0 / 60 seg.

We discovered that averaging the answers of the groups after they reached consensus was much more accurate than averaging all the individual opinions before debate.