HUMBEHV 3HB3 Lecture Notes - Lecture 9: Null Hypothesis, Alternative Hypothesis
Document Summary
Point biserial correlation: trying to see if one number predicts the other, one dichotomous and one continuous variable this is the point biserial correlation rpb. Phi: two dichotomous variables, same calculation as pearson"s r. It can be one relationship in the entire group but a different relationship in the individual subgroups: you should always consider how they gathered their samples and if there may be subgroups contained with different correlations. If the probability is high then we will say maybe it is not unusual, if it is low then we say maybe i am wrong. Null hypothesis testing: null hypothesis is zero ho r=0, alternative hypothesis true correlation is non zero h1 r does not equal 0, we assume ho is true. If p(cid:396)o(cid:271)a(cid:271)ilit(cid:455) is lo(cid:449) (cid:449)e do (cid:374)ot a(cid:272)(cid:272)ept is (cid:271)ut (cid:449)e do(cid:374)"t (cid:396)eje(cid:272)t it. If it is high then we reject the null hypothesis.