People also tend not to be very observant or, especially with parents, they have thrown their head in the sand. "No no you are/were a totally normal child".

That was my parents. "Weird" or "not normal" was bad so the reassurance was always that I was normal. Meanwhile my sister: yeah but remember how she used to watch that one DVD every single day for a year? And she read the same 5 books over and over while she ate the same breakfast every day even when you bought new books? And how you had to stock pile that cereal because if there wasn't any, she wouldn't eat anything because that cereal was breakfast and afternoon tea and sometimes dinner." 😂

My assessment involved a couple of hours of talking to my folks and my dad finally reading the end report just said "how did we miss this" 😢

it needs to be given to those people as an anonymous poll

You know exactly why this is a stupid, impractical idea. Yes, that's "better". The question in the OP is being filtered not only through what others actively say to you but also the extent to which you weigh those remarks. I understand that. Everyone here understands that.

No neuropsychologist is going to be anonymously polling a representative sample of people you interact with. At best they'll get a couple family members/caretakers (sometimes a teacher if the person is still in e.g. elementary school) to describe how you present to them. The purpose of questions like this isn't to hinge your entire diagnosis upon them; it's to get a bunch of data points and see how they cluster. It doesn't, therefore, have to be nearly as ridiculously strict as you're suggesting.

Here's the autism-spectrum quotient if you're curious what a real self-response questionnaire looks like. As an example of this style of question: "Other people frequently tell me that what I’ve said is impolite, even though I think it is polite."