Computer review methodology is terribly anti-scientific in general. For example, look at all the pomp and circumstance around reviews of desktop PC coolers - we

Computer review methodology is terribly anti-scientific in general. For example, look at all the pomp and circumstance around reviews of desktop PC coolers - we used this motherboard and this RAM with this case in a room this temperature blah blah. Whereas the physical reality is that thermal resistance (degrees temperature rise per watt dissipated) sidesteps all of that. The right way to characterize a cooler would be to use a constant heat source to measure the thermal resistance (via temperature) and noise while varying the fan RPM, and create a simple graph that could be straightforwardly compared. But then they wouldn't be able to focus on a subjective narrative supporting whatever manufacturer paid them to write the article.