1/33
Looks like no tags are added yet.
Many statistical associations are that allow _.
overall tendencies; exceptions
Response or dependent variable:
measure outcome of a study.
Explanatory or independent variable:
explains or causes changes in response variable
Many explanatory-response relationships do not involve
direct causation; one may merely predict the other
In scatterplots, always plot the ___ variable on the horizontal x axis
explanatory
A positive association occurs when
low x values accompany low y values, high x values accompany high y values
Is correlation r resistant to outliers?
No
R value falls between
-1 and 1
R is only for ___ relationships - it cannot describe strength of __ relationships (even if there is one)
linear; nonlinear
R is for ___ variables
quantitative
R does not distinguish between
explanatory and response - just indicates strength and negative/positive nature of relationship
Regression line summarizes the ___ between two variables, but only when ____.
relationship; one variable predicts/explains the other
A regression line is a straight line that describes how ___. We often use the line to predict the value of ___.
response variable y changes as explanatory x variable changes; y for a given x
Fitting a regression line to the data means creating a line that comes ___. The line will of course not go __ through each point.
as close as possible to the points; perfectly
y = mx + b: What does each letter stand for?
y (or y-hat): expected y value. m = slope. x = explanatory variable. b = y-intercept
Slope
the amount by which y changes when x increases by one unit
Intercept
the value of y when x is zero
To make a prediction substitute a value of __ into y = mx + b to get the y
x
We can use the regression line to make a prediction about the likely value of y for a given x. We interpret this as the ____
average of y for an x of multiple cases/subjects, or the likely value of y for an x of one case/subject
Least squares regression line:
the line that makes the sum of the squares of the vertical distances of the data points from the line as small as possible.
You use __ and _ of y and x to calculate least squares regression line.
mean and SD
There is a close connection between the slope of least squares line and correlation. We use r in calculation of least squares line. Along the regression line a change of ____. When correlation is perfect the predicted response of y is the same (in SD units) as the change in ___.
1 standard deviation of x corresponds to r standard deviations in y; x
Regression line passes through the ___ of x and y
means
Least squares regression looks at distance of data points from the line in the y direction, so flipping the variables results in a ___
different line
R2 is the fraction of ___ that are explained by the least squares regression of y on x. In other words, r2 is a measure of how ___
variation in values of y; well the regression explains the response.
Extrapolation
making predictions far beyond the range of collected data. Can’t be trusted.
Residual =
distance from observed value of the response variable and the value predicted by regression line. (Observed y - predicted y (y - y-hat).)
A negative residual indicates it ___, while a positive residual indicates it ___.
falls below the regression line; falls above
A residual plot is a ___ against explanatory variable
scatterplot of residuals
In a residual plot, a scatter about the fit line indicates the line is ___. But a curved pattern can indicate ___ and increasing or decreasing spread indicates ___ in when y prediction is best predicted by x
a good fit; nonlinearity; variability
Influential points can ___. An extreme point pulls the line ___.
drastically change regression line if removed; towards itself
Are all outlier influential?
No, it depends on if they are influencing slope of the line
Which outliers are often influential?
x outliers
Are residuals necessarily large for influential outliers?
No