Zur Startseite  
 
  LernStats    Glossary    Textbook    Imprint    
backIndexnext
 



Glossary

Clicking on the glossary keywords on the right will display a definition of the word in this box.

Please note that the glossary definitions at present are available in german language only.


Interdisziplinäres Zentrum für Hochschuldidaktik - IZHD, Hamburg

IEE
   / Home / LernStats / Correlation / Drawbacks of the Covariance - Deriving the Correlation Coefficient


LernStats
(Chapter 5 - Page 4 / 9)

Drawbacks of the Covariance - Correlation Coefficient r

The covariance cov(x;y) well describes the degree of variation of two variables X and Y, but it has a serious drawback:

it is dependent on the scale of the variables and their variance.

Especially in the social studies compareable measurement scales cannot always be found. As an example the values for the variables intelligence (0-35 points) and school performance (0-9 points) obviously differ much. We already know one measure to make the two compareable: the transformation to standard-z-values.

If we relate the covariance cov(x;y) to the product of the standard deviation of X and Y, the differences in scale and spread are compensated for.

This formula is equal to:

Both variables X and Y are being z-transformed here.

bzw.

The covariance of the z-transformed values of the variables X and Y look like this:



 
  LernStats    Glossary    Textbook    Imprint    
backIndexnext