Select two of the following three models.

1. Logistic Regression

1. Collect data ( n >= 30 ) on a binary dependent variable and at least four
   explanatory variables. At least one of the explanatory variables has to be
   qualitative.
2. Do a logistic regression in Minitab.
3. Email the following in a Minitab Project File:

1. the data with all relevant sources,
2. the output of the logistic regression (in the Session Window),
3. a discussion about the statistical significance of each explanatory

variable and an explanation of the meaning of each relevant odds ratio (in

the Report Pad).

II. Weighted Least Squares

1. Collect cross-sectional data ( n >= 30 ) on a dependent variable and at least
   three explanatory variables. At least one of the explanatory variables has to be
   qualitative.
2. Estimate and choose the "best" least squares regression model by determining
   which is the best model (perform the tests learned in class, e.g., choose between a
   second-order model, interaction model, or first order model). You are free to use
   alternative model specifications, e.g., log-linear.
3. Estimate a weighted-least squares regression on the model chosen in "B."
4. Email the following in a Minitab Project File,

1. the data with all relevant sources,
2. the output associated with the process of choosing the best
   model and the residual plot for the relevant model (Session Window),
3. a discussion of which model was chosen and why (Report Pad),
4. the output of the weighted least squares (Session Window),
5. a discussion of the statistical significance of the model and the
   individual explanatory variables and an explanation of the meaning of
   each regression coefficient (Report Pad),
6. a discussion of the difference (if any) between the least squares

and weighted least squares regression estimates (Report Pad).

III . Serial Correlation

1. Collect time series data ( n >= 30 ) on a dependent variable and at least three
   explanatory variables.
2. Estimate and choose the "best" least squares regression model by determining
   which is the best model by performing the tests learned in class. You are free to
   use alternative model specifications, e.g., log-linear.
3. Estimate the model by allowing for autocorrelation.
4. Email the following in a Minitab Project File,

1. the data and all relevant sources,
2. the output associated with the process of choosing the best
   model and the residual plot for the selected model (Session Window),
3. a discussion of which model was chosen and why (Report Pad),
4. the output of the model corrected for serial correlation (Session
   Window).
5. a discussion of the statistical significance of the model and the
   individual explanatory variables and an explanation of the meaning of
   each regression coefficient (Report Pad),
6. a discussion of the difference (if any) between the least squares

and the regression estimates corrected for autocorrelation (Report Pad).

Price: $29.02

Solution: The downloadable solution consists of 17 pages, 1202 words and 24 charts.
Deliverable: Word Document