Pages

Monday, October 11, 2010

Welcome to my hell

Sig. 2 tailed e.g. 0.000000000007% (p value) chance that there is a type 1 error, so we would be confident that there is a significant correlation between these two variables. anything over 5% (0.05) is too high, not a significant correlation. e.g. if p= .183, it would be 18% which is too high, and it’s a weak correlation. the closer to 0 it is, the more significant the correlation is.

probability value = 0.000007 or 0.999993 chance that there is a correlation

regression analysis is part of correlation

R square is how similar the two variables are, how much they overlap, e.g. venn diagram

constant (B) refers to the y intercept, and slope of the line underneath

beta is standardized, if we used z scores, same as R, regression not useful if you have just 2 variables


I took these notes in class last week and now have no idea what any of it means.

2 comments:

  1. I can't say that I'm a stats whizz, but I know what that all means. Flick me an email if you'd like any advice :)

    ReplyDelete
  2. Thanks for the offer. I should be okay, I'll just hassle my lecturer about it.

    ReplyDelete