Blue Mass Group

Reality-based commentary on politics.

Comments

  1. This map appears to be almost, but not quite, divided along Congressional Districts.

    <

    p>
    As a proud resident of Left Fields, in the District of the distinguished Congress Member John Olver, one of the 5 of 10 to endorse Deval pre-Primary, I am delighted to be colored dark blue.

    <

    p>
    We will continue to work hard to bring out our vote to offset the less enlightened areas of the state. Don’t despair! This used to be a very Republican area, but obviously, people have seen the light!

  3. I can’t believe, by the way, that MidMass is not in your same boat. I hope that Murray will help us keep the dream alive here!

  4. … I’m not quite sure how convinced I am of the validity of the regions.  Some clearly do make sense… “Left Field”, “Brink Cities”(mostly), “MidMass” seem to work.  However, I live in ‘Ponkapoag’ and the region doesn’t seem to make much sense.  It crosses several Congressional Districts – the 10th (Delahunt), the 4th (Frank), the 3rd (McGovern).  It also lumps vastly different communities together.

    <

    p>
    Quincy is one of the largest cities in MA.  Brockton is as well.  But that doesn’t make them similar.  Ethnically, Quincy is ~80% white, 15% Asian (one of the highest in the country) and 2% black.  In Brockton, its 60% white, 18% black, 2% Asian.  Brockton has 4 times the Hispanic population (8% to 2%) and 2/2.5 times as much poverty (5.2/7.3 vs 12.1/14.5 individuals/families) as percentages.  Quincy is almost an extension of Boston and Brockton is a city almost by itself.

    <

    p>
    And those are cities that are quite similar.  Places like Avon or Abington seem more like Hanover or Canton than Brockton or Braintree.  The piece seemed to reiterate how often these regions had voted differently… perhaps that should be seen as a mark against their 10 region paradigm rather than shifting ideology/partisanship over a short period of time.

    <

    p>
    For instance, I’ll go back to Brockton and Quincy. 

    <

    p>
    Brockton went  2,617 4,678 2,304 for Gabs, Deval, Reilly.
    Quincy went  4,889  4,983 4,219… a pretty notable difference.

    <

    p>
    Weymouth went 3,002 3,175 2,815
    Norwood went 1,995 1,762 1,632
    Braintree went  2,604  2,059  1,519
    Stoughton went 1,411 1,932 1,016
    Randolph went  1,496 2,800 1,102
    Hull went 531 926 356
    Canton went  1,256 1,582 900

    <

    p>
    You’ve got rural affluent small towns with median incomes of 80K+ that went for Patrick by a relatively small amount to a large city that nearly reached 50% to a medium population town that topped 50% for Deval, to a medium town that Deval lost to a large city that was basically a wash… You’ve got affluent small towns and a city that could easily count as one of the ‘Brink Cities’.

    <

    p>
    Perhaps its just my insomnia making me be extra picky, but how valid are these regions?

    • In most of the towns in my area, Deval’s margin of victory (or loss) was a direct result of who was organizing the town (if there was anyone) and how hard they worked. It was pretty predictable that Deval won towns that he should have lost, because of the groundwork. He did (comparatively) poorly in towns where there was nothing going on. 

    • Hi all,

      <

      p>
      Thanks for posting a link to our analysis at MassINC and thanks for the comments.

      <

      p>
      As for how the 10 regions were drawn, I mostly relied on results from past general elections, and not as much on results from primaries. Ponkapoag, for example, is made up almost exclusively of cities and towns that shifted toward Paul Cellucci in the 1998 race — as opposed to Shopper’s World towns that generally shifted toward Scott Harshbarger or Cranberry Country towns that had already shifted Republican in 1990. There are two towns in Ponkapoag, Sharon and Randolph, that did not shift toward Cellucci, and in the 2002 version of the 10 regions I put them in Shopper’s World, but I was worried that the regions were getting too checkerboard-like. They need to have some kind of geographic coherence.

      <

      p>
      I didn’t want to base the regions on demographic factors such as income or race because I think they are given too much attention in political analysis anyway. All public-opinion polls have this data, but they generally don’t take into account the role of geography in an election. (A vivid example this year was in the lieutenant governor’s primary, where Tim Murray won Central Mass. — or MidMass — by such a wide margin that he obviously transcended demographic or ethnic blocs.)

      <

      p>
      Robert David Sullivan

      <

      p> 

      • we could have these purdy 10-region MA maps colored in for Tim Murray’s results as well?  It’d be interesting to see the “overlap” of the two.  Will the relative levels of success of Patrick and Murray in the same region give their ticket the “max”, “min”, or “mean” of their individual results?

        <

        p>
        How bout some pikshurs?

      • I don’t want to seem to be completely dismissing your analysis.  As I said, some of it seems to correlate with reality and politics well.  Its just I wonder if the 10 region paradigm you set up was to restrictive.

        <

        p>
        Weymouth (my town) does certainly have Southie/Dot roots.  My family is a Southie family, my next door neighbors are Southie, my girlfriend’s family are from Dot, etc etc.  But as I said the same can’t be said of Brockton or Quincy.  Hanover (listed in Cranberry Country I believe) has similar roots however (in my experience).  There’s a more affluent to middle class city, several middle class small cities/large towns, a working to middle class city and several towns of differing economic status (the same could be said ethnically and blue/white collar). 

        <

        p>
        My suspicion is this:  Why 10 regions of approximately equal voter size?  Thats really two questions… why 10?  and why do they have to be equal size?  In fact, expecting there to be 10 areas with approximately equal number of voters seems to be such a huge and unlikely assumption that I wonder why it was made in the first place.

        <

        p>
        (There are mathematical/computer science algorithms that will “cluster” data into natural groupings.  These do not require that the population of each set includes the same number of members – although accounting for geographical location might be difficult.)

        • For example, I know that increasing my vote share 4% in one region is just as valuable as increasing it by 4% in another (unless there are significant turnout differences between the regions).  You could do the same thing with the congressional districts, but Massinc’s regions are substantially more intuitive.  In particular, they noticed that three geographically very distinct regions (the Valley and far west, Cambridge/Somerville, and parts of the cape and islands) all vote very similarly, so they grouped them as a bloc in a way that the congressional districts couldn’t.

          <

          p>
          It would be interesting to see what a clustering algorithm would give you, but it’s probably easier for a knowledgable human to come up with intuitive districts.

      • Great analysis, but I just wish you would modify the map to indicate that Left Fields includes THREE communities inside 128, not just Somerville and Cambridge.

        <

        p>
        We’re proud to be associated with our kindred spirits west of the Quabbin and on the cape!

  5. I donno, those “post-industrial” cities north of Boston have a lot of minority voters who will probably vote Dem.  Especially considering that Healy and Mihos are both shouting a lot about evil immigrants.

    <

    p>
    And my friends who live up there are generally people who have moved there in the past 6 or so years because it’s cheaper then “shoppers world” and “bigger boston.” 

  6. What a great analysis!  Thanks to MassInc for conducting it and pablo for linking to it.

    <

    p>
    The towns with 60+ Patrick vote seem to break down into (1) MassInc’s “Left Fields” region, (2) a block of towns in Northern MetroWest, and (3) a cluster of towns around Cape Ann.  I understand (1), and (2) seems to have been where the most active town-level Patrick organizations were.  But what’s the reason for (3)?  Is it just that there were fewer Democratic primary voters there?  (I notice that the Essex region came up white for a low value of Patrick primary vote divided by total November 2002 vote, for example.)

Archives