Evaluating the polls....

[Nazgul35]

I have seen some wierd stuff of late coming from the polling groups...which reflect my concerns about polling in general. Remember the shifts in public opinion in the 2000 elections?

Consider the latest Zogby tracking poll which shows Kerry leading Dean. http://www.zogby.com/news/ReadNews.dbm?ID=780

Reading the fine print about how the poll is conducted, you see the following:

Polling results will be released daily through Monday, January 19th, the date of actual caucus voting in Iowa.Zogby International conducted telephone interviews of a random sampling of 502 likely caucus voters statewide. All calls were made from Zogby International headquarters in Utica, N.Y., from Monday (1/12/04) thru Wednesday (1/14/04). The margin of error is +/- 4.5%. Slight weights were added to party, age, education, union, and gender to more accurately reflect the voting population. Margins of error are higher in sub-groups.

First, these polls are of likely caucus voters which would not capture the new voters that have been brought in by Dean and Edwards and would actually favor Kerry and Gephardt. Dean has many people who have not participated and Edwards has drawn in some young voters...

Second, how is the poll weighted?

Reported frequencies and crosstabs are weighted using the appropriate demographic profile to provide a sample that best represents the targeted population from which the sample is drawn from. The proportions comprising the demographic profile are compiled from historical exit poll data, census data, and from Zogby International survey data.

So Zogby is further skewing a poll that already relies on likely voters to begin with....

I am not going to suggest that anyone candidate is walking away with Iowa, but I am suggesting that the polling that is being conducted could in fact be supressing some candidates while inflating others....

We will all find out on Tuesday...just don't get all caught up in poll numbers folks......

[Padraig18]

1. I think Zogby is probably WAY off.

Dean's internal polls may be a BIT weighted in his favor, but they are DRASTICALLY different than Zogby.

[Nazgul35]

3. I didn't even talk about polling fatige

if you poll over and over again, eventually people stop responding, so Zogby must continually search for more respondents to avoid going below the 500 threshold....

I just remember how screwy the polls were in 2000 in the run up to Nov.....if they were true...20 million Americans were changing they minds every other week.....

[stickdog]

2. I'm still looking for any friendly wagers from folks convinced Dean will

lose Iowa.

Funny, no takers so far.

Maybe I need to give them a spread?

[Nazgul35]

4. I may not be convinced he will lose....

but I am not ready to say he will win either...and I support him....

It's going to be a close one!!!!!

I think that the surge of Kerry is not good news for the Clark campaign...Edwards either.....

If it is Dean, Kerry, Edwards in the top three, they will be getting nonstop media for two weeks....which will be flooding NH....in addition to any money those campaigns seek to spend....

If anything, it'll keep the repugs guessing as to who to attack!

[bearfartinthewoods]

5. it depends on how they determined 'likely'

if they are working from lists of previous caucus participants, i agree with your point about new voters. but if they are working by calling random numbers and asking 'are you likely to caucus' first, then i disagree.

one issue re weighting...it's cellphones. younger, poorer are more likely to use cellphones whose numbers are not yet available to pollsters.

[sfecap]

6. One thing to consider:

Zogby's "random sampling" of those 502 is only 167 people. That is, only 167 actual responses are used. The daily MOE is in the region of 10%. The tracking poll is averaged over a 3 day period in order to reduce the MOE to 4.5%, which is still quite large.

Zogby's polls are interesting, but mean little.