Throughout History Humans Only ate Once A day

[Whisper]

Your assumption that n must be greater than 30 is not sound statistics. If the researcher does good, high quality research, then they will calculate their power (beta) prior to completing the research. With that, they can calculate an n that will yield appropriate statistics with the least amount of subjects necessary. In some research studies, an n of 15 is more than sufficient with a power of .8. That number is usually considered ideal for research. Please see this: http://en.wikipedia.org/wiki/Statistical_power

Keep in mind that a power of .8 is pretty darn high, and that sample size requirements can jump considerably when you get to the more modest .5 and .3-ish ranges. Other factors come into play regarding sample size as well, although yes, a priori power calculations can be a good way to give you a ballpark of the figure for which you should be aiming, and are necessary when the recruitment and/or assessment process is going to be an expensive and time-consuming one. Neuroimaging studies are one area where this sort of sample size limitation is very common/rampant.

Although yeah, statistics somewhat aside...an N of 15 people is still only 15 people. That's going to restrict your generalizations no matter how you look at it (after all, to how many different populations could that 15-person sample truly be similar?)

 

[Wyndru]

Hmm, and everyone thinks I'm crazy for only eating dinner on most days.

Well, ok I agree it's bad for me...since if I was going to eat a single big meal it should be breakfast, but I hate eating in the morning, and I don't have time most days for lunch (I usually just eat a protein bar).

The problem is I eat WAY too much at dinner, and usually get really sleepy after that meal, which makes my workouts rough.

 

[SociallyChallenged]

Keep in mind that a power of .8 is pretty darn high, and that sample size requirements can jump considerably when you get to the more modest .5 and .3-ish ranges. Other factors come into play regarding sample size as well, although yes, a priori power calculations can be a good way to give you a ballpark of the figure for which you should be aiming, and are necessary when the recruitment and/or assessment process is going to be an expensive and time-consuming one. Neuroimaging studies are one area where this sort of sample size limitation is very common/rampant.

Although yeah, statistics somewhat aside...an N of 15 people is still only 15 people. That's going to restrict your generalizations no matter how you look at it (after all, to how many different populations could that 15-person sample truly be similar?)

I'm aware of that, but it can be acquired with specific designs, goals, and measures. In my line of work/research (physical therapy), we frequently set that power as 0.8 and design the study around it.

I understand that it does, no matter what, limit external validity. However, the greater the power (beta) and if p is < 1-alpha, then the error is pretty low. With that, we need more and more studies to maximize the summative findings and improve general external validity.

 

[Whisper]

I'm aware of that, but it can be acquired with specific designs, goals, and measures. In my line of work/research (physical therapy), we frequently set that power as 0.8 and design the study around it.

I understand that it does, no matter what, limit external validity. However, the greater the power (beta) and if p is > 1-a, then the error is pretty low. With that, we need more and more studies to maximize the summative findings and improve general external validity.

Error is generally going to be dictated primarily by a variety of things outside the statistics used (e.g., methodology), although choosing improper statistical methods/models can certainly introduce error that wouldn't have been there otherwise, can significantly limit the amount of controlling for error that can be accomplished, and/or can make the entirety of your findings incorrect and uninterpretable.

All that being said, I agree with your main point that N<30 doesn't mean the study should be written off outright. Heck, there are just as many potential problems that can pop up when your sample size is very large rather than very small. The main trouble in most areas of research, to which you alluded in your post, is the lack of replication of significant findings.

 

[SociallyChallenged]

SC, with regard to intermittent fasting, what is your opinion of the 5:2 plan? Where you eat normal calories 5 days and the other 2 days limit calorie intake to about 1/3 your normal amount?

Apparently it has some really positive effects on blood chemistry in terms of lower cholesterol, blood glucose, etc.

I just recently heard of this and was wondering what you thought of it.

I honestly don't have all the answers. I have searched repeatedly, but there is not a depth or breadth in the research yet that has allowed me to draw enough appropriate conclusions to say anything about IF, let alone IF programs.

 

[marvdmartian]

"The Romans believed it was healthier to eat only one meal a day,"

Uh, yeah.....they also practiced gluttony, followed by evacuation, followed by more gluttony. In other words, they ate until they were stuffed, then vomited it all back out, only to go back and feast some more.

Sounds really healthy!

 

[BeeBoop]

Uh, yeah.....they also practiced gluttony, followed by evacuation, followed by more gluttony. In other words, they ate until they were stuffed, then vomited it all back out, only to go back and feast some more.

Sounds really healthy!

You are the second person to talk about just the Romans. It's like no one bothered to read the article in it's entirety as the Romans is just one facet of food history.

 

[SociallyChallenged]

Uh, yeah.....they also practiced gluttony, followed by evacuation, followed by more gluttony. In other words, they ate until they were stuffed, then vomited it all back out, only to go back and feast some more.

Sounds really healthy!

It's an example, not an epitome.

 

[Kaido]

Not true - research shows no difference in baseline metabolism due to meal frequency.

I had a buddy who followed the "one meal a day" plan for bodybuilding for a long time, just to try it. Results were exactly the same, surprisingly. I'd like to argue against it, but I've seen the results and they don't appear to be any different than any other method.

I prefer the 6 meals a day method because I stay full throughout the day, which helps to combat junk food cravings a lot. Because I'm full, the temptation for junk food is there but it isn't an immediate need like it would be if I were hungry. I also like having the energy that comes from food - getting constant hits of protein and whatnot is nice.

However, I've been on the 1-meal-per-day thing for the past 3 months or so due to my work schedule. I pretty much get home in time to eat dinner, chill, and go to bed. The first month took some getting used to, but because I stayed busy all day, I mostly didn't notice it. Eventually my body evened out and I was fine throughout the day with just a bottle of water. My weight hasn't changed at all, if that counts for anything. I can tell if I had a more physical job, it'd be harder because having that constant energy input from regular meals helps quite a bit. I plan on going back on the multi-meals thing once my work projects wrap up.

 

[Kaido]

Also strangely relevant:

http://assets.amuniversal.com/167745e059d201301723001dd8b71c47?width=900.0

:awe: