Del Ali, CEO of Research 2000, who won't release data to defend his polls against accusations of data manipulation and fraud by Kos, made a bold prediction today: The demise of the Daily Kos.
Del Ali, president of Research 2000 in Olney, Md., said he could not respond to the specific allegations Tuesday and referred questions to his attorney, who did not return a call seeking comment.
"I can tell you, we're fine. What we're going to reveal, that will be the end of the Daily Kos," Ali said. "I can say, it has to do with people owing money."
Image from TPM
Then it gets even better. He threatens Kos with "criminal sanctions".
Research 2000 President Del Ali tells TPMmuckraker in an email that the explosive claim by Daily Kos Tuesday that the firm fabricated poll data has dealt a "fatal" blow to his business, and, "Several long time clients who believe in us have stated that even after we are cleared and criminal sanctions are imposed on Kos, that they can not do business with us due to perception."
Instead of making ridiculous threats, Mr. Ali might consider a more effective means of saving his business. He needs to change the perception that he is a crook.
The perception, of course, was created by the polls themselves.
The day to day variability in a daily tracking poll where a candidate is at around 50% should fit a bell shaped curve. Increasing sample size will increase the sharpness of the peak, but it should always have a bell shape if it is a random normally distributed sample. R2K's curve has the wrong shape.
But it does make for a good hat.
Nate Silver explains at fivethirtyeight.com
You only get results like these if something is orders of magnitude upon orders of magnitude divergent from random. Now, just because Research 2000's polling is extremely nonrandom, does that necessarily indicate that it is fraudulent? I suppose there are alternate hypothesis, although the jury is out (or soon will be) on how compelling they might be.
They might cite their weighting procedures, but the weighting techniques that pollsters ordinarily use would not cause this kind of underdispersion. In fact, the normal weighting methods have the effect of essentially reducing the sample size (since you're effectively double-counting some voters while throwing out others), so they would increase, rather than decrease, the amount of variance relative to the sample. (** see note) But perhaps Research 2000 is using some really avant-garde techniques that have the effect of stripping a lot of variance out of the sample, e.g. a statistical model which uses polling as one of its inputs, along with making certain other fairly strong assumptions. This could also be a bug of some kind rather than a feature. Either way, it would take a lot of explaining on their part -- but it's possible.
If Mr. Ali wants to save his business he should release the records on how he processed the data. Threatening the demise of the DailyKos will only drive more traffic to the DailyKos.
Update
This story has gone viral. Techdirt.com makes an entertaining analysis.
Gallup - Random
The other statistical analysis that I found fascinating was that when you looked at weekly changes in favorability ratings, the R2K data almost always changed a bit. But, if you look at other data, no change is the most common result. As they point out, if you look at, say, Gallup data, you get this nice typical bell curve
Note the bell shaped curve.
R2K - Non Random
For R2k the option of no week to week change was a mysterious no show.