Ran across this YouTube video on Anvil mash efficiency. Pretty well spells out everything I was thinking that was causing low efficiency
The author of the video claims that his 72% vs. 76% difference is significant, but statistically his data is not good enough to show significance. If you do a tolerance analysis taking into account the possible errors/differences in volume measurements, grain & hop absorption rates, grain bill weight measurement, SG measurement, grain moisture content, and grain potentials, you find that most efficiency measurements at the homebrewer level are only accurate to +/- about 3% - 4%. By using the same grain in both batches, he eliminates the possible errors/differences in grain moisture and potential, but the other sources of uncertainty are still there. To claim significance with this small a difference in result, he would have to repeat the experiment multiple times.
Just because the result is not statistically significant, doesn't mean it isn't real. It might or might not be real - this experiment by itself is not adequate to determine that. Since there might be a real difference, you should probably run your own side by side experiments to see what results you get. Just remember, a single experiment is not likely to prove a difference, unless the difference is much greater than the expected variation from run to run.
Brew on