Jump to content

Beating a dead horse, again... and again...


Recommended Posts

At Kits prompting, I tried to look at this article, which appeared a few days ago (the reference is Ann Intern Med. 2014;160(6):398-406-406. doi:10.7326/M13-1788 if anyone is interested). My institution does not have online access, so I would have to wait till we get the print version by post, or pay many $$$ - but I am not going to bother.

From what I can tell from the abstract and other public information:

1. The authors look respectable and come from respectable institutions, so I would tend to trust them.

2. It is a meta-study - putting together the conclusions from many other published studies (which usually have to meet certain criteria). This tends to give a bias, because people tend to only publish positive results - but the result of this study overall is negative, so that bias was clearly not dominant. In other words, a bit doubtful, but not highly so.

3. Because of not having access to the article, I cant see how well the dietary information was controlled in the different studies - it is quite difficult to measure accurately what people eat (they tend to look on the bright side and have selective memories when filling in questionnaires).

4. So overall, I'll park this in the "interesting" category and see what further research appears in future.

5. Fat easily leads to weight gain, because it is so energy-dense. However, it is slow to digest and gives a good feeling of satiety, tending to suppress appetite, so some fat seems good as part of a weight-control diet. Also it is useful for fat-soluble nutrients. For myself, I dont avoid fat completely, but only eat small amounts of high-fat food. A middle-of-the-road approach when things are uncertain. My cholesterol levels have dropped a little over the last year, when I've been eating a bit more fat, and reducing my sugar intake.

Link to comment
Share on other sites

Here's another analysis of the paper: http://wholehealthsource.blogspot.it/2014/03/new-review-paper-on-dietary-fat-and.html

I find that Dr Guyenet is always very moderate and considered in his assertion and isn't afraid to modify his view when new information arises, so I respect what he has to say.

I don't think this study is definitive, but personally I view it as another nail in the coffin of the hypothesis connecting cholesterol, saturated fat and heart disease.

What worries me more now is that we as a society can't help but go to extremes. We need to find the villain within large and complex issues. What we see now is the pendulum swinging back the other way to "carbs are evil". Historically, we seem obsessed with finding the "evil" macro-nutrient.

This tends to give a bias, because people tend to only publish positive results

Jim, have you read this? http://www.economist.com/news/briefing/21588057-scientists-think-science-self-correcting-alarming-degree-it-not-trouble

If true, the implications are a bit terrifying.

Link to comment
Share on other sites

Kit - I did read the Economist article quickly - mainly to get the gist, not all the detail.

One problem is that non-scientists think that science is absolute, true and accurate. Scientists realise that it is more of a zig-zag stagger towards the truth. Rarely does it have to backtrack completely, more usually is it a change of direction.

Yes, I think that science is by its nature self-correcting. How can I think that if results cant always be repeated, and if they are not repeatable, maybe because "performing an experiment always entails what sociologists call “tacit knowledge”—craft skills and extemporisations that their possessors take for granted but can pass on only through example. Thus if a replication fails, it could be because the repeaters didn’t quite get these je-ne-sais-quoi bits of the protocol right"?

Why do I think that science is usually self-correcting? Because if a result is false, it cannot be built on. Scientists clearly like to build on a useful result, and develop theory and understanding further. If a result is false, it leads no-where. It may sound great, ring some bells, and maybe some further successful experiments will follow, but if it is wrong, soon it will lead to conclusions which cant be incorporated into the body of knowledge*.

We also have different sorts of false results. 1. Those stemming from false or non-useful theoretical frameworks. Lots of results get published, maybe are correct in their own terms, but are all thrown away decades later because the theoretical framework was not useful (lots of older psychology, etc).

2. Errors of statistics as discussed in the Economist article: Frankly, if you have to do lots of statistics, its a sign you have a weak effect, and you should be suspicious. We always say, first use the i-test (look at the data). If you cant see the result by just looking at the data, its probably not there. Of course you have to do proper statistics later on. Another big error is due to misunderstanding conditional probability - treating events as independent when they are not. This involves Bayesian statistics to do properly and is an immensely difficult area for non-specialists. Of course, people want to do powerful advanced statistics on a data set to get the max out of it. This is done in areas such as molecular bioinformatics (genetics, where you may be looking for one base change out of the billions in the whole genome) and demographics. Because of the complexity of the statistics, it is easy for subtle errors to slip through.

3. Finally, refereeing. Referees have to take a lot on trust. To really check, they would have to repeat the work themselves, which may well take them even longer than the original researchers because they would have to get all the techniques up and running. One example – I published a paper with lots of maths and results of running mathematical models. The main referee (who I respected a great deal) did not repeat the calculations – he just looked to see if the results sort-of looked right, and to see if they were in agreement with his own calculations on a related problem some time before. So something wrong could well have slipped through. I would have done the same. In general, referees look and see if the work sort-of looks right, and whether it has made any of the obvious errors (or course, knowing what errors would be seen as obvious, a dishonest author could take care to hide them).

Overall, therefore, science will sort-of get it right eventually, but may take a long time to get there, and may look up lots of blind alleys on the way. The Economist article was implying that a scientific result somehow is or should be ABSOLUTELY correct. Real scientists know that it is just a starting point for further experiment.

JIm.

*(one of the criteria for knowledge, cant remember the philosophical term now, and am not going to leave this webpage to look it up - one of the extensions of the JTB theory of knowledge which overall fits with scientific knowledge)

Link to comment
Share on other sites

By the way, its not a dead horse, its very much a live one, which is why we keep flogging it. Its just that we dont know which way its going to jump.

"What we see now is the pendulum swinging back the other way to "carbs are evil". Historically, we seem obsessed with finding the "evil" macro-nutrient."

Yes, part of the problem seems to be that people want simple solutions - they might think that there are a few evil ingredients which can be blamed for everything, and once they avoid those, they will have perfect health. That is rarely the case - its more usually that there are many factors each of which makes a small contribution. And some which are healthy in some ways are unhealthy in others.

There are however some evil factors or ingredients. Smoking. Excess alcohol. As we now know, processed meat. Many chemical pollutants.

Jim.

Link to comment
Share on other sites

http://mobile.nytimes.com/2014/03/26/opinion/bittman-butter-is-back.html?from=opinion&post_id=1537447993_10203760591765418

Sent to be by Eric R; I don't have time to read the appended articles today, but will soon (off to Adelaide tomorrow morning for ten days).

Link to comment
Share on other sites

  • 3 weeks later...
  • 1 year later...

DW, I didn't see your addenda, but yes, please, set the time and date (and with the worms)!

Link to comment
Share on other sites

I'll bring the dead horse! The prefecture next to where I live now is known for it's (raw) horse sashimi. Did you ever happen to eat any while here, Kit?

 

The prefecture next to that one is known for its (raw) chicken sashimi, which I'd have to say I prefer over the horse both taste and texture-wise, though.

Link to comment
Share on other sites

Jim, re-reading the Economist article and your excellent commentary; but there's more: consider how simply trying to secure funding takes full time work by at least one team member (out of a team of five, according to scientist friends at the ANU, and it can be higher); it's not that waste of effort and time I am talking about, but how research proposals themselves are influenced and positioned by this competition. "Accentuate the positive", indeed. Add to that the fact that PhD candidates find it hard to get the kind of result that leads to the successful conclusion of a PhD in the same atmosphere; IIRC, the rules for the awarding of a PhD still are that the candidate's work needs to have "made a significant contribution to the field of inquiry"; this is made much harder if the results of his/her research are negative (even though the negatives in research must outweigh the positives by an order of magnitude, and often are more important, depending on previous research in the field). 

 

@Nathan: no, never horse meat nor chicken, either, but fish of astounding variety!

  • Like 1
Link to comment
Share on other sites

Kit (just in a coffee break at the contortion convention) - I think scientists are well aware of the distortions you talk about (I've not gone back to the Economist article by the way), and take them into account while trying to play the game. E.g. they will have some up-to-the minute fashionable grant-grabbing programs running, while at the same time under the radar be running some long term exploratory and very original programs. When the latter come up trumps, they will be pushed to the forefront. If a PhD students research does not produce the anticipated forefront breakthroughs expected, a quick collaboration will be set up with someone whose research does. This flexibility and freedom only exists in very large groups, that also have some core funding from a foundation that gives a degree of freedom, which is why those groups tend to be so productive in the long term. I am talking about biomedical science here, which I know most about.

 

"the rules for the awarding of a PhD still are that the candidate's work needs to have "made a significant contribution to the field of inquiry"" - theoretically - but in fact, the word significant is interpreted with an enormous degree of flexibility, to include "worked reasonably hard for 3-5 years and produced a thesis in approximately recognisable English at the end of it". A university is not going to refuse an award and the large amount of funding that comes with a completed degree, on something so basic as principles, except in extreme cases.

 

You should be at this meeting - I think you would have a whale of a time (do you say that in Oz-speak)?

 

Jim.

  • Like 1
Link to comment
Share on other sites

Jim, I would love to be there. Are you a 'Life of Brian' fan? "He 'ad an 'ell of a time!"

 

And what you write about (WRT biomedical research and how it says it operates and what really happens) is exactly what I remember from the ANU. The terminally curious can't be stopped, fortunately. And your remarks re. modern PhD research is sad and accurate in my experience. We should talk about this over a glass of wine some time soon.

Link to comment
Share on other sites

Horse is very tasty... somewhere between beef and 'roo IMHO.

 

This http://www.amazon.com/Death-Food-Pyramid-Politics-Interests/dp/0984755128 book gives a very interesting review of the history and politics of mainstream official dietary recommendations (eg USDA) and - in what seems to a layman - a fairly balanced review of the (de)merits of different dietary trends.

 

TL;DR is that the modern food pyramid was based on feeding people at the cheapest kcal/dollar cost post ww2 combined with some very dodgy science and politics.

Link to comment
Share on other sites

I'll bring the dead horse! The prefecture next to where I live now is known for it's (raw) horse sashimi. Did you ever happen to eat any while here, Kit?

 

The prefecture next to that one is known for its (raw) chicken sashimi, which I'd have to say I prefer over the horse both taste and texture-wise, though.

 

I've been to a yakatori in Tokyo with local colleagues that served chicken sashimi... tasted just like very fresh lean tuna, surprisingly tender.  Ironically, since many exotic things just taste of chicken.  That said the birds were slaughtered on the premises.

  • Like 1
Link to comment
Share on other sites

That's funny. I might compare the texture to canned tuna, but personally it tastes quite different than any fish sashimi to my taste buds. I don't mind the horse, but it's a bit tough for my taste. When eating sashimi, I'm often looking for that melt-in-your-mouth feel. Flavor-wise, I don't mind it at all. Many Japanese don't like the gamey flavor, but they traditionally have not eaten a lot of red meat, so that makes sense.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...