Monday, September 22, 2014

Analysis of Rasmussen Reports Before And After Scott Rasmussen's Departure

"Johnny Radio" AKA Hugh Hewitt
On twitter over the weekend, Hugh Hewitt, host of  the great, intelligent talk show-the Hugh Hewitt Show, and I got into a friendly disagreement with Tom Kludt, a reporter for Talking Points Memo, regarding the Rasmussen Reports polling. 

It has been my contention that the Rasmussen Reports web site has become such an extreme outlier in their polling, with regards to the Real Clear Politics average since Scott Rasmussen left the firm in July of 2013, that it is reaching illegitimacy status.  I have documented this in this post I started on Oct. 23, 2013 and have been updating until now: "What Is Going On With The Rasmussen Reports Polls?"

I have been mainly evaluating their [Rasmussen Reports] status on president Obama's approval numbers [and the spread of positive to negative approval numbers].  Hugh Hewitt pointed out in a tweet on Friday night of how their questionable numbers have crept into their polling of the generic congressional vote.  As of this writing, Sept. 20, The RCP average had the GOP ahead by a +3.5% margin.  But if you looked deeper into the numbers, of the 6 polls listed, 5 of the polls all had the GOP ahead by a 3-7 point margin.  The other poll listed, Rasmussen, had the Democrats ahead by 3 points.

Reporter Tom Kludt
Well, Tom Kludt tweeted to both of us that [paraphrasing], "What's new?"  When Scott Rasmussen was there in 2012 they were just as an extreme outlier the other way.  So, he surmises that they have always been an unreliable poll. 

I tweeted to Tom that I believed that in 2012 [the last full year that Scott Rasmussen was there] for the most part his polling firm was not an extreme outlier and for the most part they were in the mainstream of the other polls listed in RCP and that they were close to the Gallup numbers in 2012.   He countered that, "well both of those polls [Rasmussen and Gallup] were off in 2012 [on the presidential race]."  I didn't tweet this to Tom, but if I would have I would have told him the fact that Rasmussen didn't hit the election results as close as he did in 2010 did not make the polling firm an extreme outlier in relation to the other polls as it applied to the president's approval rating or as it applied to the generic congressional numbers. 

I told Tom that I would do a study of the numbers to see if Rasmussen was an extreme outlier in 2012, the last full year Scott Rasmussen was there.  I was curious myself, and if I would have found out that they were just as an extreme outlier when Scott was there, as they have been since Scott left the firm, I would admit he was right and give him Kudos.

Here is what I found out:

From January 3, 2012 [the first day of polling in that year] until November 6, 2012-when the presidential election was held, there were only 23 times when Scott Rasmussen had the president's approval rating under 46%.  That is 23 days out of 306 days.  The Real Clear Politics average of all polls during this same time period had the president's approval rating below 46%, 10 times.   In both Rasmussen and RCP, the days other when the polling wasn't below 46%, the range stayed between 46 and 50% in both polls--never below 46% and not one time above 50%.   So, that means except for 13 days out of 306, both the Rasmussen Reports poll and the RCP poll of polls almost mirrored each other, with Rasmussen having Obama's approval rating between 46-50% at a 92.5% clip and RCP at a 96.7% clip.  I think any fair person would have to say that means the last full year Scott Rasmussen was there in 2012, the Rasmussen Reports was in no way an extreme outlier, and in fact except for 13 days out of the year, they were not an outlier at all but right in the mainstream with all the other polls listed in RCP.  

Now let's compare that to what the numbers reveal for the first full year, 2014, after Scott Rasmussen's departure from Rasmussen Reports:

From Jan 3, 2014 [1st day of polling for the year in Rasmussen] until today Sept. 21, 2014 the Rasmussen Reports [sans Scott Rasmussen] has had the president's approval rate at or above 45%,  254 days out of the 262 days so far in the year.   That is 97% of the time.    The Real Clear Politics average of all the polls has had the president's approval at or above 45% during this same time period, 0 times.  As Hugh Hewitt would say, for the benefit of Steeler fans, that is 0% of the time..

Let me repeat that.  During the first full year since Scott Rasmussen left the Rasmussen Reports, the president's approval rate has been at or above 45% , 97% of the time compared to 0% of the time in Real Clear Politics.  

RCP average of presidential approval for 2014 [the black line]-as you see it never reaches  45% while Rasmussen during this exact same time period was between 45-52% approval for 97% of the timeNote:  The red line above represents the president's disapproval numbers.

Even more incredible, while Real Clear Politics for the whole year has ranged between 41% and 44+%, never reaching 45%, the Rasmussen Reports has actually had the presidents approval at or above 50% - 31 times [12% of the time].   

Let me repeat that. There is not one other poll in the universe that has had the president's approval rating at or above 50% this year [2014], while Rasmussen [sans Scott Rasmussen] has had them at or above 50%, 31 times.

So, basically for this whole year 2014, RCP has had the president's approval between 41 and 44% while the Rasmussen Reports Poll has had the president's approval between 45 and 52%.

NOTE: The RCP numbers includes Rasmussen in their average of the polls, so the president's approval rating in 2014 would even be lower in the RCP average of polls if Rasmussen wasn't included.
_______________________________________________________________
Conclusion The Rasmussen Reports during the last year when Scott Rasmussen was president of the Rasmussen Reports website, was for an overwhelmingly majority of the time, a main stream legitimate polling firm that was not an outlier with regards to the other polls listed in Real Clear Politics.

On the other hand [using a little of the president's often used phrase], the first full year [this year] after Scott Rasmussen left the web site,  the Rasmussen Reports has clearly become an extreme outlier with regards to all the other polls in Real Clear Politics.  They have become so extreme, in my opinion, that they can no longer be taken seriously as a credible polling firm. 

Just one example of dozens that shows it is hard to take the Rasmussen Reports seriously anymore since Scott Rasmussen left the polling firm to form his new digital media group, are these startling numbers from June 5, 2014:  The Real Clear Politics average of presidential approval for that date [and that even included Rasmussen which skewed their number] was 43.4%.  The disapproval on that date was 52.4% which meant there was a negative 9% spread.  Rasmussen on June 5, 2014 polling laughably showed that the president had an approval of 52% and a disapproval of 46% for a positive spread of 6%.  You would have thought the people who have taken over the Rasmussen Reports would have questioned their own numbers when, with all the other polls out there having the president's approval rating in the low 40's, they had the president at a higher approval number [and spread of his approval vs. disapproval number] than his re-election numbers.  How can anyone take a poll like that seriously.  And as I said, this wasn't a one time polling result that you could just discount as an anomaly.   I could have given dozens of examples like that, just this year. 

In my post "What Is Going On With The Rasmussen Reports" I speculated that [probably more out of hope] the new guys at the Rasmussen Reports made an adjustment in their polling methodology, because for a brief 2 month period of July and August of 2014, while they were still an outlier, they were not as an extreme outlier as they had been.  But that was short lived and they have gone back in September to their extreme unrealistic outlier status they have been for almost a year now.  

Scott Rasmussen-do you miss me now?

No comments: