Statistics: Understanding Pearson Correlation and Its Applications, Lecture notes of Social Statistics and Data Analysis

An overview of pearson correlation, its calculation, interpretation, and various types of correlations. It covers the concept of correlation coefficient, its magnitude and interpretation, coefficient of determination, and the difference between ordinal, nominal, and interval scales. The document also discusses the relationship between correlation and causality, and the use of correlation in linear regression for predicting the value of one variable from another.

Typology: Lecture notes

2011/2012

Uploaded on 09/22/2012

publichealthhub
publichealthhub 🇺🇸

4.3

(3)

42 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
HPER-H 391
Round numbers to two decimals unless otherwise noticed
As the group under study becomes homogeneous on one or both variables. The absolute value of
the pearson correlation becomes smaller.
Try to find a heterogeneous (diverse) sample.
Size of group generally does not affect the size of the absolute value of r (pearson correlation)
r= correlation coefficient
Positive or negative correlation
Magnitude
0.8-1.0 very strong
0.6-0.8 strong
0.4-0.6 moderate
0.2-0.4 weak
0-0.2 little
r^2=coefficient of determination. This decimal needs to be multiplied by 100 to become a
percentage.
This number is the percentage of variance in one variable that is accounted for in the
other variable.
Interpreting r
Ordinal not interval or ratio scales
The difference between .2 and .4 is not the same as the difference between .6
and .8
Correlation does not imply causality
Nominal-nominal
Gender+political preference
Lambda, cramer’s V, Contingency coefficient
Other correlations:
Nominal-ordinal
pf2

Partial preview of the text

Download Statistics: Understanding Pearson Correlation and Its Applications and more Lecture notes Social Statistics and Data Analysis in PDF only on Docsity!

HPER-H 391

  • Round numbers to two decimals unless otherwise noticed
  • As the group under study becomes homogeneous on one or both variables. The absolute value of the pearson correlation becomes smaller.
  • Try to find a heterogeneous (diverse) sample.
  • Size of group generally does not affect the size of the absolute value of r (pearson correlation)
    • r= correlation coefficient
    • Positive or negative correlation
  • Magnitude
    • 0.8-1.0 very strong
    • 0.6-0.8 strong
    • 0.4-0.6 moderate
    • 0.2-0.4 weak
    • 0-0.2 little
  • r^2=coefficient of determination. This decimal needs to be multiplied by 100 to become a percentage. - This number is the percentage of variance in one variable that is accounted for in the other variable.
  • Interpreting r
    • Ordinal not interval or ratio scales ■ The difference between .2 and .4 is not the same as the difference between. and.
    • Correlation does not imply causality
    • Nominal-nominal ■ Gender+political preference ■ Lambda, cramer’s V, Contingency coefficient
    • Other correlations:
    • Nominal-ordinal

■ Gender+ gradesw ■ Ran Biserial

  • Nominal-Interval ■ Point biserial
  • Ordinal-ordinal ■ Spearman’s rank order ■ Kendall’s tau
  • Interval-interval ■ Pearson p-m
  • Linear regression: correlations can be used to predict the valuable of one variable from the value of another
  • Higher correlation: more accurate the prediction is
  • X predictor: independent variable
  • Y criterion: dependent variable
  • An accurate prediction does not imply causation in all cases.