Association, Correlation, and Linear Regression Flashcards
Scatter plot
- Variable x variable
- Both _
Quantitative
Association
- (as x increases y increases) or – (as x increases y decreases)
- _
- Linear
- Straight
- Curved
- _
- Strong
- Moderate
- Weak
- Interesting outliers
Form, trend
_ _
- Independent
- X axis
Explanatory variable
_ _
- Dependent
- Y axis
Response variable
Correlation
- Correlation coefficient (r) [-1,1]
- (_ slope)
- – (_ slope)
- 0
Positive, negative
Coefficient of determination (r^2) [0,1]
% of the variance in y can be explained by the linear regression of y and x
Prediction capability
Strong
Weak
Moderate
Slope
rSy/Sx
- Line of best fit
- Least residual
- Best predictor
- Stat, calc, linear regression
- Carrot over predicted in model
Least squares regression line
Linear
- Scatter plot looks _
- _ pattern in residual plot
- Lresid x Ln
Linear, no
Make linear
- Try pairs
- Translate
- If x is time then try _
- If both not time then try _
- Check residual plot for no pattern
- Make sure to put revisions in place of y and x when describing the context of r^2
Log(y), log(x) log(y)
Data outside of range
Extrapolation
X value far from average x
High leverage
Omitting it would give a different model
Influential/outlier
Alternative variable
Lurking variable
Regression is not how y changes when x changes it’s just a model so you can’t say that they are _.
Related
Residual
Actual-predicted
Do not round just put a and b in calculations.
Tip
Pay attention to a negative slop.
Tip
Make sure if log(y) then you do _.
10^(mx+b)
