1/39
Looks like no tags are added yet.
What is correlation (r)?
Measures the strength and direction of a linear relationship.
What are the possible values of r?
'-1 to +1, inclusive.
What does a positive correlation mean?
As x increases, y tends to increase.
What does a negative correlation mean?
As x increases, y tends to decrease.
What does r close to +1 or -1 indicate?
A strong linear relationship.
What does r close to 0 indicate?
A weak or nonexistent linear relationship.
Does correlation prove causation?
No. Association does not establish cause and effect.
What is an outlier's effect on correlation?
It can greatly strengthen, weaken, or change r.
Can correlation measure nonlinear relationships?
No. It measures only the strength of linear relationships.
What does R² represent?
The proportion of variation in y explained by the linear regression model.
How do you convert R² into a percentage?
Multiply R² by 100.
What is the difference between r and R²?
r gives direction and strength; R² gives explained variation.
What is linear regression?
A method for modeling a linear relationship between two variables.
What is the least-squares regression line (LSRL)?
The line that minimizes the sum of squared residuals.
What is the regression equation?
Å· = bâ‚€ + bâ‚x
What does bâ‚€ represent?
The y-intercept; predicted y when x = 0.
What does bâ‚ represent?
The slope; predicted change in y for each 1-unit increase in x.
How do you interpret the slope?
For every 1-unit increase in x, predicted y changes by bâ‚ units.
What does the y-intercept tell you?
The predicted y-value when x equals zero, if meaningful.
What are the LSRL's two key properties?
It passes through (x̄, ȳ) and minimizes squared residuals.
What is a residual?
Observed y minus predicted y: y âˆ' Å·.
What does a positive residual mean?
The actual value is above the regression line; the model underpredicted.
What does a negative residual mean?
The actual value is below the regression line; the model overpredicted.
What does a residual of zero mean?
The predicted value equals the observed value.
What is a residual plot?
A graph of residuals versus x or predicted y-values.
What should a good residual plot look like?
Random scatter around zero with no clear pattern.
What does a curved residual pattern indicate?
A linear model may not be appropriate; the relationship may be nonlinear.
What does a fan-shaped residual plot indicate?
Nonconstant variability; the equal-spread condition may be violated.
What does a residual plot with no pattern suggest?
A linear model may be appropriate.
How do you predict a y-value using regression?
Substitute the given x-value into Å· = bâ‚€ + bâ‚x.
What is interpolation?
Predicting within the range of observed x-values.
What is extrapolation?
Predicting outside the observed x-range; often unreliable.
Why is extrapolation dangerous?
The existing relationship may not continue beyond observed data.
What is re-expression (transformation)?
Changing a variable's scale to make a relationship more linear or improve a model.
What is the goal of re-expression?
To improve linearity, stabilize spread, and make a model more appropriate.
What is the Ladder of Powers?
A guide for choosing transformations using different mathematical powers.
What transformation is often useful for count data?
Square root transformation (√y).
What does the power of zero represent on the Ladder?
The logarithm transformation (log y).
When might a logarithmic transformation help?
When data are strongly right-skewed or grow exponentially.
What must you remember after transforming data?
Interpret results in the transformed scale; predictions may need back-transformation.