Medpedia

Ah, the graphs and charts and "infographics" of today. They're pretty, aren't they? They give us through sight what the text cannot relay. In one cursory look, we get to learn about stuff. As an epidemiologist, I have to create a lot of graphs and charts for presentations and reports. For example, we can learn a lot from a graph of onset of disease during an outbreak:

It helps to formulate hypotheses

In matter of Public Health, however, there are others who will use graphs to convince you of things that are not all that true. For example, what if i was to tell you that Alcoholics Anonymous increased the birth rate in the United States from an all time low in 1935? Would you believe me?

What if I told you that the launch of Sputnik 1 lowered the morale of the United States so much that it reduced the birth rate starting in 1957? Is it plausible?

Suppose I then told you that the birth rate was on it's way down until the Pope stepped in and put an end to the insanity by banning the use of the contraceptive pill among Catholics. Would you believe me?

And what if I then added to that the creation of the US Environmental Protection Agency and the celebration of the first Earthy Day really bummed us out and lowered the birth rate again? What evidence could I possibly have for that?

How about this:

I don't need no stikin' evidence!

Those are the real rates of birth for the years listed, according to the US Department of Health and Human Services. Those are real events that happened on those years. There's a correlation, maybe even an association. So why won't you believe me?

Because you know better, that's why. I first graphed the data and then looked for events in those years where a shift happened. I then labeled those shifts with the events that I found. It's a backwards way of doing this. It's the Texas Sharpshooter Fallacy.

Unfortunately, many readers of an anti-vaccine site don't. That site describes itself thus:

The Thinktwice Global Vaccine Institute was established in 1996 to provide parents and other concerned people with educational resources enabling them to make more informed vaccine decisions. Thinktwice encourages an uncensored exchange of vaccine information, and supports every family's right to accept or reject vaccines.

The usual disclaimer then follows.

Pretty noble, right? I mean, they want you to be educated on vaccines. They want an "uncensored exchange" of information. Nothing wrong with that, correct? Well, the problem comes when they publish "information" like this PDF article on Polio. In it, you will find these two graphs:

Why, oh, why did we spend so much time and effort on a polio vaccine if death rates were on the decline already? (Some have even suggested that polio was going to die out on its own by 1990.) Because these graphs have been manipulated, that's why.

In the first case, the number of deaths was declining because of advances in the treatment of polio, including the iron lung and other ways of keeping paralyzed children alive. (The PDF article will say that the paralysis is reversible and that children recover completely from it. Bull-effing-shit. Pardon my French.) In the second case, notice that the number of cases decreased AFTER THE VACCINES WERE INTRODUCED IN THE 50's, but the author decided to label the graph to their satisfaction and make it look like it was the change in case definition.

Like in the graph I presented above, these two polio graphs take actual, real data and layer labels on it to make it look like the vaccines did nothing to reduce the burden of polio on the population. Imagine if the second graph had the first graph's markings? That would totally throw out the idea that the polio vaccines were worthless.

So, please, take a hard look at the graphs and charts that you are presented. Look for small hints that the author(s) might have hidden motives to show you those graphs. Be skeptical about them and look for other sources of data, especially official sources. Do not be deceived by their prestidigitation.

Finally, why should you believe me when I put up a graph or write something? Because I stake my entire professional career as an epidemiologist on this. I will not lie to you. I will back up what I say or write with evidence. If I make a mistake - which I have done in the past - I will correct it. I won't brush you off, call you names, or suggest that you should be fired from your job for putting me right.