ATAR Notes: Forum
VCE Stuff => VCE Mathematics => VCE Mathematics/Science/Technology => VCE Subjects + Help => VCE General & Further Mathematics => Topic started by: Ritzforlife on March 18, 2013, 06:03:53 pm
-
I'm currently doing my further maths SAC. We're required to create a scatterplot with the provided data. My correlation coefficient is a strong positive with r = 0.85
The next question requires us to do a transformation that linearises the data, however my problem is that the scatterplot is already linear and all the transformations all result in a weaker correlation and the coefficient of determination is always only around 60-78%.
Am I missing something? Do transformations always result in the correlation increasing?
Any help is appreciated?
-
That sounds a bit dodgy to me. Are you sure your data is linear? You might have potentially made a mistake, because it sounds like the data should be non-linear.
To answer your question, no, transformations do not always improve the linearity of data. Think about applying a squared transformation to a straight line, for example. :P
-
I know from my SAC and what my teacher said, they will most probably ask you to apply a transformation in order to get a stronger relationship. Rarely should you end up with a weaker correlation. Check all your working, you may have made some small error.