In the case of the Abyss, he both measured it wrong and it's also just a bad headphone, intentionally compromised to produce a certain coloration (which to many people is enjoyable, I suppose).

It's also a bit tricky to gauge what constitutes "correct" measurement because the Abyss headphones are intentionally ported, meaning even perfect seal on rig produces some bass rolloff. Even so, the distortion there is still unacceptable for the price, that much I agree with.

Another example of him measuring wrong is any of his Stax reviews. You'll see that his Stax plots all have extreme rolloff in the bass, which just is not the case when properly fitted on a human head.

It's known in the community that the contour of Stax pads makes it hard to seal properly on traditional measurement rigs which do not have malleable human skin. They're designed this way to avoid needing any clamping force, preventing pressure and damage to the delicate diaphragm.

Despite having this explained, Amir hand-waves it away by claiming this is also how he hears them on his head. Which is within the realm of possibility of course (and is a fair criticism to make that the headphone is not suitable for some head shapes), but is not representative of the real performance of the headphone in normal circumstances.

Then there's his Stax amplifier measurements, where he argues with Kevin Gilmore after being told that his measurements are totally wrong due to using an incorrect load, forcing the amps to drive something they were never intended to.

It's a common theme for Amir to double down on his incorrect takes and try to discredit people who criticize him.

I've noticed that even a lot of his less questionable measurements end up being incomplete, and he basically just measures whichever plots he feels like at the time, instead of providing a consistent test suite to compare different amps with. Just look at how many times someone in the comments has to ask him to make a 32-tone FFT, for example.

Then for the longest time he never bothered to get a balanced 32 ohm load, making it impossible to compare balanced with single-ended amps from his measurements. It would've been as simple as soldering two resistors to an XLR connector, but he never bothered.

It's lazy and unacceptable for someone claiming to be "scientific".

I wouldn't care as much if it weren't for his large following, so the misinformation ends up being propagated far and wide and defended by people who understand even less about audio gear.