Old thread but, for those encountering it in the future, as I did, it is important to emphasize that a single test or comparison of battery life proves virtually nothing!
Also, Apple's battery gauge is notoriously off in many cases, on many models, and under many conditions. That’s why battery life estimates based upon battery % and partial discharge are incredibly unreliable.
Starting from 100% and running until shut down is crucial.
Also, for a valid test, it's necessary to ensure that the devices are at the exact same brightness level — measured from the display, not based upon the brightness slider.
Also, you must make sure that the two devices are using the same video settings, ad blockers, cookie blockers, etc. They should have the same drive space. Any test should occur AFTER quitting all apps and doing a complete restart of the device. Only one app should be running. (The other viable approach would be to run all the apps one usually does, ensure, somehow, they are in the same state, do the exact same tasks, and see how battery life compares from 100% to shut down.)
Multiple tests, using different videos and different apps, are required, too, with averages and standard deviations presented.
As someone who has done applied statistics and research for a living, a 3% difference in a *single* test almost assuredly does NOT reflect what the result will be after many tests. A single measurement will likely reflect some extraneous factor (such as those mentioned) and not reflect actual battery life. The findings from multiple, well-controlled tests are needed and they might show no difference in battery life, small differences, and those could be higher or lower.
Furthermore, one needs to use common sense about battery life “measures” and real world use. Unless you are someone who routinely watches videos for more than 10 hours straight, the findings have little meaning. Most people will recharge well before then.
