Cognitive dissonance is widespread. An old roommate once informed me she had given up drinking, because she switched from vodka crans to wine. It's similar to vegetarians who sometimes eat chicken and people who "don't watch TV" because all the shows they watch are on Hulu.

Well, some of us actually -never- drink.

I personally don't like the taste, nor the side effects - I wish alcohol was the social grease it's touted as, but for me it's only bad side effects.

Nobody in my family drinks (as in, you won't find a single alcoholic beverage at a Christmas party or weekend barbecue for example), and my current girlfriend was a very casual drinker and has quit since she started dating me.

Many of the surveys I've seen in doctors offices & similar don't distinguish well between "almost never" and "never". I know a fair number of people in my family are in the "almost never" category. They might have a glass of wine at Christmas or Thanksgiving dinner, but that's about it. I don't know what they tell their doctors, but given the usual checkboxes for frequency of alcohol consumption, "never" is probably what I'd select.

I would guess that the more a person drinks, past a certain point, the more likely they are to lie about it.

they used self-reported but cite four studies case where in slightly older populations self-reporting is deemed accurate.

see section "Materials and methods" subsection "Baseline Alcohol consumption" in the linked paper.