Chap 5 There are many correlation coefficients. There are coefficients for two interval/ratio variables, ordinal variables, even coefficients for a combination of an interval variable with a nominal variable. The Pearson r is a common coefficient used for two interval/ratio variables. A negative sign indicates an inverse relationship and a positive sign a direct relationship. The strength or magnitude of the relationship has nothing to do with the signonly whether the coefficient is zero or low (like .2) or high (like .8). There is a formula in your book for computing the Pearson r. A Pearson r calculated on a sample must undergo a hypothesis test in order to determine whether this relationship holds for the population. Lastly, remember that correlation does not imply causation. Discuss. Chap. 16 Regression is related closely to correlation. Remember those scatterplots of two variables? Do they look like a line is a good representation of the shape? If so, then a Pearson r linear correlation will probably be a good fitwe could model the relationship by drawing a straight line through the data. The correlation will be high if the points cluster closely to this line, low if they do not. Any straight line has a formula of the form Y=slope X + y intercept. The regression program (or the formulas if you are doing it by hand) will give you the value of the slope and intercept. Lets say the slope is 3 and intercept 4, then the equation of the best fitting line is: Y= 3X + 4. This regression equation is very useful because it enables us to make predictions. What if we collected sales data for three years and found the best fitting regression line through the scatterplotwe could extend this line out into the future and make predictions for the next year (assuming the trend continues). In the equation above, if X is the third month of the year, March, our prediction would be Y (sales) = 3 X 3 +4 = 13. Discuss. Questions? P http://www.studentsoffortunes.com/
