mgroves

Posts Tagged ‘statistics’

Recently, Zogby and the Norman Lear center did a survey to find out the entertainment tastes of conservatives and liberals.

First off, note that Norman Lear is an outspoken liberal and friend of Meathead, but also that Zogby is a credible source of statistics and polling.

Let’s go over the concept of correlation and average, briefly. As you should know, correlation is not causation, it’s simply the statistical strength of a relationship. So for instance, just because 85% of people who drink beer are healthy, doesn’t mean beer makes you healthy. It could mean that being healthy makes you drink beer, or that some other factor external to the data causes both.

The data for the survey is attached. The sample sizes were statistically pretty good (n around 1000 for each political grouping). However, there are no confidence intervals or anything like that (I’m sure they were calculated, but they aren’t in the data), so there’s no way of knowing how accurate these point estimates are.

With that quick preface, here are some interesting points from the survey:

• Dr. Gregory House apparently unifies us all with his cynical cruelty. Conservatives and liberals all enjoy his show.
• Of all the movies surveyed about, The Departed was the only apolitical/irreligious movie that was preferred by one group (liberals) — I guess because it’s R-rated? C’mon adult conservatives, this is a good movie!
• Conservatives don’t seem to play a lot of video games, what with being employed and all.
• Out of 15 TV and film genres, “arts” emerged as the one with the highest positive correlation to liberal viewers, “arts” being a codeword for “gay”.

(Those last two are jokes, people, jokes!)

And here are some interesting things that weren’t talked about in the article, but were hiding in the data:

• People polled were all likely voters. So there are hardly any non-voters. I think this is a potential bias, as a non-voter isn’t necessarily apolitical in their opinions.
• Liberals surveyed believed that government generally solved problems (69.7%). I’m going to assume they think that liberal government generally solved problems, not government in general. Whereas conservatives (89.5%) think that government doesn’t solve problems (it doesn’t matter WHO is in there). This is strange given that the data also suggests that 80% or more of all surveyed (including liberals) were cynical about the motives of elected officials. So what’s up liberals, something doesn’t jive here.
• Nader got most of his survey votes from so-called “moderates”. Interesting.
• Liberals were observant enough to say they lived on “planet Earth” but were unable to specify any further.
• The “global economy” issue was close to split 50/50 among liberals. I’m guessing 50% are in “dey took our jerbs” camp, and 50% are in the “every other country is superior” camp?

I’ve stated in the past that the state of MLB is excellent. Excellent parity, fairness, and competitiveness. Nothing can be perfect, and I’m sure many would disagree with me.

However, as I’ve said in the past, if I were the supreme dictator of MLB, there are a few things I would do (that some might consider drastic) to improve MLB, increase parity, competitiveness, and fan interest.

Here’s the list:

• Raise the mound back to pre-Gibson days (from 10 inches now up to 13-15 inches).
• Remove the designated hitter rule.
• Stop interleague play.
• Shorten the season.
• Roll back expansion: prune teams that have a record of poor performance and low fan interest.

The basis of all of these recommendations stems from my belief that there is not enough quality pitching to go around, diluting the talent pool, and making it hard for mid- to small-market teams to compete for the very limited resource of quality pitching.

So, instead of taking the socialistic approach of salary caps and luxury taxes, I think it’s better to take a market-based approach and make a scarce commodity less scarce.

But I want to focus this post on the last item on the list: pruning of entire teams. I believe that this will make it easier for the remaining teams to compile a decent pitching staff, but forget that for a second. The more important question is: which teams? How about these:

• Kansas City Royals
• Tampa Bay Devil Rays
• Florida Marlins
• Baltimore Orioles

How did I pick these? First, I decided that attendance was a good measure of fan interest. But not all parks have the same capacity, so percentage of total seats sold should work okay. Second, I used “wins” to determine the performance of a team. More wins = better performance. Finally, I added those numbers together and sorted the teams. (See attached Excel sheet for details).

I only did this for 2006–it would be more useful to do this for, say, the last 10 years, and pick the 4 teams who appeared on the bottom 4 the most.

Or maybe both Florida teams are “combined” into one team that plays in, say, Orlando. And then Colorado (next on the list) can be pruned instead.

I came across Shattering the Bell Curve while reading one of my favorite economics blogs, EconLog.

When talking about assumptions for predictive statistical models, David A. Shaywitz suggests that there is a lot inconsistency and inaccuracy in many statistical models, due to incorrect assumptions about distributions. Specifically, using a bell-curve instead of a power-law.

Arnold Kling, from EconLog, applies this specifically to models proclaiming various scenarios of significant man-made climate change, of which it certainly applies, but that’s not what I found the most interesting.

I found this quote from the conclusion of the article interesting:

If we accept Mr. Taleb’s premise about power-law ascendancy, we are left with a troubling question: How do you function in a world where accurate prediction is rarely possible, where history isn’t a reliable guide to the future and where the most important events cannot be anticipated? [emphasis mine]

I don’t find it troubling that I can’t perfectly anticipate or predict the future. Well, okay I’m lying. I’m scared of the future just like everyone: it’s the human condition. But to somehow think that faith in a predictive statistical model was somehow holding back the now-shattered dam separating fear and functionality seems ridiculous.

How can you function in a world that can’t be statistically predicted? We all do it every day, all the time. I think rather than fearing the future, we should embrace it and look forward to it.