Skowhegan Area High School

Contemporary Mathematics In Context                                                                                      skowbutton msadbutton
Topics
Introduction
Spearman's Rank Correlation Coefficient
Scatterplot Matrix
Rank Correlation Matrix
Pearson's Correlation Coefficient
Influential Point
Lurking Variable
Cause and Effect

Least Squares Regression ("r" and residuals)

Patterns of Association

 

Influential Point

An influential point is a point that strongly influences the value of the correlation coefficient. That is, if the point is removed from the data set, the value of the correlation coefficient chages quite alot. Here is an example;
 

Now, look at the chart, do you see where the 8 year old car has a value of $14000? Notice how it doesn't really go along with the rest of the data, it is an anomoly. This would be our lurking variable because it strongly influences the value of the correlation coefficient.
Let's try substituting the values in our calculator (L1 and L2), the x values in L1, the y values in L2. Next press stat, scroll over to Calc, then go down to LinReg. Next, press 2nd L1 (,) 2nd L2. An equation for the best fit line will come up. Look at the bottom at your r value, you should have a value of -.787. Remember the (-) is just the direction of the graph, the closer to -1 the more related the graph is to the best fit line. Now, let us take out our influential point, completely erase it from the data, now do the same as above and look at your r value. It is now -.958. As you can see this is a huge difference. When looking at data and at correlation values, you must keep an eye out for data like the influential point, and decide for yourself whether that data should be regarded as an anomoly or as a legitimate piece of data.

 

 

 

Lambourghini Murcielago