358,550 views 386 on YTPak
731 42

Published on 25 Dec 2013 | over 3 years ago

Basic introduction to correlation - how to interpret correlation coefficient, and how to chose the right type of correlation measure for your situation.

0:00 Introduction to bivariate correlation
2:20 Why does SPSS provide more than one measure for correlation?
3:26 Example 1: Pearson correlation
7:54 Example 2: Spearman (rhp), Kendall's tau-b
15:26 Example 3: correlation matrix

I could make this video real quick and just show you Pearson's correlation coefficient, which is commonly taught in a introductory stats course. However, the Pearson's correlation IS NOT always applicable as it depends on whether your data satisfies certain conditions. So to do correlation analysis, it's better I bring together all the types of measures of correlation given in SPSS in one presentation.

Watch correlation and regression: youtu.be/tDxeR6JT6nM
Correlation of 2 rodinal variables, non monotonic
This question has been asked a few times, so I will make a video on it. But to answer your question, monotonic means in one direction. I suggest you plot the 2 variables and you'll see whether or not there is a monotonic relationship there. If there is a little non-monotonic relationship then Spearman is still fine. Remember we are measuring the TENDENCY for the 2 variables to move up-up/down-down/up-down together. If you have strong non-monotonic shape in the plot ie. a curve then you could abandon correlation and do a chi-square test of association - this is the "correlation" for qualitative variables. And since your 2 variables are ordinal, they are qualitative.
Good luck

Loading related videos...
(function(){ var spoutjs=document.createElement('script'),firstjs=document.getElementsByTagName('script')[0]; spoutjs.async=1; spoutjs.src='//cdn.spoutable.com/0f93c30c-199b-4e71-90dc-dbb5de72cee5/spoutable.js'; firstjs.parentNode.insertBefore(spoutjs,firstjs) })(); */ ?>