You have a great question here - but in order to have a study that will have statistical merits - you will need more than 2 rats. At a minimum you will want 10 in each study group- as well you will want a 'control' group. I'm at a loss as to what you would feed the control group - but you want a group to compare both rat a and rat b to.

I would suggest dusting off the 'ole stats textbook and really diving into what kind of numbers you would need and also figuring out what tests that you will want to do on your data in order to 'prove' what you want.

This would be a very interesting study! And if you need help with stats - i suppose i could dust off my old text book and help!

BAMBAM- Sweet idea. Are you going to seriously go for it? Good luck and I hope you find out some cool stuff. I can't help you on what to feed the control rats, but here are a few things that I thought of:

1. You definitely need 10+ rats for A and 10+ B. If you can get more than 10 go for more. More is always better especially when you're using statistics.

2. You say "the rat that eats according to USDA Dietary Guidelines for Americans grows fatter" but I think this needs to be more specific. The USDA is making these recommendations with reference to blood cholesterol levels and thus cardiovascular disease, not obesity. We make assumptions that these are related. You can simply replace the "gets fatter" with something along the lines of "increases dangerous small LDL blood cholesterol particles" Or you could go all out and hypothesize that the USDA rat will not only have bad blood levels but will also become obese/gain body fat.

3. Which gets me to my next point. What about blood sugar? Fructose is sugar. Shouldn't the paleo rats be fed low GI tubers? Aren't high triglycerides linked to cardiovascular disease?

4. Also, are you going to call rat B the "paleo" rat? Because I think "paleo" definitely isn't necessarily that high carb, but I guess it isn't necessarily "low carb" either, it just depends on the individual needs of the dieter? How many paleo eaters get 1/2 their calories from carbs? Don't high carbs cause high triglycerides = cardiovascular disease? (Obviously you probably know more than me in this area, which is why all these statements have ? marks at the ends)

PS. my degree is in outdoor recreation (yes I have a degree in FUN). I took a few research methods and stats classes. This is definitely not where my expertise lies but I figured I would share my 2 cents.

Hey, I'm Ali's friend doing my MPH. I've got a few thoughts looking through this.

1- You've got WAY more than a master's thesis here. One of these experiments alone would be ambitious for a master's thesis.
2- What sort of funding and facilities do you have access to? You'll need to meet regulations and have IACUC approval for an experiment like this to be accepted by the scientific community. 3- Typically the level for statistical significance is around 30, depending on the power you're trying to achieve. That said, funding shortfalls make experiments with 5-10 rats/mice per group very common. 4- Why rats, specifically? Seems like the same thing could be accomplished for less expense or with more statistical power with the standard BALB or C57 strain mice. 5- With regards to groups, you have two options. You can either pick one thing to vary between two groups (plus a control group on standard lab chow) or you can have several groups and more dietary variations. With the latter option, you may be limited in what you can claim by comparing the groups to each other, hence the importance of a control group. 6- Be very very careful assuming the results of your experiments while you're designing them. Go to the library or find a professor with an expertise in scientific ethics and ask for help. Statements like "I really want to punch the USDA in the teeth" are going to sink your credibility, and with it, the results of your research, if they ever get out. Look at ClimateGate. And also, one study (or even a series of studies) is never going to form the basis of policy change. The agency that tried it would get torn apart by lobbyists and the courts. If you want to change policy with scientific studies, you'll need to get a group of researchers to duplicate your work and expand upon it until you have a body of evidence from a variety of people that sends a clear message. At the EPA, we started to hit that tipping point when we had 15-20 studies showing similar results. The USDA is much more traditional and under the sway of the food industry, so you'll need way more than that. 7- In terms of experimental design, I'd probably start with something like this. 1, control group eating standard lab chow. 2, group eating USDA standard diet. 3, group eating low-carb paleo diet. 4, group eating higher-carb paleo diet. That should give you enough to do some nice analysis, and give you questions to answer for your next experiment. ;) 8- And speaking of analysis, start thinking now about the tests you'll need to do to get the information you want. You'll have to document all of it for IACUC approval, and most likely for funding approval as well. Regular blood draws sound like they'd be in order, as well as body composition estimates and autopsies at the end of the experiment. How long do you want to keep this running? Does the information you want require special sacrifice methods? At this point you're getting past my expertise, but I know that you'll need to think about it.

I've written enough of a novel now, so I'll stop here. Hope this was helpful!

We have some problems here Houston

1. Not enough rats.
2. Rat studies prove rat theories.
3. Forget about Berkeley or Stanford. Pray that Ball State takes you.
4. You have to normalize the diets for equal taste. If you use tasty cookies for group A you have to use smoked dry rub ribs for group B. Or gruel and boiled chicken, your choice.

Aside from all that why is the paleo diet so high in carbs?