« Gallup: Dems Like Hillary/Obama Ticket, But Say "No" To Obama/Hillary | Home | Harold Ickes Confirms That Wright Is Key Topic In Discussions With Super-Delegates »
Gallup: No, Our Daily Tracking Poll Is Not Flawed
Yesterday we brought attention to a possible flaw in Gallup's tracking poll, a so-called "day of the week" effect.
Now Gallup has responded to Mark Blumenthal, who first raised the subject — and they lay out a pretty good case that no such effect exists.
Giving the average figures for all the daily samples taken in 2008 — which the public normally doesn't get to see — the numbers show no significant differences for the candidates depending on the day of the week:
Gallup acknowledges one possible quirk, but even then it doesn't look like much: "Clinton has had some good Mondays in March, basically every other one, but even with that her Monday average is only 3 points higher than her overall March average."
Advertisement
How many times did HRC or a surrogate appear on a Sunday Morning Talk Show in March?
April 1, 2008 2:41 PM | Reply | Permalink
Folks: It is now official. Hillary is bonkers. She is now clearly delusional.
First you had her Fairy Tale claim about her Fantasy Warrior deeds in Bosnia.
Today she compared herself to a fictional character in the movies.
Here is what she said. She has lost all grip on reality.
PHILADELPHIA - Perhaps the analogy was inevitable: Hillary Rodham Clinton as Rocky Balboa, the scrappy underdog boxer from Philadelphia memorably depicted in the 1976 Oscar-winning film. Even if Rocky did lose his first big fight.
ADVERTISEMENT
Addressing a meeting of the Pennsylvania AFL-CIO Tuesday, the former first lady and New York senator said that she, like Rocky, wasn't a quitter.
April 1, 2008 2:53 PM | Reply | Permalink
Does she remember that when we first meet Rocky Balboa, he's a half-retarded loan shark enforcer?
Is this the analogy she's trying to strike, that she's knee-deep in organized crime and not smart enough to spell Iowa if we spot her all of the vowels???
April 1, 2008 3:10 PM | Reply | Permalink
She is no Rocky. She is Rambo the Hollywood hero,"i got no time to bleed"
Xena is the hero of Tuzla.
April 1, 2008 3:12 PM | Reply | Permalink
Interesting she would compare herself to Rocky, or even Rambo, two movies that both have conservative undertones.
April 1, 2008 4:00 PM | Reply | Permalink
Those numbers (in chart above) are from when?
April 1, 2008 2:56 PM | Reply | Permalink
The article states that they are the average of all the single day polls so far in 2008
April 1, 2008 3:19 PM | Reply | Permalink
As Gallup states in its reply to Blumenthal, those are averages of all daily polls taken Jan. 2-March 31. (The initial discovery of a supposed pattern was based only on data from February and March. So adding just one month to the data being analyzed causes the anomaly to virtually disappear.)
Like I said yesterday, you cannot slice and dice poll results this finely and expect meaningful results. Random spikes like these will always emerge. They don't require any sophisticated explanation -- because they aren't real.
April 1, 2008 3:38 PM | Reply | Permalink
Her numbers rise on Mondays.
I understand that heart attack rates also go up on Mondays
April 1, 2008 3:44 PM | Reply | Permalink
"Day of the week" has always been a known variable in polling data. It is why good pollsters make certain that attempts to reach a specific telephone number or set of numbers are made on different days. It is actually interesting that no such variation appears in the data.
The question I'd like to see asked of pollsters is: How much of your sample was "dialed to completion"-- that is had at least four attempts made to complete an interview? AND what is the adjusted level of confidence if not all sample was dialed to completion?
Failure to dial to completion could easily expain
"strange" poll results. A simple example being that those people who are reached on the first attempt are nearly always different demographically from those reached on later attempts. The level of confidence falls drastically with fewer attempts.
April 1, 2008 4:31 PM | Reply | Permalink
"Yesterday we brought attention to a possible flaw in Gallup's tracking poll, a so-called "day of the week" effect.
So, since you guys went to the trouble to say it was flawed yesterday, what are you blaming today? Gravitational forces? Too many people on the wrong side of the earth? Collective farting by contented cows?
April 1, 2008 4:45 PM | Reply | Permalink