I don't have the iPhone 4, nor any of its main Android competitors (I am considering a Samsung Galaxy S to test Super AMOLED and Hummingbird, but I still think it is far too expensive at 450€ unlocked), nor would I have the time to actually do this type of thorough testing, but here is how I would suggest that it would be best for an independent and trusted media to test the iPhone 4's antenna. I think it could be much more useful to get the real user testing data than to wait through more back and forth between each of the companies advertising and spin doctors, news conferences, blogger fanboys and other media stunts.
1. Make 100 calls of at least 2 minutes in length with each device, holding it in the left hand. Preferably make the test using people not familiar with the issue at all (ask anyone in the street who have not read the news or seen any pictures of how Apple recommends holding the phone), just ask them to make the phone calls holding the phone with their left hand. Make sure the people are speaking to each other constantly for at least those 2 minutes per call as silent calls may emit less data and generate less dropped calls.
2. Try to test this at different locations, be them central where signal reception is known to be strong as well as places where the signals are known to be weak.
3. When possible, test the different networks. For example, test other Android devices that use the same network as the iPhone 4 as well as testing the Android devices that work on other networks.
4. If possible, try to measure the bitrate and codec, thus quality of the phone calls. Perhaps one company uses lower voice call bitrate, stronger compression, maybe non standard compression technologies which may result in less dropped calls due to bad signal. For example, one phone brand on one network may have fewer dropped calls but also a lower voice call quality.
5. Report how many dropped calls there were out of 100 calls on each of the devices.
6. Launch downloads over 3G of certain different (un-cached) test 10 megabyte files hosted on specific servers. Ask the user to please hold the phone with their left hand and again with their right hand. Measure exact transfer bandwidth differences. Try to download on all the different phones at the same time and at the same location. If possible also compare upload speeds the same way if it is possible to upload something to a same server online using some browser or app based upload functionality.
I don't think this should only be an issue about the number of bars displayed in closed, open source or hacked mobile phone OSes. This is about calls being dropped or not. And about 3G data transfer bandwidth.
If the iPhone 4 is proven to have significantly more dropped calls when the phones are held with the left hand, then I do think it would be fair for consumers to claim it to be defective and to demand a global recall. And they should be entitled a new one with a fixed antenna or for a full refund. Not within any 30-day usual return schedules, this would have to be valid for anyone having purchased a defective phone within the whole 2 years of the signed contracts. And I don't think consumers have to be satisfied by just getting a $1 piece of plastic case to cover it.