SOAN 3120 Chapter Notes - Chapter 4, 5: Confounding, Standard Deviation, Total Variation
9 Dec 2015
School
Department
Course
Professor
Document Summary
Displaying relationships: scatterplots: most useful graph for displaying the relationship between two quantitative variables measured on the same individuals is a scatterplot, always plot the explanatory variable on the horizontal axis (x axis) Adding categorical variables to scatterplots: use a different plot colour or symbol for each category. The correlation r itself has no unit of measurement: the correlation r is always a number between -1 and 1. Values of r near 0 indicate a very weak linear relationship. Use r with caution when there are outliers. Because the outlier may extend the linear pattern, it increases the correlation: correlation is not a complete summary of two-variable data even when the relationship between variables is linear. Better to give the means and standard deviations of x and y with the correlation. Facts about least-squares regression: distinction between explanatory and response variables is essential in regression.
