Last week’s Republican Presidential debate and the entrance of Fred Thompson into the race seems to have reshaped the landscape for the nomination, or at least that’s what the latest CBS/New York Times polling data suggests. I don’t think that the hard liberal bend on both of these entities plays into these particular numbers, so no feedback about that please. Additionally, those of you who know me know that I am 110% against these “scientific” polls because the truth is that there’s nothing scientific about them. That’s a handy little diddy I picked up in graduate school…of course the professor disagreed, but he had his ideology to push, too!
In any event, the news reports about McCain having a good debate are right on the money. The CBS poll shows him up a few points to 18% of Republican primary voters. This may seem like a small drop in the bucket compared to the front-runners, but for McCain, this is big news. He almost doubled his support by going back to what brought him to the Presidential dance in 2000 – straight talk about every issue. It’s a nice change from some of the blabbering that you can hear in primary debates.
On the other hand, Romney came in at just 14% support, which is not par for this course. Usually, Romney gets a few more percentage points on his side after each debate because he has what some people call the “Presidential look,” whatever that means.
The biggest news in the CBS poll is Thompson jumping to 22% of the primary voters supporting his bid for the Presidency. This is huge because Giuliani is only at 27% in this poll. The difference between these two numbers will likely fall in the “scientific” margin of error (don’t even get me started). Huckabee doesn’t register on the CBS poll, which is a shame because if you look at RealClearPolitics.com’s averages over time you see that he has enjoyed a burst of support in the last few months going from 0% to 4.5%.
That’s huge for Huckabee, though it won’t get him the nomination. At this point, it’s way too early to call a clear winner on the Republican side of race. For example, the Rasmussen poll shows Thompson is actually in the lead in this race.
LOL what’s your problem with statistics?
I had a professor once who was a great guy and who I wish nothing but the best for, but he actually manipulated data to arrive at the end result that he wanted to. For example, he went through all of the regression analysis and then wound up with an end result that didn’t quite fit what he wanted it to.
Bear in mind that this is in front of the class…
So he went back in the data set and included just enough extra variables so that the data reached the percentages that he was looking for. I went ape shit. I started screaming and yelling that it was intellectually dishonest to state a hypothesis and a null hypothesis and to manipulate data to prove the hypothesis over the null hypothesis. He actually found my rant amusing and said something to the effect, “How do you think the numbers are calculated in a made-for-TV world?”
I got mad…and still am!
Well, that’s not statistics – that guy probably majored in Creative Writing. I took Stats on the upper-division undergrad level since I need it for my Math major, and if my professor heard that story, she’d go ape shit just like you did! That’s just ridiculous what he did!
As far as his statement about “How do you think the numbers are calculated in a made-for-TV world,” I can’t argue that because maybe some stations (*cough* FOX News, *cough* MSNBC) do tweaking to make certain things look a certain way. In fact one of the things I learned in statistics class is that you should only trust real polls, people like Gallup and such since they provide a full report of their testing (variables used, possible choices if polling, etc), whereas TV likes to focus on a Q & A type of approach, and obviously that’s “more sexy” as Nash would say to the common people.
It was statistics enough for this guy to be a well-respected doctor teaching graduate and doctoral students. I should mention that the other statistics professors I took at Rutgers were all in agreement on it being “okay” to manipulate base assumptions to reach your hypothesis.
They saw it as finding a means to an end – which I vehemently argued against. My position was yes, you’ve found a way to prove the hypothesis. However, you’ve included unrelated variables to manipulate the regression analysis to the point where instead of saying (for example), “Young, unwed mothers are likely to have underperforming school-aged children,” you now have to say…
“Young, unwed mothers of color who have a familial history of drug-abuse living within a census tract that is 60% of area median income in an area that has been traditionally red-lined by mainstream banking and where local schools have been cited as failures according to No Child Left Behind are more likely to have male children aged between 2 and 6 who have reading deficiencies through 1st Grade, but no problems with math so long as their teacher has attained at least a Masters level degree from an Ivy League school and is aged between 22 and 45.”
The Professor didn’t like that – none of the Rutgers professors liked that. In fact, one of them was so much against it that when he applied for another job at an institution where I have some small degree of sway, I lobbied heavily against him being hired.