3 ms·
The study linked in the article says "Information was collected in face-to-face computer-assisted interviews conducted in the participants’ homes." [1] It seem
by resu_nimda 9y ago
The study linked in the article says "Information was collected in face-to-face computer-assisted interviews conducted in the participants’ homes." [1]
It seems like this is a study where people might be likely to lie about their consumption, or even earnestly estimate it incorrectly. Is there any way to control for that? The distribution feels too extreme, like a significant number of people underestimated and overestimated at either end of the spectrum.
[1]https://pubs.niaaa.nih.gov/publications/AA70/AA70.htm https://pubs.niaaa.nih.gov/publications/AA70/AA70.htm
- sjg007 9y agoMultiply by 1.5 or 2 usually. Much like software estimates.
- pavement 9y agoOne sure-fire way to control for this sort of loose testimonial data is to collect many volunteers among those who affirm such habits, and follow many or all of them longitudinally, and engage with them and make observations, hopefully without influencing behavior too much. It's not a short-cut by any means, if that's what you had in mind.