One of my most frustrating pet peeves is when scientists don’t understand what it is they are measuring.

And, let me assure you, this is a much bigger problem than anyone is willing to admit.

My background – going way, way back, before my ventures into the business world or even into parenthood, I studied Science.  And, while I never sought a doctorate or any such thing (I had done my due diligence on child-bearing statistics in preparation for parenthood and realized that if I wanted to optimize for my children’s intelligence, I had to conceive my first child at no older an age than 25 – and my last one at no older an age than 30:  and since my then fiance – now husband – agreed that we did not approve of the ‘daycare’ model of child-rearing, somewhat to my now hubbie’s chagrin, I chose not to pursue further studies), I do have a degree in Physics in there somewhere….

What I specialized for (though I did not realize at the time that this was ‘soooooo Aspie’) was data acquisition, test and measurement.  I made a career out of helping other scientists (and industry, military etc.) figure out how to measure what it was they were really trying to measure, from designing the data acquisition systems to telling them if they were actually measuring what they thought they were measuring.

As such,  am somewhat sensitive to ‘sloppy science’.

Which is why I so happy that my son has forwarded me a link to an absolutely excellent essay about how statistics – especially in the medical field -(where, when I was finishing my degree, I was heavily lobbied to go into post-grad, so that I could ‘clean-up’ the methodology in a prominent Canadian immunology University lab – so I really, really understand the criticism here…) are misunderstood not just by the public, not just by the media people who are reporting on it, but especially by the scientists themselves who are carrying out the studies/experiments!

‘Open a random page in your favorite medical journal and you’ll soon be deluged with statistics: t tests, p values, proportional hazards models, risk ratios, logistic regressions, least-squares fits, and confidence intervals. Statisticians have provided scientists with tools of enormous power to find order and meaning in the most complex of datasets, and scientists have embraced them with glee.

Many of these tools are misapplied or misinterpreted.

In fact, most published research findings are probably false.’

Aye, aye.

The essay is written with the layman in mind:  it explains things, from first principles, without jargon but with examples of just how easy it is to manipulate results, even without realizing one is doing so.

IF you are interested in science…

IF you have not taken a lot of courses in statistics – but want to understand the real-life meaning of statistics…

IF you want to keep ‘science honest’ ….

IF you question ‘politicized science’…

…you would benefit from/enjoy reading this simple essay.

H/T:  Tyr