Replies

I just made some phone calls, got the run-around, but finally got in touch with the statisticians that actually worked on the polls.

Come on Mod's - this has got to win an award!! - That is too F'ing funny -

Funny as it is - facts are you are still wrong - Newsweek polls are junk - They simply adjusted their weightings to get their latest results -

You can ignore the link which I provided in my last post which will take you to where the Newsweek poll has been discredited by fellow-freepers -

And you keep thinking it is "how many they called" - They weighted it for a turn-out model that isn't close to any REAL turnout model....What don't you get about that??

And you keep thinking it is "how many they called" - They weighted it for a turn-out model that isn't close to any REAL turnout model....What don't you get about that??

Would you like the names and phone numbers? Or do you just not believe them? Well, you can deny the facts, but they're there. As for the number called, that was in fact your argument. See your own previous posts. And if you're arguing that they weighted the phone calls based on a model, that now tasks you with debunking that model. Your argument was that there was NO legitimate reason to weigh the phone calls, remember? Nice attempt to change the argument, but I'm not falling for it. Again, it is incumbent upon you to show:
Normalization was not used
AND/OR the normalization (in this case, the models to which you refer) was invalid.
Have at it, save the rant.

A recent story from another thread:

As CrushKerry observed, GOP respondents outweighed Democratic respondents in the poll, 374 to 303. With 300 Independents noted,that works out to 38.3% Republicans, 31.0% Democrats, and 30.7% Independents of the 977 total from the three. But the Newsweek poll states that 1,188 respondents were polled, with 1,088 Registered voters. And while the press release takes 20 pages to print and read, Newsweek doesn’t say, exactly, how those remaining 111 to 211 respondents breakdown. Also, I took the time to break down the respondents’ percentages: If the Rep/Dem/Ind response was unweighted, Bush would have led Kerry in the poll 58.0% to 42.0%, a sixteen-point lead, not eleven. So, while they don’t say exactly how they weighted their poll, Newsweek did weight the poll to balance things more towards the Democrats. In other words, you should consider deeper issues before accepting or rejecting a statement, whether it’s news you dislike, or news you do like.

Note the reference to "we don't know how Newsweek weighted their poll". Exactly what I've been pointing out. You don't know all the details of the weighting so guesses about what their poll really should be is pointless.

Yet more from Rasmussen - an admission of error:
September 6, 2004--We have been flooded with e-mails asking (in varying tones of politeness) why our poll results seem different from those released by Time and Newsweek.
There are two basic explanations, one involving our polling data and one involving the newsmagazines. For those who need to know the answer before the explanation, the bottom line is that the President is ahead by 4 to 5 points at this time. That's a significant improvement over the past few weeks, but not a double digit lead.
Our current poll (showing the President ahead by just over a point) includes a Saturday sample that is way out of synch with all the days before it and with the Sunday data that followed. In fact, Saturday's one-day sample showed a big day for Kerry while all the days surrounding it showed a decent lead for the President.
It seems likely that Saturday reflects a rogue sample (especially since it was over a holiday weekend). But, it remains in our 3-day rolling average for one more day (Tuesday's report). If we drop the Saturday sample from our data, Bush is currently ahead by about 4 percentage points in the Rasmussen Reports Tracking Poll.

Perhaps this explains Newsweek and Time's choice of normalization? Looks like all the contradictions may lie in faults of the Rasmussen poll -- the Saturday confound.