Correlation and Regression

Correlation and Regression

1. Different types of correlations exist, perfect+ve, perfect-ve, moderate+ve and-ve, unreleated

2. 2 variables of co variance, X and Y

   • X increase= Y increase Covariance +ve.
   • If X decrease and Y increase= Covariance -ve
   • No relation of X and Y= Covariance is 0

3. +1=Perfect +ve correlation. -1= perfect +ve correlation.

   • 0=no correltaion

4. R values, 1=High correlation, <0.3=little or no correlation

   • measures strength of correlation

5. Correlation does not imply causality

6. Don't use regression if no problem.

7. Association of 2 linear correlations found by bivariate.

   • No extrapolation

8. Pearson correlation used, to find strength of relation (rvalue) and signif P value

   • R values range from -1 to +1
     • if value -ve x increases and y decreases
     • If value +ve X&Y increase
   • When reporting 3 things to mention, Direction, strength and significance
   • Pearson Correlation test, 1.Analyze, 2.Correlate 3. Bivariate, 4. transfer all to right 5.First value is R value and under it is P value

9. Regression, describes relationships fully

10. Don't use regression if no sig correlation

    • Multiple regression= 2 or more variables are predictors.

11. Curves to straight line, like line weaver burke plot= easy calculations