Eating things that say "hydrogenated" on the label may piss you off.

Peer Review

All GOOD science depends on independent verification and peer review. You need to share your data, so that other scientists in your field can examine it and, essentially, throw darts at it, to see how good the experiment was, how convincing the data is (given the size of the study), and so on. They need to suggest ways the experiment can be improved, to isolate other effects, or exclude other possibilities.

And yes, even if everything else checks out -- the data, etc. -- you still have to be open to the possibility that all you may have discovered is just correlation, and not a proof of cause and effect. That doesn't mean that you cannot prove cause and effect, just that the existing experimental evidence may not actually do that.

Correlation (by itself) does not prove cause. It still might be interesting and useful, but it's not the same as proving cause.

 

http://en.wikipedia.org/wiki/Cargo_cult_science

Also, here's the foundation:

http://en.wikipedia.org/wiki/Cochrane_Collaboration

 

As I've recommended in the books thread:
Surely You're Joking Mr. Feynman

But also there's this:
The Feynman Lectures on Physics This was required reading. It's a very qualitative approach to physics, easy to read, etc. It made you feel like you understood the subject. Unfortunately, then there would be the homework, and you'd realize just how little you really understood. These were not our primary textbooks (the Berkley series was one of them -- which was a lot more helpful when dealing with the math). But it is something that is good for a qualitative understanding of physics.

I curse the fact that I got to Caltech in 1978 -- in that year, Feynman taught Sophomore Physics. If only I were a year older... It wasn't his last time teaching undergrads, but I was already at least a Junior or a Senior when he did it again. He didn't have to teach an undergrad course, but he had a great rapport with the students and loved to do it.

Anyway, I DID get a chance to enjoy a dinner with Feynman (well, he was a guest at our student house one evening, so it was me and about 25-50(?) or so fellow students and Feynman). He regaled us with all sorts of stories in our library before the meal. Besides being a Nobel-winning Physicist, he also was an amateur artist, musician, lock-picker/safe-cracker, practical joker, fighter for the civil rights of brothels (lol-- no joke), and had a great sense of humor. He actually was in several Caltech productions, musical, artistic, and otherwise. He also was on the panel of people who looked into the Challenger disaster, helping to point out the problems that led to it.

When he was just starting out as a Physicist, he was put on the Manhattan project, and was one of many scientists working on building the bomb, where he first gained his reputation for being a practical joker, where he'd flaunt the rules in order to teach valuable lessons about the ridiculousness of some of the security and rules there (as well as the bad habits of scientists and their safes).

Anyway, one of the things that Feynman did later on in life, was to look at other scientific fields. He was appalled at the lack of understanding of statistics and math in general, in the fields of medicine, biology, social sciences, and so on. Basically, if you lack understanding of math, you are likely to look at your data and draw the wrong conclusions, and he pointed out a whole lot of bad science and wrong conclusions. I would not be surprised if a lot of what he saw was still going on today.

 

So it begins. I ran into this today without even looking for it.

http://www.cracked.com/article_20177_5-shocking-side-effects-foods-you-eat-every-day.html

The media takes these laughable studies and run with them. It's no wonder that so much of what we know is wrong. This study on hydrogenated oils and aggression has already been mentioned in articles around the world.

 

In other words, this was based on a survey, and there's no absolute statement that there ever was an experiment, blinded, double-blinded, or otherwise, where they fed different groups different kinds of fats. If there was an actual experiment done, there's no indication of how it was conducted.

Did they interview people over the phone, bring them in for physical exams, pass out questionnaires on campus, sent out e-mail spam, or what?

 

Cracked is where I ran into the story, but if you do a quick google search "fatty acids cause aggression" you'll see this study is being reported everywhere. I noticed it was reported on the huffington post and on the dailymail. Granted both of those are pretty garbage media sources.

​

Bad ideas propagate quickly.​

 

Now, hold on a minute. Just because the study might not be a sterling example of science doesn't mean the idea is bad -- as I've mentioned, there are precedents behind the idea that trans fatty acids screw up important things like neurotransmitter production. If you're going to be scientific about your science, you need to be equally frank and rational about how you treat your non-science.

 

True. But even so, it does mean that it still needs more work before it should be trusted as fact.

 

Agreed! I'm no fan of bad science myself -- I'm just weirded out by the violence of jhmmfn's reaction.

 

Imagine you're at work and a coworker tries to explain to you that mermaids are in fact real because he/she watched mermaids: the body found on animal planet.

Or how about this, your son comes home from school one day and explains to you that the earth was created 5000 years ago because his teacher said so.

I dunno, maybe it is just me. I did say its my pet peeve. Is there anything that just sets you off?

 

The former doesn't annoy me much (its something we technically can't disprove, but you'd need some very good evidence to believe - we have a very limited understanding of the oceans.)

The latter makes me want to stab people with knives. There's nothing worse than teaching something because I said so. The specific fact doesn't matter, but not teaching people how to think, and how to evaluate and analyze is one of the most awful things a teacher can do to a child. It's not the same as a questionably conducted study in any way shape or form.

 

I want to give Mining some knives. Stabbing idiots makes my arms hurt...

 

This is not quite the same thing -- there's a difference between questionable science and bad science. This falls more on the questionable side, mostly because we don't have the details. In other words, we should be skeptical, but still asking questions and not dismissing it out of hand.

True story -- you've heard of Millikan's Oil Drop experiment, where he measured the charge of the electron. Well it turns out that for a long time after that, no one could actually get the same results as Millikan did. Everybody was getting different values for the charge of the electron. What was unusual was that if you looked at the measurment over time, the next time the experiment was done, the value was just a little bit bigger than what Millikan got. And the time after that, it was just a little bit bigger. And so on. Eventually the values converged on what we now accept as the true value.

The problem wasn't bad science. The problem was an error in the methodology that Millikan used, and scientists unwillingness to accept that Millikan actually made a mistake -- he didn't take the viscosity of air into account. What was happening though was that each experimenter, when he didn't quite get Millikan's results, they looked into their data and looked specifically for reasons why their numbers were too big, and they shaved the data (for mistaken reasons). It wasn't until much later that anyone discovered the error in Millikan's methodology. Everyone else was doing it right but they were fudging the numbers each time.

That's faulty science not bad science, in the way that you are using the term.

We don't know that we have bad science here, so we need to continue to be skeptical, but that doesn't mean throwing it out out of hand.

 

Millikan's methodology is something I wish literally everyone knew. It's the coolest experiment ever in a lot of way - you're describing something that is ridiculously small with observations you're making with your eyes! It's one of the only quantum effects that is trivially easy to observe, too.

 

I don't really see the distinction between bad and questionable science.

While we can't look at the study and say that the claim is false, We can look at the study and see it provides no evidence that the claim is true.

I don't see a whole lot of difference between making wild claims and making false claims. At least not enough difference to say that the science is not bad.

 

bad science = complete distortion of data set, unwillingness to revise conclusions based on new data, and no conversion on a 'correct' value. questionable science = selection out of data set according to described criterion that you have published and an open method that is reproducible, coupled with conversion on a true value.

 

Your definition here doesn't specify criteria for being questionable.

 

Questionable is relative. You set the conditions for the situation. Trying to have absolute rules for what is questionable is plainly counterproductive.

Some say we evolved as can be demonstrated by such things as people getting taller over generations. Some people say that is questionable and spew their belief of choice as undeniable proof.

I do not care who is right. But questionable is relative.
(Not trying to start an evolution verses anything else argument, just pointing out the first thing I can think of that such rules will run right into.)