Yes, Chuck Todd is a confirmed cherry picker of polls.

[grantcart]

On an earlier thread I posted that Chuck Todd was becoming increasingly unreliable in his handling of poll numbers.

I specifically accused him of 'cherry picking' polls.

Remarkably two members (maybe more) have coresponded with Todd and he has responded with a question back with the following,

"who cherry picks polls? where is this coming from? can you name a segment where this happens? not sure there is anyone on air who is MORE careful with what polling they use;"

Rather than responding on one to one I'll make the facts public so that everyone who wants to respond can follow what Mr. Todd said and what his response will be.


On February 11th Chuck Todd ran a graphic on MSNBC that showed the President at 44 points and running against an "unnamed Republican candidate" at 40 points. It showed the results from February 1st Gallup poll.

His commentary with that graphic was that the President was losing support among "Democrats" and liberals.


Here are the problems with Chuck Todd's use of polls and why the charge that he 'cherry' picks polls can be easily substantiated.

1) 'Cherry Picking
The use of a poll with any named candidate against any "unnamed" candidate is a highly questionable technique. The reason is that polling exaggerates negatives. The first thing that you learn in political science is that when you are polling in advance of a political campaign throw away all of the positive polling and concentrate only on the negative numbers. Candidates who have high negatives almost always lose.

The reason for this is that people are generally willing to express a positive opinion about someone even though then know little about that person. These people, obviously, are vulnerable to change a rather weak alliance to a superficial opinion. People expressing a negative opinion, however, have gone through a more exacting thought process to form that opinion and it is much harder to change a negative into a positive than the reverse.

The classic example is to take two candidates, one with high positives and high negatives versus another candidate that has low postives and low negatives. Almost always the second candidate will win. This was the scenario with Obama and Clinton in the primaries. President Clinton started with very high positives - 43 and high negatives. The end result was that Senator Clinton fulfilled Poli Sci 101 by remaining flat and finishing about where she started.

Now when Todd uses a poll using Obama against an 'unknown' Republican he is using a poll that will use all of Obama's negatives and none of the personal negatives that any of the leading Republican candidates have. The only value that such a poll would have is if it was done regularly and you could graph movement up or down. Using it on a one off situation is almost useless. Using it when there are hundreds of other polls is professionally irresponsible. Gallup itself has two other tracking polls, a dailly and a weekly, that are of much more value.

So using a one off poll versus a tracking poll, especially when the same polling company has tracking polls is, by definition, 'cherry picking'.

2) Todd's conclustions are clearly wrong

Todd's editorializing that Obama has experienced loss of support among "Democrats" and "Liberals" is not supported by the poll that he quoted (which didn't show tabs for those groups) and is in fact 100% in contradiction against what the empirical data is showing. How is that possible? I don't think that Todd 'has a dog' in this but rather,

Todd is listening to the punditry and not looking at the numbers

The numbers show that over the last six months the President's numbers aren't moving much at all.

Pollster.com shows that in their "Poll of Polls" Obama had a 50% approval rating on August 18 and 6 months later his approval rating is 48.4%. (This includes the ridiculous Rasmussen polls.)

However even more telling is the breakdown by self identification.

Compare the numbers from the end of November and today

. . . . . . . . . November. . . Today

Overall approval. . . 50. . . . . 51
Democrats . . . . . . 82. . . . . 83
Liberals ... . . . . . 78. . . . . 79

November (cannot find 'gif' link)
http://journals.democraticunderground.com/grantcart

Today (cannot find 'gif' link)
http://www.gallup.com/poll/121199/Obama-Weekly-Job-Approval-Demographic-Groups.aspx

Last Week




Other well established polling firms confirm gallups numbers.

Here is Pew's numbers and comment:


http://people-press.org/report/589/midterm-electionchallenges-for-both-parties

The latest nationwide survey by the Pew Research Center for the People & the Press, conducted Feb. 3-9 among 1,383 adults reached on cell phones and landlines, finds continuing public dissatisfaction with the economy and disapproval of major policies to address it. Yet President Obama’s overall job approval ratings have remained steady in recent months. Currently, 49% approve and 39% disapprove of the way Obama is handling his job as president, which is largely unchanged from surveys since October.

3)The news that Todd is missing.

There is a major underlying story in all of this.

While the Democratic brand has lost a 20 point advantage (see Pew article) going from a +22 to a +2, while Palin poll numbers plummet. While no other leaders' numbers are going up and while Obama's negatives on specific issues (health care and the economy) go down. His overall numbers have been stuck on cement.

Here is what the numbers show and what Chuck Todd doesn't get and doesn't report:

Despite volatility in the polls for other individual leaders, the party, or Obama's approval numbers on a particular issue, his overall numbers have shown a rather remarkable dose of stability. The reason is that people have decided how they feel about him and most people trust him and like him, even when they disagree with him. This is the big problem that Republicans face in 2012, even if they show improvement in other areas. Unless they come out with another candidate who is going to be more trustworthy and more likeable than the President they are in trouble.

[goclark]

1. Thank you grantcart

Your response is crystal clear to me.

[Cha]

2. Thank you, grant, for the open letter to Chuck Todd and what

is meant by his "cherry picking polls" for whatever reason.

[goclark]

5. Isn't "Lil Chuck the one that was crunching the numbers

for "Lil GW?

For some reason I have never had Todd as ~ one to trust.

He reminds me of a school boy that looks innocent and is doing bad things behind the teacher's back.

[cliffordu]

3. Excellent work, my good friend.

Sent a link to Chuckie just now....

[goclark]

4. K and K and K