bivariate correlation research

bivariate correlation research

1. comparing two variables

   • r
     • -.1 to .1 weak
     • -.3 or .3 moderate
     • -.5 or .5 or higher strong
   • both variables quantitative
     • scatterplots
   • 1 quantitative 1 categorical
     • bar chart
     • pie chart
     • t-test
       • comparing the means of both samples?

2. vs. univariate

   • looking at one variable, which has multiple levels
     • e.g. weight

3. statistical validity

   • effect size
     • strength of a relationship (r)
   • outliers?
     • Q1 - 1,5(Q3-Q1)
     • Q3 + 1,5(Q3-Q1)
   • restrictions?
     • not showing all possible values
       • correlation might look weaker
   • statistical significance
     • is the correlation caused by chance?
       • probabilistic probability p
         • the chance that the association came from a population in which the association is zero
           • statistically significant when p <0.05
             • the stronger r, the higher the chance that it's statistically significant
     • depends on:
       • effect size
       • sample size

4. curvilinear relationship

5. third variable causing correlation

   • spurious association
   • lurking variable
   • moderator