In: Economics
Just the Fax, Inc. (JTF) hired a consultant to estimate the demand for its line of telecommunications devices in 35 different market areas for one month. The available data set includes observations on the number of thousands of units sold by JTF per month (QX), the price per unit charged by JTF (PX), the average unit price of competing brands (PZ), monthly advertising expenditures by JTF (A), and average gross sales (in thousands of dollars) of businesses in the market area (I). The consultant provided the results below without the estimated coefficients. However, she provided the following:
- Standard errors (in parenthesis) - t-stats [in square
brackets]
- R square.
QX=b0 +b1PX +b2PZ +b3A+b4I (250) (1.4) (0.8) (0.05) (0.04) [1.5] [-2.5] [1.5] [3.0] [2.5]
R2 = 0.80
JTF has hired you as a second consultant to produce the
estimated coefficients, and to determine whether they are
statistically significant.
Answer the following questions using the table below:
i. Calculate the missing coefficients for the intercept and the four independent variables in column 2 of the table below.
ii. If the critical t value at 5% significance level is 2.04, indicate in column 3 whether the calculated coefficients are significantly different from zero.
Variable |
Coefficient |
Significant/Not Significant |
Intercept |
||
Px |
||
Pz |
||
A |
||
I |
iii. What is the interpretation of the R2 value of 0.80?
i)
Estimated value of bo=t-stat*Standard error=1.5*250=375
Estimated value of b1=t-stat*Standard error=-2.5*1.4=-3.5
Estimated value of b2=t-stat*Standard error=1.5*0.8=1.2
Estimated value of b3=t-stat*Standard error=3*0.05=0.15
Estimated value of b4=t-stat*Standard error=2.5*0.04=0.1
ii)
Absolute value of t for observed value of bo is 1.5 which is lower than critical value of t i.e. 2.05. We can say that intercept is not statistically significant.
Absolute value of t for observed value of b1 is -2.5 which is higher than critical value of t i.e. 2.05. We can say that there is a significant correlation between Qx and Px. Observed value of b1 is statistically significant.
Absolute value of t for observed value of b2 is 1.5 which is lower than critical value of t i.e. 2.05. We can say that there is not a significant correlation between Qx and PZ. Observed value of b2 is not statistically significant.
Absolute value of t for observed value of b3 is 3 which is higher than critical value of t i.e. 2.05. We can say that there is a significant correlation between Qx and A. Observed value of b3 is statistically significant.
Absolute value of t for observed value of b4 is 2.5 which is higher than critical value of t i.e. 2.05. We can say that there is a significant correlation between Qx and I. Observed value of b3 is statistically significant.
Variable |
Coefficient |
Significant/Not significant |
Intercept |
375 |
Not significant |
Px |
-3.5 |
Significant |
PZ |
1.2 |
Not significant |
A |
0.15 |
Significant |
I |
0.1 |
Significant |
iii)
We get coefficient of determination equal to 0.80.
It means that the model explains 80% variability of the response data around its mean.