3 ms·

So for attentiveness and anxiety, the 5 and 20 mcg doses both had a statistically significant effect vs. placebo, but the 10 mcg dose didn't. Should that lead m

So for attentiveness and anxiety, the 5 and 20 mcg doses both had a statistically significant effect vs. placebo, but the 10 mcg dose didn't. Should that lead me to believe that there's really a dose-response curve with two peaks near 5 and 20 and a valley in between, or should I be wondering whether they got "lucky" finding a significant result with either the 5/20mcg doses? (P-value was <1% for the significant results.) Is the former (multiple peaks in the response curve) known to happen in other drugs?

Faint 6y ago
With only 24 participants it's no wonder if there is noise in the results.
vikramkr 6y ago
They tested so many factors and so many data points that you could start to consider it unintentional P hacking. 24 people might also be underpowered. Best to view the findings as hypotheses to study further rather than concluaive results. I personally can't think of a drug that stops working and starts working again with dose like this, but if LSD did have that sort of profile, that would be a really fascinating new mechanistic insight. Probably just noise though