As an electrical test engineer and lifelong audio geek, I've found Stereophile's measurements rudimentary. For the most part, the testing strategies, reporting, and the depth of analysis hasn't evolved in decades. And they don't tell us a whole lot; today all equipment should measure great to Stereophile standards, and, as mentioned, the attempts at correlating the minimalist measurements with the reviewer's impressions are weak at best.
Looking back at some magazine equipment test reports from the '70s and '80s is instructive. People are right to complain the reviews were 50% or more a thorough reporting on a multitude of valid measurements, and at best a paragraph on the (particularly electronics) subjective sound. BUT the measurements were solid and for the most part far more comprehensive than JA's. The instrumentation and printing of the day didn't allow for as many pictures of graphs, but the authors got the information across, at least to those who could interpret the specs for what they were. A lot of the measurements that had the most relevance are of little use for today's gear: tape decks, tuners, tone controls. But two pieces that were very well-tested that seems entirely lacking today were turntables and phono cartridges, where the measurements generally did do a reasonably good job of reflecting each component's sonics when contextualized with listening comments. Today all vinyl playback gear I've read about is entirely subjective. A frequency response plot of a properly aligned and loaded cartridge is VERY relevant but for one example.
Finally, to the points on this forum, even the speaker testing was often more involved. This was before FFT instrumentation and RTA analyzers, but Hirsch-Houck and CBS labs, among others, used the best techniques they had available to get a handle on loudspeaker measurements, and were clear where they encountered measurement limitations that may not correlate to actual listening. They published measured distortion throughout the frequencies at various levels, often toneburst reproduction and interpretations of it, impedance vs frequency, farfield frequency response with pink noise from various (and combined) positions, and more, that for 30-40+ years ago gave a good idea of how a particular speaker might perform. It doesn't seem this has evolved very much at the consumer level in any publication I'm aware of. Though I rarely read much of the audiophile press with any regularity!
For me, like the rest of us, -- in this forum -- how my Thiels actually sound with proper setup and upstream hardware is far more important than how they measure. But the engineer in me still wants to connect with my chosen gear from the standpoint of thoughtful, solid, and often clever, engineering I agree with, long-term durability, fad-free design, and as far as I'm able to discern, specs that don't reveal shortcomings or shortcuts in the design or execution of the end product.
To quote Michael Fremer, "I'm a Giant Walking Opinion."
