MKTG 3596 Lecture Notes - Lecture 7: Null Hypothesis, Tegra, Covariance
Document Summary
Ex. linear regression, logistic regression: data driven. A statistic that expresses the magnitude and direction of the association between two variables. Does knowing score of one variable help tell us the score of the other variable: ex. direct/ indirect relationship. Correlation is not causation (only indicates that a relationship exists: multiple causal patterns could be possible, outside factors could be cause, can only really assess causality if you manipulate a variable. + 1 indicates perfect positive correlation (direct relationship) 0 indicates no correlation either way (this statistic assumes data is normally distributed and the variances are homogeneous) You must determine whether the r is statistically significant (i. e. , different from zero) Computer programs will print out the p-value, which tells you the probability the r is just a fluke (i. e. , not statistically significant) We continue to use the 95% confidence level for our conclusions, so we need p-values smaller than . 05.
