My apologies for the long post but, holding public figures accountable for their misuse of statistics is important. This post isn’t meant to be partisan, only to critically think about statistics that are being used by politicians. Any examples of politicians of any kind misusing statistics like this are welcome in the comments and may be fodder for future posts.

Recently, the Obama Administration has been publicizing a graph by Nancy Pelosi’s team comparing the last year of the Bush Administration with the first year of the Obama Administration.

It’s an interesting graph and it’s meant to show that Bush Administration (and Republican) policies were losing the nation jobs while Obama Administration (and Democrat) policies are bringing those jobs back. At first blush, it’s pretty damning.

However, there are a number of problems with a comparison like this: Continue Reading

“Those states that do consume the most porn tend to be more conservative and religious than states with lower levels of consumption…”

‘Some of the people who are most outraged turn out to be consumers of the very things they claimed to be outraged by,’ Edelman says.”

Conservatives and religious people consuming more porn than others?  That would make them hypocrites!  Let’s see how they arrive at this conclusion.

“Edelman spends part of his time helping companies such as Microsoft and AOL detect advertising fraud. Another consulting client runs dozens of adult websites, though he says he is not at liberty to identify the firm.

That company did, however, provide Edelman with roughly two years of credit card data from 2006 to 2008 that included a purchase date and each customer’s postal code.”

Great, we can see the number of porn subscriptions in a given zip code.  While this may be interesting, it does nothing to link porn subscriptions to religious beliefs and/or activity or political leanings.  It may be that the liberal elements of the zip code are the ones purchasing the porn.

“The biggest consumer, Utah, averaged 5.47 adult content subscriptions per 1000 home broadband users”

Again, the problem with this result is that it does nothing to link behavior to an actual person.  All we know is that in towns that vote conservative and have religious beliefs, more porn is purchased than in other towns.

The article goes on to quote other statistics at the state level that fail to link porn use to individual religious and political beliefs.

In other words, this data tells us everything except the conclusion the study’s authors have drawn.

It would be more useful if they reported results at the zip code level and linked that to voting districts.  At least then, we would have a closer link to an individual than an entire state.

I thought scientists were unbiased and believed in the integrity of their research.  Who’s the hypocrite again?

The Seattle Times has an article entitled “Mistakes hospitals don’t want you to see“. In it, Carol Ostrom cites some statistics that are supposed to sound terrifying:

“Over the past year, hospitals in Washington left ‘foreign objects’ in 36 surgery patients. And 21 people got surgery on the wrong body parts.”

I didn’t wet my pants in fear at those numbers.  In fact, I thought they were kind of low.

So, I checked to see how many surgical procedures are done in this state every year. The best I could find was the number of surgeries done nationwide by privately-funded in hospitals in 2003. That number is 7.1 million, making Washington’s share 142,000, give or take an spleen or a colon.

Let’s restate Ms. Ostrom’s sentence with meaningful statistics instead of biased, agenda-based ones:

“Over the past year, hospitals in Washington left ‘foreign objects’ in .025% of surgery patients. And .015% people got surgery on the wrong body parts.”

.025% and .015% failure rates? Toyota would happily rip it’s own arms off for that kind of a failure rate.

Come on Seattle Times and Ms. Ostrom.  I know misunderstanding statistics is de rigueur in the news business these days, but I thought Seattle was supposed to be counter-culture.