NUR 80A/B Lecture Notes - Lecture 6: Phi Coefficient, Univariate, Contingency Table
Document Summary
Measures of central tendency and variability are critical for describing a distribution. Sometimes, we are interested in describing the relationship between variables. Cannot make any conclusions with univariate stats tho. Describes the __correlation__ between two variables (iv and dv) Dv) must be continuous can be shown graphically (scatterplot) indicates __direction and _magnitude__ of relationship between the two variables. Other tests of relationships include phi coefficient and spearman"s rank-order correlation. Continuum runs from -1 through 0 to +1. 1 and +1 indicate perfect correlations variables are so perfectly correlated, they share everything in common. Don"t expect to find a perfect correlation between any two variables in the behavioural or social sciences extremely rare. Prediction one variable is used to make prediction about the other (more on this later!) ie: prediction uni scores from high school scores. Validity measure correlation between new test and other measures to demonstrate new test is valid.