Exercise on Significance

• What do I do about that? Test for significance.
• Research Methods Knowledge Base says to do a two tailed test, which I couldn't figure out quickly.
• Firstly, this is really asking the question how likely is it that this correlation is just random.
• Trawling through the socialreserachmethods page, I note that they pick alpha as .05. That does seem to be a standard significance level.
• They say to use a two tailed test, which I think has to do with a normal distribution both ways.
• They have degrees of freedom = 20-2=18 in their example because there are 20 pieces of data, and I think 2 variables.
• So in our data we have 7-2=5 degrees of freedom.
• My stats book doesn't have their table.
• I googled significance of correlation and came up with davidmlane.com
• This required me to find Pearson's r. I followed the links and got the formula.
• I then looked in Excel and used the function and got 0.584033704
• I then used Lane's t formula t=((r(sqrt(N-2)))/(1-r*r)), plugged it into excel and got t=1.608835467
• I then looked at Lane's table and found the magic number of 2.5706. Since 1.608835467 < 2.5706, this correlation is not significant at .05