PSYCO212 Lecture Notes - Lecture 18: Phi Coefficient, Effect Size, Null Hypothesis
3 Mar 2017
School
Department
Course
Professor
Document Summary
Bivariate correlation: a correlation between exactly two variables. R=0. 10 is a weak correlation, 0. 30 is a moderate correlation, and 0. 50 is a strong correlation. Point-biserrial correlation: a statistical test used for evaluating the association between one categorical variable and one quantitative variable. Phi coefficient: statistical test to measure association between two variables. Bar graphs show the average/mean between each group instead of the exact scores. Statistical validity in association claims: we need to consider the effect size, any sub groups or outliers, the statistical significance of the relationship, and whether a zero relationship but actually be curvilinear. Effect size describes the strength of an association. In some cases, effect sizes indicate how important the result is. Depending on the context, even a small effect size can be important. Statistically significant: a conclusion that a result is extreme enough that is is unlikely to have happened by chance if the null hypothesis is true.
