PHORM Correlation

Correlation does not equal

causation

but it is necessary for causation

correlation measures

the degree to which the variables are related

Simple linear correlation

two continuous normally distributed variables

example: age and accommodation

Pearsons Correlation Coefficient

-1.0 > r < 1.0

0 means no correlation

-1.0 or +1.0 means perfect correlation- no spread from the average

Closer to +/-1.0, the tighter the spread

R (correlation coefficient) only gives the degree of correlation

Not the spread!

R does not tell us

if the correlation is statistically significant

(likelihood a bunch of points would just land there by chance)

if X causes Y

Given a point X, what is Y

Simple Linear Regression

regression analysis involving one independent variable and one dependent variable in which the relationship between the variables is approximated by a straight line

y= mx+b

we can say, at age 35 we expect 6D of AA

Need to test if r is statistically significant!

do a t test

Variance

r^2

represents the percent of all data points that can be attributed to the regression

Pearsons correlation coefficient is for ____ data

normal!

What are ways to see if a data set is normal

Skew

big difference between mean and median

Shapiro (small p means not normal)

Smirnoff

What is the non parametric equivalent of Pearson's

Spearman's rank formula

Spearman's rank formula

rank x and y

Logistic correlation

for data non linear (polynominal cirve) and non normally distributed spread far apart

conversts to linear

Multiple Regression

a statistical technique that includes two or more predictor variables in a prediction equation

overall model and each coefficient can be tested for its correlation and significant

What famous study used Multiple Regression

Glaucoma risk calculator

OHTS and EGPT

5 most highly correlated risk factors for glaucoma

age

c/d ratio

IOP

central corneal thickness (thicker - lower risk, thinner - higher risk)

pattern standard deviation on VF