Sunday, February 17, 2013

Ah, Data How Do I Know Thee?


Every news program you'll see in the evening will always feature a segment on research findings and what you should know that is either currently available or will be available at some time in the future. While this sounds very encouraging and, I agree, I love to have an optimistic outlook on research, how do we know that the person writing those news blurbs really has interpreted the data correctly? Well, maybe I should say adequately because correctly might be too strong a term.

In science and in research, there is really little in the way of "correctness" as I see it. It's in the way someone chooses to analyze the data and, as one very wise NYU professor told my class, there are many ways to look at data and not all of it may be useful. Recently, I heard of a clinical trial for Alzheimer's researcher, which indicated that a certain medication would be helpful. The operational word here is "helpful" and I'd look a bit closer.

Helpful actually meant that it might make some small difference, after the data were reanalyzed, for an even smaller sample. Now am I saying that this sample was specifically cherry picked in order to provide a more positive outlook on this research? Far be it from me to do that. I ask you to do that.

I can tell you that when I was working on a major, national research protocol of a medication, I found that the only difference the medication made was that people were more likely to engage in limited, social interactions. What social interactions am I referring to? They would say, "Thank you," or "Good morning." That wasn't a really big deal, but it looked to the researchers to be promising. Promising was really stretching the point and despite the fact that persons in the research were experiencing pretty awful side effects, it was a small price to pay. The product went to market.

So what do you need to know about interpreting research data? Can I give this to you right here and now? No, of course I can't, and what I would recommend is that, if you are so moved, that you take a wonderful free course on Coursera.org at http://bitly.com/YgMjrT (Making Sense of Statistics).

I know, I know. Statistics may not be a favorite of yours, but this course may just provide you with a greater understanding and appreciation of how people use statistics in either appropriate or inappropriate ways. I'm sure they'll tell you about sample size (very important) and about that all-important inclusion and exclusion criteria in research studies. Take it and you will be a more informed consumer. Not bad since all of us have to deal with healthcare and making determinations about what might be best for us or for those we love. It may even tickle you a bit to be able to speak the language of research, if only in a limited manner. It will make others begin to listen to you.

I'm a big fan of MOOCs (Massive Online Open Courses) because they provide us with incredible educational opportunities and they are FREE. So go take a look, find out what appeals to you, and begin the journey of learning again.