Chapter 11 Flashcards
What is the minimum level of measurement for correlation and regression?
Interval
When would use r correlation in regards to data?
When you want to know the linear correlation between variables
When would you use regression in regards to data?
When you want to use one variable to predict a score of another
What is the non directional null hypothesis equation for r?
H0: r = 0 ( = 0)
What is the alternate hypothesis equation for r?
HA: r ≠ 0 ( ≠ 0)
What is the directional positive null hypothesis equation for r?
H0: r ≤ 0 ( ≤ 0)
What is the directional positive alternate hypothesis equation for r?
HA: r > 0 ( > 0)
What is the directional negative null hypothesis equation for r?
H0: r ≥ 0 ( ≥ 0)
What is the directional negative alternate hypothesis equation for r?
HA: r < 0 ( < 0)
How can you tell if the correlation is positive/direct? (2 characteristics)
- If X and Y are both low or both high values (the same)
- typically the first points on the graph are on the bottom left and at the top right of the graph
How can you tell if the correlation is negative/inverse correlation? (2 characteristics)
- typically the first variable is on the top left and the bottom right
- values of X are low and the values of Y are high. Vice versa (opposite)
In general how to most outliers affect the r coefficient?
reduce the correlation coefficient
How do the outliers of X with the middle values of Y affect the r coefficient?
It affects the slope by pulling the regression line closer to the X values. Not the intercept
What is regression toward the mean?
variables much higher or lower than the mean are often much closer to the mean when measured a second time
What are the different effects of regression toward the mean? (4 things to know)
- Predicted Y values will not always be perfect unless r = ± 1
- Predicted Y values tend to be closer to the mean of Y than the observed values of X are to the mean of X
- Predicted Y values will have less variation than actual observed values of Y
- zY >zX THIS WILL NEVER HAPPEN
What is the residual? And what does it signify?
the vertical distance away from the predicted Y score and their actual Y. It signifies an error
How do you calculate the residual?
It is the difference between the predicted Y score and actual Y score
Predicted Y score - Actual Y score
What is the independent variable for correlation and regression?
It is X. It is on the X axis across the bottom
What is the dependant variable for correlation and regression?
It is Y. It is on the Y axis
How do you calculate predicted values for standardized variables (z scores) in correlation?
Y = intercept + slope times X
How do you know if the data on the scatterplot is strong or weak between variables?
Strong: The dots or data are close together or close to the line or close to 1
Weak: The dots or data are spread apart or far apart from the line
What is the rand for R squared?
0-1 IT CAN NEVER BE NEGATIVE
What is R squared?
It is where you square correlations to get the better more correct value for the correlation. THE VALUE WILL ALWAYS GET SMALLER WHEN YOU USE R SQUARED
How do the outliers of Y with the X having middle values affect the r coefficient?
the intercept by moving the entire line up or down without dramatically affecting the slope.
How do bivariate outliers affect the r coefficient?
It where there are outliers on the X and Y affect the intercept and the slope by moving/distorting them