By and large, I think most of us would agree that a simple total BTU consumed measurement by nation (the image used by New Scientist as an example) is only superficially useful; big countries will easily use more energy in total than smaller countries, even if the smaller countries are more wasteful.
The next more complex version is energy use per capita. This is better, but still misses quite a bit. What do they do with that energy?
"Energy intensity" seems to answer that by comparing energy use not to population, but to GDP. In these posts at WorldChanging, I called this value the "use efficiency of energy," as it tries to show how much use-value you get out of a given amount of power. I know that some environmentalists dislike intensity/use efficiency as a metric, as it makes the US position a bit more ambiguous -- yeah, the US uses a lot of power, but the US does a lot with it, too.
But GDP sucks as a metric, for a variety of reasons. I've played around a bit with using slight modifications (such as purchasing-power-parity valuations), but it struck me today that it's simply not the right category to examine. We should, instead, look at standardized quality-of-life metrics as the point of comparison to energy use. Lifespan, healthiness, availability of health care, leisure time, Internet access, creative work publications, education levels, voting percentages -- the variety of factors that tell us not whether a society is wealthy, but whether the society is thriving. Most of those metrics wouldn't aggregate in a population-linear way (as GDP more-or-less does), so we'd probably need to compare not to raw energy consumption, but energy consumption per capita.
There's undoubtedly quite a bit of room for debate about how to measure these various quality of life factors, but the goal is uniformity/standardization, so even if they combine in somewhat weird ways, as long as the combination is consistent across countries being studied, it's okay.
I haven't actually started this study; it just occurred to me today. But I'm wondering if any of you know whether (a) it's already been done, or (b) there's some Very Good Reason why it wouldn't work...