I did a blind trial on myself by getting liquid multivitamin (bargain large European superstore brand, powdered), and creating a sequence of vitamin or placebo based on a pseudorandom hash function to take every day with breakfast in drink. The sequence is prepared a month ahead so I don't know on any specific day what I'm getting.

I also record my mood and excercise at some time most days with a survey.

After 2 years, I saw a significant decrease in general happiness on days I took the multivitamin, and a lesser (not p<0.01) effect the following day. I was less likely to leave the house, and less likely to plan a social activity on those days.

I've now stopped taking the multivitamin for that reason.

I'd still like to know if I have stumbled on a p<0.01 fluke, or if I'm unique somehow, if the multivitamins have a negative short term but positive long term effect, if these particular multivitamins are harming me somehow, or if all multivitamins have this issue.

Life isn't long enough for me to find out, but if anyone else is happy to take on a bigger trial, results could be interesting.