I have an idea for an experiment. I will be moving and doing intensive home renovations for the next few months, so I may not get to it for a while, but here it is:
1) Divide a large yeast starter into 16 smaller ones, measuring as precisely as possible, and begin multiplying 16 new starters
2) Divide the starters into 2 groups of 8 starters each, shuffling them so that the groups don't contain starters that were divided one after the other.
3) Add olive oil to 1 of the 2 groups.
4) Purge 16 sanitized 1/2 gallon jugs with CO2 and plug to keep in CO2
5) After boiling and chilling a 6+ gallon wort, divide equally into the jugs, taking care not to agitate the wort.
6) Shuffle the jugs of wort and divide into 2 groups of 8
7) Aerate 2 of the four groups of wort with an airstone or other method
8) Pitch the yeast into the wort to make 4 each of the following 4 combinations
---a) olive oil + aeration
---b) olive oil + no aeration
---c) no olive oil + aeration
---d) no olive oil + no aeration
9) Add airlocks and ferment as usual, labelling groups and shuffling them together to account for any small differences in temperature in different parts of the room.
This way, you have a control group with no aeration or olive oil, as well as a group with both methods to see if there is any additional benefit in using both. You could even make a lid for the brew kettle with notches for siphon/chiller lines and purge the kettle with CO2 so any oxygen that was driven off while boiling wont be reabsorbed. The OG reading could be taken from the original wort or if you really want to be anal, from each of the sixteen samples. During active fermentaion, record daily observations of each sample such as krausen depth, airlock bubble frequency, etc. Also watch trub depth, and record when each sample begins to flocculate and clarify. Check gravity readings at the end, bottle and perform blind testing with a group of people, each privately recording their own subjective observations.
Having randomized test groups instead of single samples for each technique would help minimize any affect of minute differences in yeast cell count in the starters. Doing the experiment this way gives a lot more numbers to compare: min, max, mean, median, mode and range for each measurement involved. The more test samples in each group, the less significant the variations become to the end result, the more accurate and objective the experiment can be.
I must admit, I haven't read every last post in this thread yet, so someone may have proposed an experiment like this already. If so, ignore this post. It seemed to me like the hang-up was over not being able to accurately measure the cell count in single samples, whereas using randomized test groups to a large extent compensates for lack of high tech lab equipment.
Like I said, I'll barely have any time to brew in the next while
so I probably wont be able to tackle anything like this at least until the summer, so feel free to beat me to it and post your results. In fact the more people who do an experiment like this, the more (scientifically) conclusive our results will be (assuming there is a positive correllation between our results).