Originally Posted by Randomoneh
And what is!? Even asking "which image looks better" (double-blind, lot of participants, lot of separate measurements for every participant) would be enough. There is absolutely no reason to ask participants "in which image do you perceive higher resolution" because every difference between can be attributed to resolution, since everything else stays the same.
And as I've said before: differences should be even more obvious when comparing between different panels with different resolutions - instead of simulating resolution.
Doing some sort of resolution test charts on the TV screen. eg. at one point do two lines blur into one, basically switching between original image and video and downsampled images video, without telling the viewing which is which. Both test chart type videos as well as lots of other video types could be tested. You ask the viewer which is showing the most detail (not most real/being there).
Another way is like how they did with the EBU test - have 3 TVs positioned at the same distance from the viewer, one showing each resolution. But unlike them, you could randomly put the different resolutions on the different TVs. Also, unlike with the EBU test, you should make sure the test is unbiased. Also, you should probably do an independent standard eye-test of each of the participants, and take that and age into account in the tests (eg. do they all have at least 20:20 vision)?.
showed that the preferred viewing distance did not depend on the resolution or content
It's hard to believe it didn't depend on the content - surely how it was shot would affect how you viewed it. It would also have depended on the fact that the TV was only 60Hz too.