Module 10: Correlation and Regression Flashcards
Correlation test use
-evaluate whether there is an association between 2 numerical variables
-whether one variable trends up or down as the other changes
Correlation coefficient values
-0 means no association
-1 means positive association
- -1 means negative association
Correlation test null and alternative
-Ho: p = 0
-Ha: p =/ 0
Correlation test null distribution shape
-t distribution
Correlation test t score calculation
-To: r - p / SE
-SE: square root of 1 - r^2 / df
Correlation test degrees of freedom
-n - 2
-n is for sample size
Correlation test reporting
-symbol for test (r)
-degrees of freedom
-observed correlational value (2 decimal places)
-p value (5 decimal places)
Linear regression test use
-evaluate whether changes in one numerical variable can predict changes in a second numerical variable
Linear equation
- y = mx + b
-y is for response variable
-x is for predictor variable
-m is for slope
-b is for intercept
Statistical model for linear regression
-systematic component: the slope line
-random component: standard error of each data point, mean changes but standard deviation does not
-link component: states that mean of normal distribution is the same as predicted variable from linear equation
Linear regression intercept null and alternative
- Ho: a = ba
- Ha: a =/ Ba
-a is for intercept
-Ba is for reference value
Linear regression slope null and alternative
- Ho: b = Bb
-Ha: b =/ Bb
-b is for slope
-Bb is for reference value
Linear regression null distribution shape
-t distribution
T score calculation for intercept
-T0: a - Ba / SE
T score calculation for slope
-To: b - Bb / SE
