Data transformations.

[EagleWings]

Ok I know this might be a silly question but I don't know how to do it...
When it comes to data transformations I understand that you need to match the graph given with the suitable quadrant on the data transformation circle thing. The problem is, how do you know which transformation within that quadrant to make? For example, say I needed to use the top right quadrant, how would I choose between y squared and x squared?
Cheers:)

[Shandbagers]

You apply all of the transformations in the spreadsheet (Cas)
Then do linreg, the transformation which gives you the highest r and r^2 value is the one you want.
Also double check that it actually is linear by doing a residual plot.
Hope that helps!

[EagleWings]

Ok cool, thank you:)

[Shandbagers]

No wuckers

[EagleWings]

Ok I'm trying to get a residual plot up on the calc but when I press analyze in the menu it doesn't want to give me the option of pressing residuals:/ is there any other way to do it?

[Shandbagers]

You have to do a regression line first

[EagleWings]

Oh whoops

[EagleWings]

Thanks:D

[Shandbagers]

np

[Yacoubb]

You apply all of the transformations in the spreadsheet (Cas)
Then do linreg, the transformation which gives you the highest r and r^2 value is the one you want.
Also double check that it actually is linear by doing a residual plot.
Hope that helps!

Don't say the largest r value. Because, if the two variables are negatively associated, you're going to want an r value closest to -1. So, you'd pick the transformation that yields:
~ A correlation coefficient closest to 1 (if linear relationship is positive) or -1 (if linear relationship is negative).
~ A coefficient of determination closest to 1.

^^ Assessors can get extremely picky with expression for these types of questions, so just make sure your wording is accurate

[EagleWings]

Ok got it.Thank you!:)