We had an 8th grade social studies teacher who almost seemed to want us to cheat: he would set very difficult essay tests and then leave the room for nearly the entire class period. People discussed the answers.

I don't call that cheating. I call it 'cooperation'. Calling it cheating would be an insult to the term.

10th grade french, I remember some people suggesting cheating on a test because it would be easy and at the time I went along.

Yes, cheating.

Also I remember someone suggesting I read "L'etranger" in English translation and I did that, it was way easier.

Mere common sense. If a test in no way distinguishes between knowledge gained by different methods it has no right to call one method 'cheating', no matter what it may claim.

My point: if 1/3 I believe it more likely that people will mistakenly report they didn't cheat when they did than vice versa. And I believe it is easy for people to "forget" they cheated.

Absolutely. This particularly applies to sexual 'cheating'. I am referring explicitly to reports that are genuinely mistaken, not deliberate lies. This is having sex with someone who is not your partner. That's not something that isn't a big enough deal to remember. But people can compartmentalize this knowledge. There are also people that "don't count". When talking to friends who have their confidence it is not unheard for people to say "I've never cheated". When prompted with the example the genuine response is a double take and the impulse to say "Oh, but he doesn't count!"