BIOSTATS- CORRELATION AND REGRESSION

Correlation

testing for RELATIONSHIPS between variables

ANOVA

Comparing MEANS/DIFFERENCES between effects of variables

Regression

Testing for DEPENDENCE ON INDEPENDENT VARIABLE

Correlation does not

look into effect In c

In correlation ,

variables are numeric

Height and weight

example of correlation

As X increases

Y will decrease

As X decreases

Y will decrease

r

Pearsons product of movement, correlation coefficient

r ranges from

-1 to 1.0

Negative 1 means

negative correlation

Positive 1 means

positive correlation

• Less scattered

  • positive correlation if significant

Stronger relationship

P>0.05

no correlation

There is

no drawing in conclusions

Regression does not

have a correlation with correlation

Y= a+bx

Y is dependent variable, X is independent variable

a is

y intercept b

b is

slope of the line

if statistically significant,, equation for the best fitting line can be

used for prediction of Y based on X

A flat slope line means that

Slope will be 0

R²

ranges from 0 to 1

R^ means

proportion of variance in y variable is explained by variations of X

If R is 0.65 then

65% of variance