairhaven voted yes on both. This map shows were people disagreed between the two issues, which I felt is a much more interesting data point.

        In response to your comment, I added a text explanation beneath the maps. I hope that helps.

  2. I see, and thank you. This is very interesting indeed. Nice work, Riley.

    • Thank you!

      • Sorry Riley, this kind of analysis (one based on ‘difference scores’) isn’t very useful and rarely is. There are many reasons why one variable will score higher or lower than the other and you don’t know what they are. Worse, as several commentators note, it’s difficult to read presentations like these. Nearly fifty years as professor of psychology I learned what not to do with data … alas, this is an example.

        It would be better to analyze each factor separately and lay the data out precinct by precinct — the differences will be easy to read and (relatively) easy to interpret.

        Criticism offered in friendship.


  3. On it’s own it’s not too useful, However if you combine it with the knowledge you already have of the conservative vs liberal counties, then you are left with a clearer idea on the break down of conservative votes.

    As Riley pointed out the more farmland of conservative votes for 502 (using R74 as a baseline) by including both variables you effectively eliminate the people who voted for both or against both. So this graph could be used to find out where there are conservatives that have some push on financial issues and where less flexible conservatives live.

    I think this tells the same story on for more liberally minded city dwellers, where it isolates areas of the city where they are fine with gay marriage but something about the 502 law didn’t fly with them….again why?

    I am trying to think of how you could come to those conclusions with the variables separated,

    You could show them both independently but it wouldn’t screen out the people who voted for both or neither as easily.

    But then again I am not a Psych Prof so maybe Arthur can enlighten me on what I missing?

  4. I am missing?*

  5. While the percentage differences are if some interest, they are mostly useless without the raw numbers. An area with 80% approval of marriage equality and 70% approval of marijuana is radically different than one that breaks 40-30. Without the numbers, I have no way whether your young mothers hypothesis holds. Were this areas precincts below the city average on marijuana? If not, they would have to be more pro-equality

    • Mr. Burr is correct, which I why in the social sciences “difference scores” are not used much. Yes, as Devlin notes, you can dig and dig and find interesting stuff in Riley’s presentation but there are far easier ways to present and interpret these kinds of data.

  6. Great chart Riley. I would challenge the assumption that pro-weed votes outside Bellingham was a financial issue. Probably a reflection of the visceral dread those folks have against gay anything… And the libertarian case for weed.

    • Both solid points. I wanted the chart to be a jumping off point for discussion as to why certain communities voted the way they did.

  7. […] was a ton of requests for the two data sets that I compared in my recent post, Examining Social Issues in Whatcom. I grumbled and complained that, really, they breakdown of those two initiatives just show a pretty […]

  8. […] promised in “Part TWO!”, here is the final data set from my post “Examining Social Issues in Whatcom”, the marijuana legalization maps. I have used the same scale as the other two maps to keep […]

  9. Appreciate the effort put into these charts, Riley! It is at least SOMETHING about the county (including B’ham.) voters. I thought it was interesting that almost all of the cities( I prefer villages,which is probably not acceptable – ) anyway, cities are White. Not sure what that means though. Living on Lummi Island, our vote turned us White. And, I know we are always about 95% Democrat voters. I like that correlation – particularly as we begin to make efforts to change the conservative/ tea party/libertarian, etc. votes in the 42nd Legislative District. It has been a long time since we have had Democrat/progressive/liberal representation at the State level, or even at on the County Council. I’d like to see more demographics giving a breakdown to actual voters. Why you ask? We need PCOs in ALL the precincts and it will take pin pointing actual people who have our leanings in order to make that happen. Hey, the Republicans have already done this kind of charting. I see that as a CHALLENGE! Si se puede!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s


%d bloggers like this: