I too have been curious about this and devised a way that I think you can capture fairly accurate data with a single set of dual sided power meter pedals. My interest was both curiosity and to convince myself that my stokers claims that she was working just as hard as I was were indeed true. Always looking for a bargain, I invested in the IQSquared cheapo power meter on Kickstarter and while I am still sitting by the mailbox waiting for them to arrive
I will share what I had in mind.
With just 1 set of dual sided power pedals, do several representative rides with both pedals on the front and capture the data. Looking at the Left/Right power balance, calculate an average L/R balance for the captain. Then move the pedals to the stoker and do the same - calculating an average left/right power balance for the stoker. Here is where the brilliance comes in: Put one of the pedals on the captains crank and one on the stokers crank. After capturing a rides worth of data, import the data into Golden Cheetah where you can see and edit both both pedals power data. For the captain, copy the captains pedal data to the other pedals data and apply the appropriate captains L/R balance. This should give a pretty close overall picture of the captains ride power. Then do the same for the Stokers data. You should now have a good data set for both riders power to compare (and settle those outlandish claims once and for all
).
There is some off bike work to do here that may involve spreadsheets and graphs, and it requires that both captain and stoker use the same, or compatible, pedal cleats. If you didn't want to go the route of importing and manipulating the data in a spreadsheet, you could do the same and just apply the L/R balance to the average power from the ride. This number would not be too interesting to me as all it would show is that the power from both riders is different.