View Single Post
Old September 16, 2018, 10:29 AM   #27
Glenn E. Meyer
Senior Member
 
Join Date: November 17, 2000
Posts: 20,064
No, that's not really an example of calculating a correlation.

You need at least three pairs of values. The degrees of freedom to establish significance for a regression is Number of Pairs - 2.

It is also true that you couldn't do a non parametric with just one observation pair.

We need some more households. If we could establish a predictive, significant relationship then experimental methods might show us causality.
__________________
NRA, TSRA, IDPA, NTI, Polite Soc. - Aux Armes, Citoyens
Glenn E. Meyer is offline  
 
Page generated in 0.02574 seconds with 8 queries