Frazz by Jef Mallett for April 19, 2015

@Nabuquduriuzhur Oooh, I bite! Do you have a link, or at least remember what it was about and what the variables were?

Another problem recently has been perfectly ok studies with misleading reports being sold to the media, or worse, lately, someone put an incredibly misleading statement in their abstract and then acted all offended later and said “but of course we didn’t mean x, look, anyone can see the data!” Dude. Your abstract is meant to be a part of the actual science!

Or… sometimes the media just runs in circles on their own, the crossing fingers one made me want to cry.

But there are times when a small sample size et al can be an ok first step. If it does show something, it can help you get funding for a large version. If it doesn’t, it doesn’t really mean anything… but especially for some populations, where its hard to get a lot of the people in them together, a well designed study with a small number of data points is still better than just randomly writing case reports and hoping for the best. (Well, potentially.)

And sometimes you have no idea what a study is showing… there’s a recent one with a pretty damn good p value and a population of 2 million (which is about 50% of the total possible population if I recall.) But… but… why. A small part of me goes “even the best p value is sometimes chance.” Another part of me is just fascinated… its almost certainly a correlation. What on earth is underlying these two facts? But we need another 20 years and another 2-4 million people just to see… not even if it was true, just if it is continuing. Argh, health takes too long to figure out….

And actually both can be true. At the end of the day, coffee could be good for some people and bad for others. Or neutral, and coffee drinkers could just tend to do some other significant good or bad habits. I can’t stand the stuff, so I get to be amused from the sidelines this time.

Pschearer’s Laws of Studies:1) For every study there is an equal and opposite study.2) If Law 1 doesn’t hold, just wait.3) Every study must conclude with a call for more study.

Every high school curriculum should include a unit called “Causality vs. Correlation.” But then most of the perpetrators of “studies” would have to find more meaningful work.

I stopped worrying so much about so called “studies” when I heard on the radio that ORANGE JUICE caused cancer in female rats. Really? Orange Juice? sigh My thought is that anything will be bad for you if you use/eat it in excess.

Even aside from the whole discussion about studies, I find it refreshing that Mr. Spaetzel got to be the smart one, for once.

Gosh, I must be a scientist! But I doubt it.

any of these studies can be skewed to prove whatever the “researcher” wants by cherry picking the results

Sigh. I just knew Nab and Puma would have something negative to say about Mallet or Spaetzle while ignoring Frazz.

The latest studies shows that living is bad for you. The longer you live the more likely you will die.

One more of the many instances of Jef’s intelligence and wisdom.

Dow breast implants, the Alar apple scare (compete with ignorant celebrity testimony before Congress), etc.

Scientists relying on statistics to publish reports, when they don’t know how statistics work.