Section 2.3 Solutions
2.42
- The
slope of the regression is 2.59. Hence, as the revenue of an NBA team
increases by 1 million, the predicted value of the team increases by 2.59
million.
- Predicted
value for Los Angeles Lakers:
million dollars. Then
the residual (i.e., the error) is 447 – 407.31 = 39.69 million dollars.
- Note
that
. Hence, the least-squares regression line
explains 85.84% of the variation in NBA team value. Because this value is
so high, we can conclude the regression line is pretty successful in its
predictions.
2.43
- For
each additional year, the predicted discharge of the Mississippi
River increases by 4.2255 cubic kilometers.
- Predicted
value for 1780:
cubic kilometers.
It’s impossible to discharge a negative amount of water, indicating that
this extrapolation is nonsense. The regression line should be used for
predictions only within the range of the data collected.
- Predicted
value for 1990:
cubic kilometers.
Then the residual (i.e., the error) is 680 – 617 = 63 cubic kilometers.
- There
are high spikes in the timeplot in the years
1973 and 1993, indicating an unusually high discharge (which would go
along with a flood).
2.55
We are interested in predicting a husband’s height from his
wife’s height. Hence, husband height is the response (y) variable and
wife height is the explanatory (x) variable. Then the slope of the
regression line is
and the y-intercept
is
Thus, the equation of
the regression line (shown in the graph below) is
For a woman 67 inches
tall, the predicted height of her husband is 33.67 + 0.54(67) = 69.85 inches.

2.56
The equation for the least-squares regression line is
(plugging
in the formula for the y-intercept).
Note what happens when we plug in
as out x-value:

Hence, when
is our x-value,
is our y-value. That is, the point
is always on the
least-square regression line.
2.58
Professor Friedman wants to predict final exam scores from
pre-exam totals. Hence, final exam score is the response (y) variable
and pre-exam total is the explanatory (x) variable.
- The
slope of the regression line is
and the y-intercept
is 
- For
Julie’s pre-exam total of 300, the predicted final exam score is 30.2 +
(0.16)(300) = 78.2.
- The
least-squares regression line only explains 36% of the variation in final
exam scores. This is not very high, so Julie’s final exam score may be
quite different from the predicted value. She has a legitimate argument.
2.59
Since
is 0.16, r is either 0.4 or –0.4. Since there is a
positive association between class attendance and grades, r must be 0.4.