Chapter 19 Correlation Flashcards
Biserial (correlation coefficient)
Correlation used where one variable is artificially dichotomous; formed by categorising from an underlying continuous and normal distribution
Correlation
A (standardised) measure of relationship of co-variance between two variables
Correlation coefficient
Number signifying strength of correlation between two variables
Curvilinear correlation
Correlation between two variables with low r value because the relationship does not fit a straight line but a good curve
Negative correlation
Correlation where, as values of one variable increase, related values of another variable tend to decrease
Positive correlation
Correlation where, as values of one variable increase, related values of another variable tend to increase
Correlation strength
Measure of the closeness of the relationship, indicated by a number between 0 and 1 (positive or negative). Strength does not indicate significance.
Dichotomous variable
Variable with just two exhaustive values (e.g. male/female)
Linearity
Extent to which a relationship between two variables can be represented by a straight line rather than, say, a curved line
Pearson’s product-moment correlation coefficient
Parametric measure of correlation
Phi coefficient
Measure of correlation between two truly dichotomous variables
Point biserial correlation
Measure of correlation where one variable is truly dichotomous and the other is at interval level
Range restriction
A selection of cases form a larger potential data set, which has the effect of distorting the true population correlation
Scatterplot
Diagram showing the placement of paired values from two variables on a two-dimensional chart
Spearman’s rho
Non-parametric, ordinal level measure of correlation; Pearson correlation on ranks of the paired raw scores