An example of how tough it is to study humans

(Last Updated On: March 27, 2016)

As part of the aerospace engineering and human spaceflight course I recently took, we were looking at disturbances in human equilibrium after spaceflight. Funky things happen to people in space due to microgravity. One of those things is a decreased ability to orient yourself when coming back into earth’s gravity. You can incur some odd sensations in your head, feel like falling over, etc.

NASA has studied this by using a phone booth-like machine (red circle below) to assess people’s ability to essentially balance themselves. Below, time = 0 is when astronauts have come back into earth’s gravity after being on the space shuttle, which was in microgravity. You can see before this time, most score right around 100%.

Immediately after landing though, the astronauts are worse. Then they acclimate fairly quickly. The blue trend line shows people coming back to baseline:

“Don’t give me the headline; give me the trend line.” But the trend line doesn’t always say how you’re doing. Look at everyone before the 30 minute mark:

Everyone did worse the first 30 minutes after landing…except one did better!

Not only that, another person scored way worse, all the way down at 75%. One person scores 110%, another 75%. And a bunch of others are all over the place. Before they left though the difference was, at most, 10%. That’s a huge variance between two people. 30 minutes after landing on Mars one person may not even be able to walk, while another is as balanced as they’ve ever been!

As the professor teaching this lecture mentioned, this is a great example of how hard medical research is. I’d broaden that and say it’s an example of how hard human research is, and how hard it is to predict human responses. For instance, having a history of motion sickness on earth does not correlate to having motion sickness -space sickness- in microgravity.

If you ever wanted a clear indication of our inability to predict human responses, look at the stock market. The thing has, in some shape or form, been around hundreds of years. So the sample size is enormous. Not only that, the incentive to create a predictive model is as big a financial incentive that’s ever existed. It still can’t be done. Never mind a person trying to predict it, no amount of people have been able to create a computer which can predict it. Sure, some good picks can happen, but it’s not as if the power of F=ma is going on. Warren Buffett just lost ~450 million dollars on Tesco. Yes, that’s not much money to him, but the point remains: He’s not predicting in the strictest sense; he’s estimating. And he certainly didn’t envision Tesco lying about their quarterly earnings to the tune of $400 million.

You may know how most people do on something, but you’ll probably never know how everyone does, until you test them. “Give me the headline; the trend line; and all the data.”

 

–