T Nation

How to read stats in studies

In many studies I find on the net it says things like:

“Lean body mass was increased in men taking 10g/day of testosterone gel by (2.74+ 0.28 Kg, p=0.0002)”

Can anyone explain the numbers in brackets? I hate to have to ask a question that probably has an obvious answer, but I’ve been wondering about this for months and never found an explanation.

Thanks in advance.

The number 2.74 +/- 0.28 is the mean and the standard error. In the case you provide here, the 2.74 is the average increase for the entire group. The 0.28, the standard error, is a measure of the variability between the subjects. Larger standard errors mean that subjects varied quite a bit in their response to the treatment provided. Small ones indicate that everyone responded in a similar manner.

The P=.0002 is the probability that the results observed were due to chance alone. The lower the P value, the lower the probability that the results seen in the study were due to random chance. In this case, there was only a .02% chance that the results observed were due to random chance.

I’m still trying to forget my experience with sadistics-oops I mean statistics-but that means 2.74 give or take 0.28, so 0.28 is an error value that gets spat out of some formula that I can’t remember. I’d have to look up what the p means though, I think it’s the same as r squared. (it couldn’t be p-value right? Isn’t that for probability?)

(2.74 +/- 0.28 Kg, p=0.0002) The 2.74 is the mean change from baseline. So this group gained an average fo 2.74 kg. The +/1 0.28kg is the standard deviation from the mean. The standard deviation demonstrates the range of differences seen (not exactly correct, but you get the idea). Basically if you look a the range created by 2 standard deviations, you find that 95% or so of the subjects fit into this range. For your example, 95% of the subjects would have gained 2.18-3.3 kg. The p=0.0002 is the statistical significance. Each researcher has to set what p value creates a significant relationship, but generally if you see a p value less than 0.05 that is considered statistically significant.

The 2.74 in your example is the average increase. The plus or minus value of 0.28
kg is either the standard deviation or the standard error. You cannot tell which unless
they say. If it’s the standard deviation,
this means that (approximately) 65% of cases
fell within 0.28 kg one way or the other
of the average, approximately 95% fell within
0.56 kg (twice that), etc.

If it’s the standard error, this is looser
and the significance depends on the sample size. So an 0.28 kg standard deviation is
pretty consistent results, whereas 0.28 kg
standard error might be pretty inconsistent.

The p value is the probability that an average
result as large as this could have occurred
just by chance, not from actual cause. In this case, it’s only a chance of 0.0002, or one chance in 5000. (Hypothetically, if it was found
that a few men in the control, untreated group gained muscle mass without receiving testosterone, it could be that the treated group all or mostly were composed of men who happened to be like that.)

Wow! I’m humbled by the swift and thorough responses. Thanks go out to all for the explanations. I’m off to find some more studies with my new found knowledge :o)