Advice: Linear Regression and Correlation


Interval (normal) vs. Interval (normal)


SIMPLE LINEAR REGRESSION AND PEARSON'S PRODUCT MOMENT CORRELATION: These are listed under the regression and correlation section of the analysis menu. Regression provides information about the nature of the relationship between the variables but correlation describes just the strength of the relationship. The residuals about the regression line must be from a normal distribution and have a uniform variance. The hypothesis test for correlation assumes normality of one variable but the correlation coefficient requires that both variables are from a normal distribution.





Copyright © 2000-2017 StatsDirect Limited, all rights reserved. Download a free trial here.