Download as DOCX, PDF, TXT or read online from Scribd
Download as docx, pdf, or txt
You are on page 1/ 6
Question 1:
a. Run the regression using both R and Excel. Do you get the same results? Make sure
you include both outputs.
These two output are the same
b. Is the explanatory power of the model significantly greater than zero? Explain. Set the significant value alpha = 0.05 The p-value of this model < 0.05, so we reject the H0 and conclude that explanatory
power of the model greater than zero
c. Interpret the R2.
R2 is the fraction of the explained variation / total variation, in other words, how much do
we explain the dataset by applying this linear regression. In this model, multiple R2 is
0.4249, which means 42.49% data has been explained by this model. d. What could we do to make the R 2 higher? We can probably try to change this model from simple linear regression to multiple linear
regression, to have more factors to explain the variation.
e. If we increase task interdependence by 1 unit, what happens to team performance? Team performance will rise 1.242 units.
Question 2: a. According to the model, how much additional market value is added per each additional
100 square feet of living space?
$1,860 market value is added per each additional 100 square feet of living space
b. Test the null hypothesis that the square footage slope coefficient equals zero. Use a 1%
significance level. Do you conclude that square-footage has any effect on market value?
p-value for this regression is 0.000 < 0.01, so we reject the H0, and conclude that
square footage does have effect on market value.
c. A new property has become available and Ms. Walk wants to predict the market value
that particular property will be able to command. The home contains 900 square feet of
living space. Provide a prediction for Ms. Walk and an approximate 95% confidence
interval for the prediction.
Prediction: Y hat = 20.26 + 1.86*9 = $37k Approximate = CI(y hat +/- 2Se) = 37 +/- 2*4 = [29,45]
d. What can you say about the overall ability of the model to explain variation in market
values? Since R^2 = 0.56, which means 56% data variation has been successfully explained by
this model. This is OK but not excellent.
e. Ms. Logan has added a new room that measures 200 square feet to her house. The
room cost her $3000. She now is asking Ms. Walk to sell the house for her and knows
that she must pay Ms. Walk 6% commission (6% of the selling price). At the 95% level
of confidence, can we say that this investment will pay off?
CI = 23.98 +/- 8 = [15.98,31.98] 0.94*15.98 – 3 >0, so this investment will pay off
Question 3: a. Use regression to determine, is there a relationship between CO2 concentration and
temperature? From your model, how much does temperature change per thousand parts
per million concentrations change in CO2 concentration?
P value < significant level 0.05, so we reject the H0 and conclude that there is a
relationship between temperature and CO2 concentration.
For 1 positive unit change in per thousand parts per million concentrations in CO2
concentration, there will be a 9.6223 positive rise in Celsius degree in temperature
compared to the avg. temperature of 1910-2000.
b. CO2 concentration equaled 0.40871 thousand parts per million as of August 17, 20183.
What global temperature would you predict using your regression model? What is your
95% confidence interval for that temperature?
Prediction: 0.865 degree
95%CI = [0.802, 0.928]
c. Assume you are a government policy-maker and your goal is to decrease global
warming. Mindful that CO2 concentration has been increasing over the years, you are
preparing for a summit meeting among world leaders to discuss a target level for CO2 concentration at which the Earth would be relatively “safe”. Do your results convince you
that CO2 concentration is a cause of global warming? Explain. What more would you like
to see before taking action at the upcoming summit meeting in part b above?
CO2 could be a good cause of global warming, since the R2 shows that CO2 explains
more than 85% of data variation.
Question 4: a. Run the above regression by entering the data onto a .csv data file (open Excel, copy the
data to an empty sheet and just save the file as a CSV file) and then using R. Overall,
does the model fit the data well? Explain.
R2 = 0.9941, which means this model explains more than 99% variation of data, so we
think the model fit the data well.
b. Interpret b1, the estimated coefficient for β1 . In managerial terms, state what b 1 means. B1 means for the sample data we held, we predict that 1 unit increase of month will
increase the unit of sales by 45.275.
Managerial: after each month, the unit sales will increase by 45.275 units.
c. Use the estimated regression to forecast unit sales for February, 2018. Provide a
prediction and a 95% confidence interval around that prediction.
Prediction: 14*45.275 + 1388.846 = 2022.692
95% CI for Feb. 2018 = [2004.26,2041.124]
d. Use the estimated regression to forecast unit sales for February, 2019. Provide a
prediction and a 95% confidence interval around that prediction. How does the width
of this confidence interval compare to the width of the confidence interval in part c?
Why is this so?
95% CI for Feb. 2019 = [2521.02, 2610.958]
The width of the confidence interval is larger than that of part C. Because the confidence
interval for the independent variable “Month” is pretty high, like approximately 5. When
we turn the month from 14 to 26, it will be a large number to add up or subtract.
