I have a friend that owns a beverage stand.  He needs to find out on which days he sells the most lemonade so he does not waste lemons and Ice.  He notices that on really hot days he sells out of lemons and runs out of ice really quick.  On days where it is not so hot, he runs a risk of ice melting and lemon halves go bad.  Regression is defined by Welkowitz as; the line connecting each of the medians (Welkowitz, 2012).  I would predict that he would sell more lemonade on days where the temperature is above 80 degrees.  Using regression would allow me to “make predictions by measuring two variables and forming a prediction equation from that data and then using the equation to predict a future event” (Welkowitz, pg 255). In this scenario the variable being predicted or the criterion is sales of lemonade.  The predictor or the independent variable would be temperatures above 80 degrees.  Once me friend confirms the linear relationship, he can analyze the Pearson’s correlation coefficient which is (r).  This will state if the variables are correlated, and how strong the correlation is between them.  We can then take the correlation coefficient which is (r) and square the results (r2) to tell us how the variables relate to one-another.