I disagree.

It's not insufficient data, it's uncharacterized data.

If more-or-less everyone with exposure to both cultures is in broad agreement about the differences, that's not "insufficient". It's probably actually more data than any formal study would generate.

But it's also lower-quality data, with a greater chance of being distorted by various cognitive biases. Which means that it doesn't meet current regulatory standards[1] for use on large protected-class[2] groups of humans.

.

[1] as loosely and collectively defined by the morality police, loudest / most extreme non-ignorable activists, Internet Lynch Mob, etc

[2] which includes pretty much anything these days