Why the Searcy Paper’s Judge “Poll” is Useless

Well, the Searcy paper finally found a way to run the headline that they’ve been dying to run for months: “LINCOLN LEADS JUDGE RACE.”

Why do they say this?  They ran a poll.  Why is the poll crap?  I’m about to tell you, because this so called “poll” is an insult to poll-lovers and statisticians everywhere.  I would accuse them of being ignorant, but they have been proven to be much more deliberate at their attempts to influence opinion for me to believe they are really just this stupid–or that they think the people of White County are this stupid.

(I know the paper is trying to hide behind a local university political science professor that they mention in their story to loan credibility to their poll, but I can almost guarantee he would agree with my analysis below.  He may have drafted the questions, but I am highly confident that he didn’t endorse this methodology.)

I ran this poll by a Republican political consultant who has worked on gubernatorial campaigns, ran targeting on congressional campaigns, and has run a targeted 527. Needless to say he has done more than a few polls. He agreed that the poll “has multiple issues with accuracy, and cannot be used to conclude that Lincoln has any kind of a lead.”

The alleged results of the poll:

  • Lincoln: 49%
  • Haynie: 34%
  • Undecided: 17%

And that huge “15 percent” number the paper tosses around can be a little deceiving.  The margin represented by that 15% is only 63 votes.  They called 410 folks, who we can only assume are actually registered voters, but based on the rest of their “methodology,” I’m not sure that’s a smart assumption.

1. They didn’t poll ‘likely voters.’  This is kind of a big deal.  The paper, according to their own story, didn’t make any effort to identify people who were actually likely to vote in the primary.  Sure, they asked people ‘do you definitely plan to vote,’ but that’s essentially crap.  There is much more that goes into determining likely voters than asking people on the spot, who will almost all say yes out of fear of being considered a ‘bad citizen.’  “Likely voters” should only be defined as people who, based on their voting history, are actually likely to vote.  I know, I’m a conspiracy theorist.

2. They didn’t poll identified Republicans, meaning those who are either registered as Republicans or have consistently voted in Republican primaries.  Oh, I’m sorry, you’re doing a Republican primary poll and including Democrats?  I’m sure my ‘ultra-partisanship’ will blind me to why this is a good idea.  But seriously guys, this is crap.  You aren’t getting meaningful results here and, once again, you are misleading the public by purporting crappy poll results as credible.

3. Lincoln’s name was placed first in the poll question.  See, the paper knows this is shady because they preemptively defend any attacks by saying, “Well, he’s going to be first on the ballot!”  More bull.  This isn’t the same as someone going into a polling booth and looking at two options.  The results will be slanted heavily towards the first person identified because people want to get off the phone.  The order of the names should be randomized.  Lincoln easily gained 5-10% from this trick.

Now, just because the results are ‘crap,’ that doesn’t mean we can’t still glean something from them, both statistically and politically speaking.  What have we learned?

1. Despite the attempts to slant this poll, Lincoln doesn’t even receive 50%.  This is the real story here.  Think about this:  a 6-year incumbent judge cannot even break 50%, despite being listed first on the poll which easily gave him 5-10%.  Subtract 10-points and a nearly 5% margin of error, it is very possible that this race is actually tied (Lincoln -15%, erasing his “+15%”).  As an entrenched incumbent, Judge Lincoln should be easily polling above 60% right now, and my consultant friend agreed.

2. This is a very close race.  Look, if this wasn’t a close race, the paper wouldn’t be running sketchy polls in an attempt to help out their favorite judge.  It’s really that simple.

Despite what you’ll read in the paper, these poll results are very good news for lovers of liberty & transparent government in White County.

Advertisements

36 comments

  1. Groan

    Why don’t you just call the good professor and ask him? His name and place of employment is listed in the Citizen’s story (Which, oddly, was not linked)

    Speaking of which, I’m sure folks would like to ask your Republican political consultant friend how he came to the conclusion he did. Think I’ll call him in the morning.

    What’s that? I can’t because he’s another one of your unnamed sources? Huh.

    So your buddy was either not confident enough to reveal himself, or you made him up. Which is it?

    • Arkansas Patriot

      1. I have too much respect for the professor to insult him so greatly.

      2. I don’t link to tabloids.

      3. I didn’t reveal my friend’s identity to protect him from unwarranted attacks from trolls. Do you really think you can scrutinize his anonymity while hiding behind a fake name? Hypocrisy at its finest.

      • Groan

        To recap:

        1. You respect the paper’s methodologist too much to even ask one question, despite repeatedly bragging about asking the tough questions.

        2. You respect your friend too much to be transparent in your analysis, despite being a self proclaimed a champion of transparency.

        And it is the Citizen that is a “tabloid?”

  2. Jake Pearrow

    I seem to really find this interesting that you simply throw this poll under the rug and stomp on it. This poll may very well be an “insult to poll-lovers and statisticians everywhere” but I don’t see any actual facts here that would truly discredit this poll and deem it the absolute crap that you make it out to be. You stated, “we can only assume are actually registered voters, but based on the rest of their “methodology,” I’m not sure that’s a smart assumption” but I ask why is that not a smart assumption? You have in no way displayed their methodology used in their poll. So basically either point 1 and 2 are just Nic Horton philosophy 101 or you know this for a fact, which if that is the case it is not represented here at all. And you say Lincolns name being first would sway the entire results 5-10% because people want to get off the phone? Oh please, that is a joke, if people want off the phone they are going to hangup long before these questions get asked. And who is going to really think of themselves as a bad citizen if they hangup on unexpected/unwanted poll? I can tell you no one in their right mind or with any lick of common sense is going too. You have given me and the other readers nothing more to go on than the word of an unnamed political consultant that just so happens to agree with you. Then you go on to say that ” it is very possible that this race is actually tied.” Whoa?! Where did that come from? From what I can tell it came from that extreme of the 5-10% that I’m still trying to figure out where came from, and from the extreme of the margin error. Now I’ve done more than my fair share of math and statistics and I just cant seem to see how it is “very possible” that this race is tied. And another thing, how can you in one sentence say that the results are complete “crap” and then turn around in the next and use these same numbers prove something? If the results and statistics are “crap” in one direction they are going to be “crap” in another direction, it doesn’t matter which way you turn them around. Although, you did present a few legitimate claims about the poll, there is nothing here that supports your reasoning or conclusion other than an unnamed expert and a conspiracy theory. There is a lot of talk and accusations of falsity here with little or no facts to give it any credibility. And Geez, this is the Daily Citizen in White County, AR conducting this poll, not Rasmussen or Gallup by any means; you know that and so do the readers of the paper, if they are as intelligent as you give them credit for in the beginning of your article.

    • Arkansas Patriot

      The results are “crap” because they are slanted. I’m not claiming that these were people polled from Faulkner county and the whole thing is irrelevant. I’m saying that there was an unfair advantage given to one candidate and the methodology was flawed, assuming the paper was looking for an accurate result. But that doesn’t mean that can’t glean a few things from it. Thanks for reading!

    • Scott Biddle

      There is a concept in statistics called “answer order bias”. This is why the state of Arkansas specifically calls for a random draw to determine ballot position instead of allowing it to be in party order or name order or incumbents first or any other structured order (in Arkansas code 7-5-207(c)(1)). Different statistical sampling methods reveal that there is a measurable difference in the effect of answer order bias in different scenarios (elections, polling, customer surveys, multiple choice tests). Whether this is the 5-10% that Mr. Horton has asserted is debatable. The existence of the impact of the answer order bias to varying degrees in varying situations is not debatable.

  3. Ben Williams

    You buried your lead Nic. The story is that this thing is wide open, not that you think the the Citizen sucks. We know how you feel about the issue and I doubt the above is going to change anyone’s mind. I know its like the GOP thing to get all up in arms over every slight, but its kind of tiresome, and you are better than that crap you listen to on the radio.

  4. Free American

    I agree, I think the poll is crap! The daily wipe is a liberal rag. Do you think that they are trying to sell any ads????
    Do not buy anything from these people!
    They will not let the truth get in the way of a good story!!! They are worthless! I am surprised that they are able to stay in business!
    Searcy needs a new newspaper! One that will print the truth and stands for freedom!!
    I am voting for Bill!!!!!

      • Free American

        My Pappy always told me if the shoe fit, then wear it!
        The only thing that the daily wipe is good? Yes, you are correct- wipe!

        Groan, why don’t you quit the paper and get a real job!
        Your welcome!

    • Free Thinker

      Reading this reminds me of the TEA Party meeting from a year ago, when Lincoln was pushing the tax increase for the bypass. His ‘consultant’ volunteered, but he was also being payed by the county to ‘help out’. Something does seem a bit fishy with this poll to me, as well.

  5. Simplemente Conservador

    Regardless of whether or not the poll is a good one, the fact a sitting county judge with considerable tenure can’t garner even a paltry 50% is the real story of this poll. If the poll is spot on, he has problems. If it isn’t, the likelihood is his numbers are worse than the likelihood the numbers are better than the poll indicates. Don’t be surprised, if this race stays tight, to see a large number of absentee votes in the upcoming election. Compare to the last few election absentee voter numbers. If this one has considerably more, then it would be a good exercise to obtain a list of said absentee voters and make a phone call thanking them for voting in the recent election. See how many profess to not know what you are talking about. Close races often end up with a margin well within the number of absentee votes cast.

  6. The Consultant

    To get the most accurate results you always want to limit the possibility of error by filtering your target universe. There are a multitude of things that can skew results. Most of which is just human nature. You also want to limit the errors caused by the respondent not hearing the question correctly or misinterpreting the information based on perception.

    First, when conducting a poll about an election you only select active voters. These are more likely to actually go vote. You don’t just take their word for it. You select them from a list of active voters registered with the Secretary of State’s office.

    Secondly, when polling for a primary you further filter you voters by who has actually voted in the target primary in the past or is registered within that party. This again increases the statistical probability that they will, in fact, show up to vote.

    You can then further filter by asking them if they are going to vote. All of these together give you the most accurate base of voters for your poll.

    Thirdly, you ALWAYS randomize the placement of the answers if you want an accurate result. The first answer will be chosen between 5% and 10% more than it would have if it was placed somewhere else. This is just part of human nature….doesn’t really have to do with getting off the phone…It’s just psychology, and common knowledge among politico’s who actually run polls for a living.

    The poll was presented with an almost 5% margin of error. If you assume that they somehow stumbled into the right polling group then you could look at the poll this way. 5% error plus a 10% error in the poll question would equal 15%.

    I tend to think the poll question probably created an error of somewhere around 5% – 7%. Although it is easily possible for it to be as high as 10%. I think the bigger assumption was that the poll somehow got the right voters to ask the questions. I just have no confidence in their targeting.

    Lastly, a long term county level incumbent should easily have 60% or more in a poll. The fact that this poll was skewed in the Judge’s favor and he did not get 50% should alarm his campaign.

  7. Scott Biddle

    Regardless of the actual polling methodology (questioning style, answer order, etc.) Mr. Horton pointed out that the single biggest flaw is that they asked people who said they planned to vote instead of asking REGISTERED REPUBLICANS. It doesn’t matter if the folks are randomized properly or are asked the questions in a different order. The Citizen chose to ask people who won’t necessarily even get a chance to vote in the election because they will be voting in the other party’s primary. This is akin to Europeans who were surveyed in 2004 regarding their choice for U.S. president and oerwhelmingly chose John Kerry. The U.S. media made a big deal about this at the time and the response ought to have been “So what? These folks don’t get to vote on this.” If you ran a poll asking several hundred andomly selected registered republicans who voted in the 2010 republican primary you could get a much more valid and reliable result, whereas this poll is structured in such a way as to not be valid or reliable (and I use the terms “valid” and “reliable” in their statistical sense).

    • Groan

      Only asking registered Republicans would be extremely flawed. Most registered Democrats and registered independents in White County will be voting in the Republican primary. If you don’t believe me, just ask them. The only action on the Dems ballot is one JP district (only 7.69 percent of the county) city of Kensett races (only 2 percent of the county, much of which is in the aforementioned JP district.) and Obama vs. some sacrificial lamb no one has heard of. Meanwhile the GOP ballot has County Judge for all the marbles, County Clerk for all the marbles and Romney/Gingrich/Paul. Anyone outside that 7 percent would be silly to pull a Democrat ballot.

      Do some research on how Gallup, etc., does its polling and you will be surprised how spot-on the Citizen is on this one. And for Pete’s sake, if you’re going to repeatedly call their methodology “crap,” would it kill you to — I don’t know — publish it?

      • The Consultant

        I agree that you would not just ask registered Republicans in an open primary. You would however, only ask known active registered voters who have actually voted in a Republican Primary in the past.

        If you look at the voting trend for White County over the last 8 years you will see very few crossover voting(despite what each side would like to believe). Voters who have voted in the democratic primaries are very unlikely to vote in a Republican primary and visa versa.

        Gallop, Rasmussen, and those of us who depend on accurate polls for our livelihood, only poll actual voters from a voter registration list.

        You are making assumptions that don’t actually play out based on known data.

        By the way… how did that Mayor’s race turn out?…. Didn’t look anything like the poll in the paper, and it used the same methodology.

  8. Groan

    I looked it up. The mayor race did turn out exactly the way the paper had it. 1. LaForce, 2. Morris, 3. Reeves. They didn’t do a poll for the runoff.

    Saying that names must be randomized is just something you just made up. Are names going to be randomized on the ballot? No. Who’s first is first, who’s second is second. Does it create a bias? Yes. The exact same bias that the ballot itself creates.

    Also, saying Galup, Rasmussen, etc., goes off registered voters based on voting history is hogwash. From the good people at Gallup themsevles, “The findings from Gallup’s U.S. surveys are based on the organization’s standard national telephone samples, consisting of directory-assisted random-digit-dial (RDD) telephone samples using a proportionate, stratified sampling design.”

    http://www.gallup.com/poll/101872/how-does-gallup-polling-work.aspx

      • The Consultant

        Stratification is the process of dividing members of the population into homogeneous subgroups before sampling. The strata should be mutually exclusive: every element in the population must be assigned to only one stratum. The strata should also be collectively exhaustive: no population element can be excluded. Then random or systematic sampling is applied within each stratum. This often improves the representativeness of the sample by reducing sampling error. It can produce a weighted mean that has less variability than the arithmetic mean of a simple random sample of the population.

        Meaning they filter their sample universe to achieve the most accurate results.

    • Free American

      Groan, if you love Mike Lincoln as county judge, then please tell what has this guy ever done for White County?
      Looks like you are trying to deflect on his issues! We can cuss and discuss polls all day. But that misses the real issue, White County needs a NEW County Judge!
      Again, please tell me what Lincoln has ever done, good for the citizens?
      He needs to go home!!!!

  9. Groan

    Come up with some proof to support your theory, other than your invisible friend’s opinion, then we’ll talk.

  10. Mr. Independant!

    You know Nic, I have been reading your stuff, for a bit and at first I tended to agree with alot of what you had to say, but it really seems after awhile now that if anyone disagrees with you then they are all just stupid and don’t understand, you attempt to belittle them and there remarks and thats pretty pathetic and childish if you ask me. Your attempts at trying to make Lincoln into a criminal are pretty pathetic as well, maybe one day you will grow up, but I’m not going to hold my breath on that, Good Luck becoming the Politician you want to be, you sure got a good start at being just like what we have now, all talk and no action,

    • Arkansas Patriot

      Thanks Skip. I have no aspirations of becoming a politician–who would aim for such a low mark?

      And I’m sorry you don’t like the law, but it is what it is. I don’t write it, I just try to hold politicians (who you claim to dislike) accountable to it.

      Thanks for reading! NHH

      • Mr. Independent!

        Unethical sure, illegal? Where are the charges? So far it’s just your opinion, petty stuff such as a sign being in the wrong place is not what I would deem a hanging offense. Nic didnt you just run for public office? Alderman I think it was. Hmmmm I think you qualify with that reply. I recall some old lady sayin she wouldn’t vote for Anyone with hair on there face lol
        Nice job on calling me out as well;)

  11. Groan

    Did you see the Citizen’s county clerk poll? Randall Young, whose name was read first, is dead last. Cheryl Evans, whose name was read last, is 24 points ahead of him.

    So much for you and your invisible friend’s “randomization” theory.

    • Arkansas Patriot

      Groan, shouldn’t you get back to work? Don’t you have a hit piece to write for the paper?

      Speaking of paper, it looks like they included the clerk’s poll in the judge’s poll. This hurts the poll’s credibility even more. And since you apparently can’t read, I never, not once said that being first in the poll would put you first in the results. I said it would add 5-10 points. But you know that.

      • Groan

        So, instead of 10% support, you’re saying that Randall Young’s support is actually 0-5%. Those same people who couldn’t wait to get off the phone during Question 3 just had to stick around to hear Cheryl Evans name read at the tail end of Question 4 because she would be such a gosh darn good clerk! Yeah, that’s the ticket!

        I’m also amused that, yet again, you see fit to question the literacy of someone who disagrees with a theory that is supported by absolutely nothing, except your blind hatred of all things Daily Citizen and Michael Lincoln. Oh, and your invisible friend, of course. After all, only an employee of the big, bad, evil newspaper could challenge such a sound and thoroughly-researched hypothesis. Maybe you should FOI Lincoln for all those uncounted Haynie responses he’s hiding in Tamara Jenkins’ new OEM building.

        Now, if you don’t mind, I think I’ll take your advice and get back to my work of destroying documents proving that Oswald didn’t pull the trigger, and video of the moon landing being staged. Keeping champions of “general knowledge among the people” like you busy is a tough job, but somebody’s gotta do it!

      • Scott Biddle

        Wait, did Groan just ask about FOI responses from Michael Lincoln’s office where he claims he is providing all documents and is provably lying? I’m fairly certain that the Patriot has solid evidence in hand that Michael Lincoln has been failing to comply with FOI laws. I wonder when we’ll see an article that lays all of that out for the public. What say you Mr. Horton? Will you lay your cards on the table and let the voters see the truth about Judge Lincoln’s true record of transparency before early voting starts?

    • Groan

      Apples to nuts? Read the article. Exact same poll, exact same respondents, exact same pollers, exact same nights, exact same methodology. Couldn’t have been more similar.

  12. Pingback: The Paper Turns «

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s