We've seen how using NNTs can help to decide whether the results of a trial are clinically important.

However, you'll recall that, at the start of this section, we stated that there are three possible explanations for the results of research:

1) Bias
2) Chance
3) Truth

Before we decide to act upon the findings of research, we need to consider the second of these possibilities. Can these results be explained by chance variations between the groups?

Or, more precisely, could these results have happened even if there were no difference between the treatments?