Python Programming for Biology: Bioinformatics and Beyond

Statistical tests

Significance and hypotheses

Many  statistical  tests  are  tests  of  significance,  but  we  will  require  different  levels  of

certainty in different contexts before we say that a group of observations is ‘significant’.

This is often dependent on the cost of the outcome if we’re wrong. More generally though,

statistical  tests  are  about  probabilities,  and  indeed  it  can  be  useful  to  estimate  the

probability  of  an  event  occurring,  given  a  hypothesis,  without  formulating  any  notion  of

what is, or is not, significant.