Check out the two posts by Galit Shmueli over at Bzst on hypoth­e­sis tests: one on the value of p-​​values and another on one-​​sided tests.

She says “Shock­ingly enough, peo­ple seem to really want to use p-​​values, even if they don’t under­stand them.” That mir­rors my expe­ri­ence too.  Con­fi­dence inter­vals are much more use­ful because they pro­vide a mea­sure of the size of an effect, rather than test­ing if it is equal to some pre­spec­i­fied value.

Related Posts:

• Why I don’t like sta­tis­ti­cal tests
• Happy World Sta­tis­tics Day!
• Sta­tis­ti­cal tests for vari­able selection
• How to avoid annoy­ing a referee
• Why every sta­tis­ti­cian should know about cross-​​​​validation