The “March For Science” would not bother me so much if it weren’t for bullshit like this.

by w3woody

“Mindless Eating,” or how to send an entire life of research into question

If you are truly a supporter of science, the following things would worry you:

P-hacking and trolling datasets for correlations. If–as was done by Dr. Wansink–scientists start trolling datasets for correlations in order to “salvage research”, you get crap correlations. (And there are a lot of spurious correlations which have no meaning whatsoever.)

Statistical abuse, and the fact that a lot of scientific research is being done by people who have no clue how to properly use statistics to find correlations.

The replication crisis: remember, if a knowledgeable individual with experience in your field cannot replicate the results of your paper–either because you do not fully publish the data set (or at least refer to the data set), publish the tools you use to process the data (and a lot of research uses custom software to crunch data), or publish the methodology you use (such as the apparatus setup)–then you have not published a scientific paper. What you’ve published is a fucking press release.

Confirmation bias, which leads people to accept results because they sound right rather than results that were actually proven by replicatable and carefully recorded observation. In this case, we have a researcher using bad techniques to draw conclusions about nutrition which have driven policy and marketing–including the marketing of 100 calorie snack packs and the creation of food programs designed to “nudge” children into making better nutrition choices–implemented because they align our preconceptions and the profitability of food corporations.

(Take 100 calorie snack packs: we accept paying more per unit of food and significantly increasing the amount of waste because we think this will help people lose weight. But what if that isn’t the case? What if people don’t actually lose weight? All we’ve done is increased the cost of packaged foods and increased the amount of waste going to landfills.)


Of course most of the people marching in last week’s “March for Science” probably couldn’t solve a fucking quadratic equation much less discuss coherently the problems with the replication crisis, the proper use of statistics, or how to present a scientific paper in a way which allows practitioners in the field to replicate the results. They probably don’t understand the need for testing and publishing null hypothesis or the need to properly trace data back to its source. (Fields such as nutrition are especially vulnerable because they often rely on participants accurately recalling what they ate in the past 24 hours–a method which is fraught with significant errors. Yet results in the field of nutrition are used to drive public policy affecting all of our lives.)

If those marchers want to do something to help science, they may want to look at shit like this, rather than smugly watching Bill Nye saving the world by having Rachel Bloom sing about her vagina.

Seriously.

Advertisements