There is an app on the market called battery drain...
Be careful using this, it will drain the battery VERY fast, potentially heating up the battery to beyond safe temparatures. DONT SELECT EVERY OPTION!!!
Here are the options you can select to have it do all at once.
-Turns on Permanent Vibration
-Enables Wifi Send Data
-Constantly Max Loads CPU
-Sets Screen Brightness to Maximum
-Enables GPS Location Services and Constantly asks Device to Refresh Position
-Turns on Camera Flash
-Turns on Bluetooth and Constantly Scans for New Devices
I would also grab battery monitor widget pro by 3C (the paid version is $4) or another battery app that has detailed logging.
Consistency is the key, what I would do is power down the phone and let the battery completely charge (use the volume rockers to turn the screen occasionally to see the battery charge level, its fully charged when there is a 100% shown inside the green battery).
Set the battery logging app and the battery drain apps either on the lockscreen if you have something like widget locker or set them on the first homescreen that appears on an initial boot so you can access them quickly.
Then unplug the phone, note the current time (open the clock on your PC for seconds), power the phone up, wait a set amount of time (say 1 minute from when you initially turn the phone on) and launch the battery drain app, turn on maybe 4-5 of the options (dont use all of them, you dont want to damage the battery), say max screen brightness, camera flash, wifi, bluetooth, & GPS and wait til the battery dies and note the time when it dies (shouldnt take more than 1 hour).
You can test each battery this way (make sure you pick the same options in the battery drain app) and if you have an app doing battery logging, you can pull up the stats after the phone dies and you plug in the charger to reboot it.
Its also key that nothing changes (dont install any new apps or remove any apps) between testing. What most testers do is factory reset the phone, leave all the settings at stock and install only the apps they need for testing. This might be a bit extreme, the key tho is to have consistency.
Using both batteries and testing them in this fashion would give more information rather than anecdotal. However, I can understand that it is a time consuming process.