Minitab Data Project (Part 1) for STAT 108 Honors

Describe the dependent variable (also called the response variable) and the predictor variables (also called independent variables) and what you are trying to do

including how many observations you have and how many independent variables you have

Which independent variables would you expect to have an impact (and which less of an impact) on predicting the dependent variable

Which independent variables are quantitative and which are qualitative

Make a histogram of each of your quantitative variables (including your dependent variables)

Do a boxplot of all your quantitative variables on the same plot (that is all boxplots together).

What is the shape of your quantitative variables (including your dependent variable).

Do any of your quantitative variables look approximately bell shaped?

Do descriptive statistics on your quantitative variables (min, Q1, median, Q3, max, mean, mode, variance, standard deviation, IQR)

Are there any outliers in your quantitative variables (show why or why not)

For qualitative variables do a frequency distribution table.

Make a barchart (or pie chart) of your qualitative variables.

Try making a new variable which is made by binning (change a quantitative variable to a quantitative variable) your dependent variable.

Minitab Data Project (Part 2) for STAT 108 Honors

Do a 2 sample hypothesis test

 use dependent variable and break into 2 parts based on a categorical independent variable (low, high)

 then test if mu low of dependent variable = mu high of dependent variable at alpha=0.05 and at alpha = 0.01

 would it make more sense to do a left tailed test or right tailed test instead of a 2 tailed test?

Interpret your results.

Do a correlation between 2 continuous independent variables (or a continuos independent variable and the continuous dependent variable)

What does the correlation say about the relationship between the two variables, interpret your results?

Do a chi squared test of independence between 2 categorical variables.  If you have only 1 categorical variable then bin one of your continuous variables into a categorical variable (such as low/high, or low/medium/high). Interpret your results?

Do a sinple regression using your continuous dependent variable and only one of your continuous independent variables. 

Now, try another simple regression with only one continuous independent variable (but a different independent variable than before)

Interpret your results.

If you want, also try using 2 continuous independent variables in the same regresion model (multiple regression) and interpret your results.