AP Stats Unit 2 - LSRL interpretations

correlation (r)

There is a (DUFS) association between [x] and [y].

• Direction (+/-), Unusual features (outliers), Form(linear or non-linear), Strength (close to pattern)

coefficient of determination (r²)

About r²% of the variation in the [response variable (y)] is explained by the [explanatory variable (x)].

Slope of LSRL (b)

For each additional [x], our model predicts the [y] will [increases/decreases] by [b].

y-intercept of LSRL (a)

When [x-variable = 0], our model predicts [y] to be [a].

Standard deviation of the residuals (s)

The actual [y] is typically about [s] away from the number predicted by the LSRL.

residuals (y - ŷ)

residual < 0: prediction was too high

residual > 0: prediction was too low