BROWN v. COAKLEY: LIES, DAMNED LIES, and STATISTICS

UPDATE FRIDAY, 01.14.10: A just issued Suffolk University poll now puts MA State Senator Scott Brown at 50% v. 47% for MA AG Martha Coakley in their bid to become the next U.S. Senator from the Bay State. The margin of error is + or – 4.38%. http://www.suffolk.edu/39994.html (Note: check out the cross tabs. You will see, this poll groups Independents together with Unenrolled’s, producing a far more accurate prediction as to how independent minded voters, both registered Independents and registered Unenrolled’s, will vote. This contrasts with previous polls mentioned below, which find Ms. Coakley besting Mr. Brown by separating Independents from Unenrolled’s and then excluding the voting preferences of these Unenrolled’s from the totals.)

If you had only accessed the Boston Herald for their report of this Suffolk University poll, you could have ended up as confused as the reporter, Jessica Van Sack. She writes, “Unenrolled long-shot Joseph L. Kennedy, an information technology executive with no relation to the famous family, gets 3 percent of the vote. Only 1 percent of voters were undecided.” See what she did? She called Joe Kennedy the “Unenrolled” candidate. He is not. Joe Kennedy, while often referred to in the press as an Independent Libertarian, is listed on the MA ballot as representing the Liberty Party.
http://www.sec.state.ma.us/ele/elespeif/senatorincongressmaelecan.htm
(Just to confuse you even more, when the Suffolk University poll asks voters for which candidate they intend to vote, it correctly refers to Mr. Kennedy as the Liberty candidate. But when it asks voters for favorability ratings, it calls him the Independent candidate.)

Contrary to the designation ascribed to Mr. Kennedy in the Boston Herald report on this just-released poll, there is no Unenrolled Party. In Massachusetts, there is only a category of registered voters not aligned with any party; these voters are registered Unenrolled. Yes, Unenrolled voters tend to be independent minded; but this does not make them members of either the Liberty Party or the Independent Party.

Read the Suffolk University summary. They got it right, even if the Boston Herald did not.
*************************************************************************************************

I think I might have ‘figured out’ ‘what’s up’ with the apparent disparity in 2 (two) of the recent polls for U.S. Senate in Massachusetts.

Results of the latest Boston Globe poll put MA AG Martha Coakley up by 15 points in the race for U.S. Senate against State Senator Scott Brown. This hit me as really odd. The PPP poll just out put SB up by 1! So, I investigated further. I saw, the Globe poll ran from January 2-6 and has a margin of error of 4.2 whereas the PPP poll ran from January 7-9 and has a margin of error of 3.6. (Generally, the larger the sample, the smaller the margin of error.)

(The Globe runs liberal and PPP is paid for by the D Corporation, so I discounted the political bias as effecting these skewed results. And they both used likely voters, which tends to make the poll more reliable.)

In other words, in order to figure out what the skewed results in these 2 polls mean, I considered factors such as who paid for the poll; what was the sample used; whether this was a one-shot poll or a rolling poll, taken over a few days (and look at those days); and what was the margin of error.

But questions posted on the blogosphere point to another problem with accurately interpreting these polls as an indicator of voter preference.

Texasdarlin writes,

I am confused by the party affiliations cited in the Globe poll. I’ve seen on various blogs and news sites that the percentage of independents included in the Globe poll was very low (around 18%), but that’s not how I read their tables.
Can anyone double-check this and figure it out? Here is the survey:
http://www.boston.com/news/politics/2010/senate_race/0108010_poll_senate_race/

http://texasdarlin.wordpress.com/2010/01/10/boston-globeuniv-nh-coakley-up-by-15-really/#comment-112419

And Hot Air writes,

The new Boston Globe poll showing Democrat Martha Coakley beating Republican Scott Brown by 15-points in the race for Ted Kennedy’s Senate seat, when contrasted with Public Policy Polling showing a dead heat, has people scratching their heads. So what’s up with that?

http://hotair.com/greenroom/archives/2010/01/10/whats-up-with-the-mass-senate-polls/

I think I figured out what happened to MA Independents in these polls, too.

The Globe poll asks this question (p.3): ‘Are you registered D, I, R, or something else?’ How would you answer if you were registered as “Unenrolled”?

Massachusetts has a category of voter Registration called “Unenrolled.” The biggest difference in these categories is this: Independent voters may vote only in elections but Unenrolled voters may vote in primaries, too. They declare which ballot they want; vote; and leave the voting booth an Unenrolled voter, the same registration status they held when they went in.
http://www.sec.state.ma.us/ele/elespeif/us_senate_info.htm

When pollsters asked Unenrolled voters who participated in the Boston Globe poll, ‘Are you registered as D, I, R, or something else,’ these voters would have answered, ‘something else,’ right? However, the answer categories to this question are D; I/Unaffiliated; R; and Other. Did you catch that? The “I” in the answer category includes a designation “Unaffiliated,” but the I in the question does not include this word. So, when pollsters polled Unenrolled voters who answered “something else,” these responses likely were recorded in the category that reads, “Other.” This means that, the “I/Unaffiliated” answer category would only count those voters who answered that they were registered “I’s” but it would not count those independent-minded registered Unenrolled voters who answered “something else.”

And more than 50% of registered voters in Massachusetts are Unenrolled.
http://www.sec.state.ma.us/Ele/elepdf/st_county_town_enroll_breakdown_08.pdf

Looks like perhaps the Globe undercounted those independent minded voters who are registered as Unenrolled.

On the other hand, the PPP poll specifically categorized Independent/Other as a single grouping. Thus, “Independents” in the PPP poll would include all of those voters registered as both “Independent” and as “Unenrolled.”

Advertisements

3 Responses to BROWN v. COAKLEY: LIES, DAMNED LIES, and STATISTICS

  1. TeakWoodKite says:

    Hey jbjd, unrelated but you were cited and I thought you my wish to know.
    http://www.facebook.com/note.php?note_id=276398546214&id=100000555737985&ref=nf

    I get back to reading your article. Thanks.

    TWK: Thank you for letting me know this! ADMINISTRATOR

  2. TeakWoodKite says:

    If I understand the outcome is the Globe is inaccurate due to a vague question as to affiliation?

    TWK: Technically, the problem was, one of the questions they asked about affiliation had no corresponding answer category to fit that question. So, given the poll results, what I figured out must have happened is that, the Unenrolled respondents to that question were grouped into an answer category that effectively masked not only their numbers but also their stated voting preference for Brown. And this ‘undercounting’ of voters against Coakley resulted in an over-sampling of D’s, which skewed the results toward her.

    This particularly irked me because, even assuming the UoNH, which conducted the poll, was confused that their neighbors have a discrete category of voter registration called “Unenrolled”; surely, the Boston Globe knew better. And I cannot imagine they released the poll BEFORE scrutinizing the results. (I cannot imagine they did not scrutinize the survey in advance, given that they commissioned the poll.) As if the reputation of this once great newspaper had not already been sufficiently eroded with their lopsided coverage of the 2008 election cycle… ADMINISTRATOR

  3. jbjd says:

    There is speculation from some R lawyers that current U.S. Senator Kirk from MA cannot legally cast votes in Congress after Tuesday’s special election to pick his successor. http://www.weeklystandard.com/blogs/barnes-massachusetts-senatorial-race-and-obamacare. They based their conclusions primarily on the wording in the new law, which reads in part, an appointed senator remains in office “until election and qualification of the person duly elected to fill the vacancy.”

    I read the new MA law differently from these R lawyers, with reference to the Definitions at the beginning of the chapters dealing with elections. In that section, there is no definition for “elected” or “duly elected.” However, there is a definition for the word “election.” Accordingly, using the word “election” in the new law means, the candidate who will now fill that US Senate seat will do so after the voting has ended. But there is an additional requirement in the new law that restricts this requirement “until election.” That is, the candidate must also be “duly elected.” The R attorneys above construe this as meaning, s/he satisfies the Constitutional requirements of eligibility. But this makes no sense. Because in MA, no law requires any state official to verify such eligibility either before the candidate’s name is printed on the ballot; or before the Governor certifies the results of the election.

    It makes more sense that “duly elected” in the new law means, elected according to provisions of law. And under this law, it looks like, the results cannot be certified before the 15th day after the election.
    http://www.mass.gov/legis/laws/mgl/54-116.htm

    Because by law, a candidate has 5 (five) days to file a petition for recount. Here is the ballot certification procedure spelled out.
    http://www.sec.state.ma.us/ele/elerct/rctidx.htm

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

%d bloggers like this: