## It’s Just a Shell Game, Really

I learned to distrust statistics when I first took a class on it twenty-plus years ago. As an engineering student (then), I tried to boil it down to a simple question: will the bridge (or whatever) fall down? Stats, I thought, would give me more info on the probability of the bridge falling down or not. Boy, was I mistaken.

Turns out that statistics, as a mathematical discipline, is really all about appearances. Yes, you can figure out if the bridge will probably stay standing…but, if you want it to look MORE probable, use statistics method X; if you want it to look a little more shaky, use statistics method Y. I’m sure there are sound mathematical reasons for different methods, as well as different approaches, but it seemed suspect to me at the time and I’ve never shaken that attitude.

Political motivations behind statistics? Shocking!

Statistics, as presented as facts, definitely deserve the same level of distrust. For instance, take these statistics which I read in the newspaper yesterday:

There are 1.2 million cases of HIV/AIDS in the U.S. and 17,000 people die in a year, with 50,000 new cases in 2011; the government funds \$7.6 billion in research annually.
There are 11.96 million cases if cancer in the U.S. and 577,190 people die in a year, with 1.64 million new cases in 2011; the government funds \$4.9 billion* in research annually.

Now, nothing against AIDS research…it’s a nasty disease and a lot of progress has been made in the last thirty years. Being HIV-positive is no longer a guaranteed death sentence. Outside of AIDS itself, the whole science of anti-viral drugs has basically been created, resulting in benefits even for those without the disease. But, looking at those numbers, you might feel that there’s some sort of disparity; almost twice the funding given to cancer research, a disease suffered by ten times as many victims…with a death rate thirty times higher??? What logic is this? My tax dollars should be researching the cancer that killed my sainted grandmother, not some totally preventable [multiple pejoratives redacted] disease! I’m writing my Congressman! The President sucks!

Slow down there, hoss…you can’t just believe statistics as pure fact. Especially when they appear on a newspaper’s editorial page (hence the simulated politically-charged mock response above). This sort of outrage is exactly what the writer & editor are aiming for; after all, they showed you the numbers…and numbers can’t lie, right?

Well, yes and no. Did you notice the * next to the \$4.9 billion? Turns out, it leads you to the fine print which states “National Cancer Institute funding, which spends the most on research. Doesn’t represent other agencies.” Hmmm…so that’s only the money given to the NCI. So we assume that they spend the “most” (and that could mean anything from 51% to 99%…split the difference and call it 75%) so that means the government could fund about \$6.5 billion, in a conservative estimate. And that \$4.9b number isn’t even a lie…it’s just a statistic.

There are other factors to consider, too. Cancer isn’t like AIDS; one is a pretty specific result (immunosuppression) of infection by a certain type of virus while the other is a whole host of diseases attacking various and sundry parts of the body, with barely-understood causes ranging from genetics to environment. You could “logically” lump in any anti-viral research funding to pad that \$7.9 billion…but “conveniently” neglect to include associated research funding (artificial organs to replace cancer-ridden ones, for example) into cancer’s \$4.9 billion number.

And what about the private sector? Maybe the government funds so much AIDS research because not much money comes from elsewhere.  Celebrities are great at raising “awareness”…but I don’t see them putting billions into research. Meanwhile, what’re the numbers on private funding into cancer research? Doesn’t “Big Tobacco” have to pay billions per year as part of the settlement from years ago? I haven’t done the research…but my gut tells me there’s more to the story than these paltry few numbers, all based on the shadiness of that star’s fine print.

If they called textbooks what they really were…

That’s the thing, really: you can’t trust the numbers when people with an agenda, left or right, use statistics. The long-dormant engineer in me hates saying that but it’s true. People have learned that there’s power in numbers, so much power that they can use them without lying…they just have to pick the numbers they use carefully. The old joke is that 87.4% of all statistics are made up; the scary part is that they don’t have to be, they only have to be chosen for the maximum propaganda effect.