Skip to main content

View Diary: Wow:These Ohio Maps Show Supression... (72 comments)

Comment Preferences

  •  Some Graphs of machines w/ turnout (4.00)
    I was working on a Diary about the voter turnout compared to Machines per voter, but it looks like this thread is a better place to put it.  The source for this data is on another machine, but yesterday one of the diaries linked to an article with a list of precincts with voter turnout, number of registered voters per machine and percentage of voters for Kerry.  I put them up in a graphic to try to get an idea of what was actually happening.


    Pay attention to the right half of the graph.  Once we get past 250 or 300 voters per machine, the numbers clearly start to go down.  The running average is clearly depressed by about 10%, to me this is clear evidence of voter suppression.  There isn't enough data to be certain, but my intuition tells me that the graph falls as 1/x, saying that there is a fairly fixed limit to the number of voters per machine who can and will vote.  So, a lack of machines depressed voter turnout, a fairly obvious conclusion.  Glancing at this graph, I would say that all else being equal we should fight to make sure that all precincts had no more than 250-275 voters per machine.


    This graph correlates the number of voters per machine with the percentage vote for Kerry, it is much more noisy and harder to draw conclusions from.  As has been previously discussed, the worst precincts for machine limiting voter suppression are pro-Kerry.  The article I got my data from did not include actual registration numbers, but it we were to look them up we could "normalize" the results to determine about how many votes Kerry should have gotten if these precincts hadn't suppressed the vote.

    Of course, it isn't possible to statistically model elections.  They won't even let us model the census.  What we could calculate with a robust enough model is how many votes were prevented at those precincts by insufficient machines and determine if it accounts for enough votes.  I assume that there is no way to prove that someone was disenfranchised by an excessively long line and find them to permit them to vote.

    Here is the raw data that I used to make these graphs.  Please note that one precinct had no data for Kerry percentage this turns up as a 0 in my data that accounts for that blip in the Kerry Vote graph.  also note that the two precincts with less than 150 voters per machine only had two machines, if they had only one then they too would have been suppressed.

    The man who trades freedom for security does not deserve nor will he ever receive either.-Benjamin Franklin

    by Luam on Tue Nov 30, 2004 at 02:12:50 AM PST

    •  can you (none)
      try and cut the top graph in half, compare to each other, and see how much of the top graph's right side readings came from democrat heavy precincts?

      eg. is there a clear case of suppression in certain districts?

      •  The short answer is Yes, but... (4.00)
        It could be done easily.  I decided not to make a case that Kerry voters had a lower turnout.  The right has contended that they did better GOTV when in fact they did better KOTV (Keep or Kick).  So there is little point in saying that we had a lower turnout.

        But if you look at the tail of the top graph and compare the same points on the Kerry graph, again look at the right sections, you will see the correspondence yourself.

        BTW I messed up the data link

        The most suppressed precincts are:
                             V/M        % Turnout    % Kerry
        MIFFLIN                   322.5    47.44    65.88   
        WHITEHALL_2        324.8    53.81    53.86   
        REYNOLDSBURG_4    325.6    52.73    50.44
        FRANKLIN            348.9    49.77    49.32   
        CLINTON            351.1    48.73    57.07   
        WHITEHALL_4        368.9    51.99    60.31

        The man who trades freedom for security does not deserve nor will he ever receive either.-Benjamin Franklin

        by Luam on Tue Nov 30, 2004 at 03:02:26 AM PST

        [ Parent ]

    •  Great post, thank you (none)
      I found your argument easy to follow and you stated clearly what you thought the data and graphs showed and didn't show, which is helpful for people like me whose eyes glaze over at graphs of this type.  

      These clips from your post were especially meaningful to me:

      ... The article I got my data from did not include actual registration numbers, but i[f] we were to look them up we could "normalize" the results to determine about how many votes Kerry should have gotten if these precincts hadn't suppressed the vote.

      ...What we could calculate with a robust enough model is how many votes were prevented at those precincts by insufficient machines and determine if it accounts for enough votes....

      "Now watch this drive."

      by tompaine2004 on Tue Nov 30, 2004 at 04:05:47 AM PST

      [ Parent ]

    •  I disagree with one point... (none)
      "I would say that all else being equal we should fight to make sure that all precincts had no more than 250-275 voters per machine."

      But the way I figure it, if you are allowing each person 5 minutes to vote, then you allow one minute to get the next voter in place...in 12 hours of voting, only 131 people can vote on one machine.

      12 hours x 60 mins per hour divided by 5 1/2 minutes per voter.

      250-275 voters per machine will still leave long lines and massive disenfranchisement.  I'm thinking more like 120 voters per machine should be the ideal.

      •  Different math (none)
        I thought about running the numbers the way you have, thanks for doing that.  What I was looking at is that the turnout seems unaffected by the number of machines until you get to that point.  Perhaps the reason is that most precincts only has a 60-70% turnout which means that the machines go that much further.

        We need reform which makes sure that the lines never get long enough to discourage voters.  The reason I said all else being equal is because different machines and ballots will change the critical number.  We could start with something like, whenever mathematically possible no more than 10% variation in the number of voters per machine.

        BTW I realised that I said the numbers were for Ohio, but they are only for Franklin County (Columbus area).  Also I found a copy of the article I was referencing.

        The man who trades freedom for security does not deserve nor will he ever receive either.-Benjamin Franklin

        by Luam on Tue Nov 30, 2004 at 05:47:43 AM PST

        [ Parent ]

        •  Okay! (none)
          Now I see your point, the difference between registered voters and actual turnout.  So, say we have 70% turnout and 131 is 70% of 187.  Maybe the ideal should be 180 voters per machine?

          I'm just thinking that in negotiation you rarely get what you ask for.  If we demand 275 voters/machine, maybe we'll get a standard of 325 voters/machine.  If we ask for 180 voters/machine, maybe we'll get a standard 225 voters/machine.

          By the way, I should have said this in my first post, good analysis and thanks for your efforts!

    •  Little error. (none)
      The horizontal axis is mislabelled in both graphs: It should be (I presume) voters per machine.

      But the point is clear.  Terrific work.  I'm sure it's worth a diary.

      We're just getting started.

      by jem6x on Tue Nov 30, 2004 at 09:57:58 AM PST

      [ Parent ]

    •  You inspired me... (none)
      ...to make a graph of my own

      Ohio 2004 turnout

      This shows both the top 60 and bottom 60 precincts for provision of machines, along with whether the precinct went for Kerry or Bush.

      Clearly the bottom sixty are much more likely to be majority Democratic.  I'd be interested to see the data for the precincts in between, to draw a pair of regression lines through the red and blue dots.  

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site