B1.+Compute,+use+and+or+estimate+correlation+coefficients+(Association,+clouds+of+points+Spearman's+rank+correlation,+scatterplot+matrix,+Pearson's+correlation+coefficient)


 * 1. What you need to understand.**

Correlation coefficient = r A measure of the linear association between two variables. A number between -1 and 1 that tells how closely the points on a scatterplot cluster about the regression line. The correlation is negative if the cluster slopes from left down to right and the correlation is positive if the cluster slopes from the left up to the right.

You calculate r by going to data analysis under statistics at the top of the CPMP tools page. Then entering the data and putting it into a scatterplot. From the scatterplot, on the top of the page you click on the little bubble with the i in it then check the box for show correlation.



Scatterplot matrix - a matrix whose entries are scatterplots. For example by entering a matrice data into CPMP tools you can then plot it into a scatterplot to show it's correlation.



Describing Scatterplots - trend, strength, variability, pattern, influential point(s)


 * 2. Example problem.**




 * 3. Common mistakes or misunderstandings:**

NOT regression line.

Common mistakes: Changing the units does not effect the correlation. The correlation would stay the same because if the unit of the variable was changed that whole variable would be changed by the same amount. Therefore the correlation between the two variables will still have the same strength as it did with its initial units.

Sometimes it may be hard to judge the strength of the correlation, for example if there are a lot of data points.
 * 4. For more information:**


 * link on class website "Guess r"
 * http://sites.google.com/site/prhsmrbywater/guess-r
 * cpmp tools
 * unit 4 from book