Professional Documents
Culture Documents
1. You want to see how the demand for your product changes when you change the price. You
can find the sales data below
Price
Sales
4528
4012
3445
3002
2420
1985
1506
1025
476
93
1
2
3
4
5
6
7
8
9
10
a.
b.
c.
d.
Regression: Sales
constant
Price
4972.2 -495.09091
coefficient
std error of coef 32.82706295.29056023
151.4665 -93.5801
t-ratio
0.0000% 0.0000%
p-value
-0.9995
beta-weight
10
8
2.3060
c. d. The value of R-squared statistic implies that 99.91% of variation of the demand is explained
by the price. This is very strong relationship. We can say that only thing that really matters in
determining the demand for your product is its price.
e. It is enough to check p-value of the Price coefficient. In this case we have to reject the
hypothesis that the Price coefficient is equal to zero at any reasonable significance level since the
p-value is close to 0.0000%.
2. You performed a regression of sales vs. advertising expenditures. In the dataset the
volume of sales was expressed in hundreds of units while the advertising expenditure was
in thousands of dollars. The results are presented in the table below.
a. Using only the results below, write down the regression equation.
b. What is the mean increase in sales if you spend extra thousand dollars on
advertising?
c. How much do you expect to spend on advertising in order to sell one more unit?
d. Suppose that the profit from each unit is equal to $5. Is it profitable to increase the
advertising expenditure?
e. What part of the variation in sales is explained by changes in the advertising
expenditure?
f. Do you find coefficients of the regression significant?
g. If you spend 10,000$ on advertising what will be the predicted volume of sales?
Regression: Sales
constant Advertising
4.35628507 1.82308202
Coefficient
std error of coef 0.69708473 0.11552581
6.2493
15.7807
t-ratio
0.0000% 0.0000%
p-value
0.9347
beta-weight
standard error of regression 2.04152815
87.37%
R-squared
87.02%
adjusted R-squared
number of observations
residual degrees of freedom
t-statistic for computing
95%-confidence intervals
38
36
2.0281
Solution
In the table we have results of the estimation of the following regression equation:
Sales = 0+1Advertising + Error
b. After spending additional $1000 on advertising the mean sales should go up by 182.31.
c. In order to increase the mean sales by one unit we need to spend additional 1000/182.31=5.49
dollars.
d. From the c. results we know that to sell one more unit on average we need to spend additional
$5.49 but we will get earn $5. Therefore the expected profit from increasing the advertising
expenditure is negative and it is not profitable to increase the advertising expenditure.
e. 87.37% of variation of the demand is explained by the variation in the advertising expenditure.
f. p-values for the constant and the Advertising coefficients are close to 0.0000%. Therefore we
can conclude that both of those coefficients are significantly different from zero.
g. We have to remember that in the dataset, the sales were expressed in hundreds of units while
the advertising expenditure was expressed in thousands of dollars. Using the estimated values of
the parameters we have
The predicted mean volume of sales is equal to 2258.73 units.
3. Influential wine critics such as Robert Parker publish their personal ratings of wines, and many
consumers pay close attention. Do these ratings affect the price? The data in this exercise are a sample
of ratings and prices found online at the Web site of an Internet wine merchant.
(a) Does the scatterplot of the price of wine on the rating suggest a linear or nonlinear relationship?
(b) Fit a linear regression equation to the data, regressing price on the rating. Does this fitted model
make substantive sense?
(c) Create a scatterplot for the log of the price on the rating. Does the relationship seem more suited to
regression?
(d) Fit a regression of the log of price on the rating. Does this model provide a better description of the
pattern in the data?
(e) Compare the fit of the two models to the data. Can you rely on summary statistics like r2 and se
Solutions
Quiz 2 Soltuion
Life Expectancy
An actuary wanted to develop a model to predict how long individuals will live. After consulting
a number of physicians, she collected the age at death (y), the average number of hours of
exercise per week (x1), the cholesterol level (x2), and the number of points that the individual's
blood pressure exceeded the recommended value (x3). A random sample of 40 individuals was
selected. The computer output of the multiple regression model is shown below.
THE REGRESSION EQUATION IS
yhat= 55.8 + 1.79x1 0.021x2 0.061x3
Predictor
Coef
StDev
Constant
55.8
11.8
4.729
x1
1.79
0.44
4.068
x2
0.021
0.011
1.909
x3
0.016
0.014
1.143
S = 9.47
RSq = 22.5%
ANALYSIS OF VARIANCE
Source of Variation
df
SS
MS
Regression
936
312
3.477
Error
36
3230
89.722
Total
39
4166
Is there enough evidence at the 5% significance level to infer that the cholesterol level and the
age at death are negatively linearly related? Explain with reasons
Comments
H0:2 =0 vs H1 : 2 < 0
Rejection region: t < t0.05,361.69
Test statistic: t = 1.909
Conclusion: Reject the null hypothesis. Yes, there enough evidence at the 5% significance level
to infer that the cholesterol level and the age at death are negatively linearly related.