Crest Toothpaste Case:

The data corresponding to this homework is given in the Minitab file named "Tpaste data.MTW". The data consists of weekly time-series (1958 through 1963, 276 weeks) with the following order of variables: Week/Time index (1,2, … , 276); Market share of CREST toothpaste (CRESTMS); Market share of COLGATE toothpaste (COLGTMS), Unit Price of CREST toothpaste (CRESTPR), Unit Price of COLGATE toothpaste (COLGTEPR). In this exam we will be interested in the Market share of CREST (the dependent variable, CRESTMS) and we will consider the other three as potential independent variables.

Define,

CRESTMS= Market share of CREST toothpaste (weekly),

COLGTMS= Market share of COLGATE toothpaste (weekly),

CRESTPR= Unit Price of CREST toothpaste (weekly),

COLGTEPR=Unit Price of COLGATE toothpaste (weekly).

1. . Provide a scatter plot matrix of all four variables.
2. . Identify and discuss possible linear relationships among all these (dependent and independent) variables.
3. . Use Stepwise regression method (based on α = 0.05 significance level), determine the best fitting model.
   You must provide correct Minitab stepwise regression outputs to support your decision.
   c-1. Please write down your model in its correct format.
4. . Investigate the overall significance of the best fitting model you find in step (c), list your hypotheses (the null and the alternate hypotheses) and the test statistics you use for making the statistical conclusion.
   (Note: If you are using Global F test, then you must run the regression model in its regular format in Minitab, which will provide F test statistics. Be sure to provide the Minitab outputs if you are using the F test.).
5. . What is the R2 of your best fitting model, and explain its meaning.
6. . Interpret the meaning of the partial regression coefficient estimate for independent variable CRESTPR.
7. . Using Minitab, obtain the predicted values of CRESTMS based on your model in (c), please provide an overlay plot of observed and predicted CRESTMS values.
8. . Save the residuals from your model in (c), and investigate whether the residuals is a white noise series (here you only need to test whether there exists autocorrelation). In so doing, show your Minitab ACF function outputs, list hypotheses you use and your test (use either Bartlett or Portmanteau test) statistics as well as your conclusions.

Price: $11.47

Solution: The downloadable solution consists of 6 pages, 547 words and 6 charts.
Deliverable: Word Document