Instructor: Larry Musoino! "mai: musoino#northampton$edu Chapter 4 Summary Concept %hen to use Procedure Scatter &iagram Page 1'( You want to create a graph which plots the (x, y) datavalues. The scatter diagram allows you to determine if there is a linear relationship between x and y. 1. Set up a graph consisting of x and y axes. . Scan the x and y data values to determine an appropriate scale for each axis. !. "lot the independent (explanatory) variable on the hori#ontal x$axis. %. "lot the dependent (response) variable on the vertical y$axis. &. 'or each (x, y) pair, plot a point at the corresponding (x, y) location on the graph. (. )etermine if the resulting (x, y) points generally align along a straight line, that is, would a straight line form a reasonably good fit to the datapoints. Sampe correation coe))icient! r Page 1'* You want to determine a numerical measurement which assesses the strength of the relationship between x and y. 1. The sample correlation coefficient, r, always falls in the range from $1 to *1, inclusive. . +f r , *1, there is perfect positive linear correlation. !. +f r , -, there is no linear correlation. %. +f r , $1, there is perfect negative linear correlation. &. The closer that r is to *1 or $1, the better the linear fit to the data. (. .alues of r close to - indicate a poor linear fit to the data. /omputation 'ormula for r (note this can be determined using your calculator)0 +est ,it Line -Least S.uares Line/ Page 151 You want to create a best fit line to a set of (x, y) data where a linear model is a good fit to the data. The linear model will be of the form0 y = slope(x) + yintercept +n the notation below, 1b2 is the slope and 1a2 is the y$intercept but your calculator may use a different notation so beware 33 MAT150 Chapter 4 Summary Page ( Ma0ing Predictions using the 1est )it ine Page 15* You want to ma4e a prediction for the dependent (response) variable based on a given x$ value. 1. /onstruct the best fit line0 y = slope(x) + yintercept 2. 'or a given x$value, substitute this x$value in the e5uation from step 1. 3. /alculate a value for y based on this e5uation. 6ote0 You should only use x$values for predictions based on the range of observed x$values. 'or example if the observed x$values ranged from & to &, then you should generate predictions only for x$values in this range between & and &.