POL 322 Lec Dec 3, 2012
Correlation and Regression
line of best fit
-lease squares line
-like SD, if we add up all diff’s from the line = 0
-so we square the diffs to rid it of negatives
-line allows us to predict beyond sample data
-accuracy increases as strength of relationship increases
-direction: are IV’s positive or negative?
-are they significant in anyway?
-base significance on t-scores, more than 1.96 is significant
look at paper’s theme
look at direction of relationships after a multivar