Null hypothesis significance testing

Null hypothesis significance testing

1. Parameters show effects/what we are looking for

   • I.e relationship between variables or difference between means
   • all parameters have a related distribution
   • For any parameter (b value) we can work out probability of getting outcome if null hypothesis is true

2. sample size can affect significance, bigger sample creates more significance

3. effect size

   • Standardise parameter = B = r
     • 0.1 = small. 0.5 = large
   • Cohen's d = standardised difference between 2 means