Friday, September 11th, 2009

Today's assignment :
Correct p. 21-26
Do p. 30-32

Homework :
Finish p. 30-32

The correlation coefficient spreadsheet (using Open Office Calc) is available at :
http://metisbeachschool.wikispaces.com/Mr.+Degrace

You can try with up to 10 set of variables. Try a few to see how the coefficient is influenced by the new data.

Mickael asked me at the end of the period how you can approach n. 4 on p. 31.

There can be more than one approach, but here is the one I came up with.


We know that estimating a linear correlation using measurements involves the following equation :

r = +/- ( 1 - length of shorter side / length of longer side )

For a) -0.2 , the first thing you can do is ignore the sign. The sign is used only after, depending on the type of slope your scatter plot takes.

So, we are looking for the following equation, knowing that r = 0.2.

0.2 = 1 - length of shorter side / length of longer side

You can decide any length for this exercise, I will say that my longer side is 5 cm.

I can replace my unknown length of longer side with 5.

0.2 = 1 - length of shorter side / 5

There is now only one unknown variable. Isolate it and you will have your two lengths :

0.2 - 1 = 1 - 1 - length of shorter side / 5

0.2 - 1 = 1 - 1 - length of shorter side / 5

-0.8 = - length of shorter side / 5

If you divide both sides by -1, you get :

0.8 = length of shorter side / 5

So multiply both sides by 5 and you're done !

0.8 * 5 = length of shorter side * 5 / 5
0.8 * 5 = length of shorter side * 5 / 5

0.8 * 5 = length of shorter side

In my example, my longer side = 5 and my shorter side = 4

Using the value 8 for longer side your result will be 6.4 for the shorter side.

Hope this helps. Reply if you have any questions.