Math 153- DB1, DDH PRACTICE MINITAB14 REGRESSION
9/05
The following data came from a statistics class. The questions were “How many hours do you plan to watch TV per week?” and “How many hours do you plan to study for this course per week?” The x variable is TV Hours and the y variable is Study Hours.
|
TVHrs |
StudyHrs |
|
70 |
0 |
|
3 |
3 |
|
2 |
3 |
|
8 |
2 |
|
7 |
14 |
|
5 |
4 |
|
13 |
6 |
|
4 |
4 |
|
10 |
2 |
|
1 |
4 |
|
7 |
1.5 |
|
50 |
3 |
|
5 |
5 |
|
2 |
7 |
|
10 |
4 |
|
|
|
1) Make a scatterplot. Graph, Scatterplot. Select the third option, With Regression. For the x-variable, select TVHrs, and for the y-variable, select StudyHrs. OK You can print the graph.
2) To find the correlation coefficient, Stat, Basic Statistics, Correlation. Select the two variables. OK The correlation coefficient is – 0.352.
3) To find the linear regression line, Stat, Regression, Regression. Select the response variable, which is the y variable or StudyHrs. Select the predictor variable, which is the x-variable or TVHrs. OK The equation is StudyHrs = 4.93 – 0.0579 TVHrs.
What is the slope of this line? ___________ What would be its interpretation?
4) Minitab gives the regression equation. Remember we are considering the y-variable StudyHrs and the x-variable TVHrs. In the following table, note the number under the column p (for p-value) and in the row for TVHrs.
If the p-value < .05, we can generalize from the
sample to the population.
If the p-value 
.05, we CANNOT
generalize from the sample to the population.
In this example, the p-value =
0.198, so we CANNOT say that in the population of Math153 students there is a
relationship between TVHrs and StudyHrs.
5)
Skip the Analysis of Variance section.
6)
Observation 5 is marked with an R, an observation with a
large standardized residual, i.e. it is far off the line. The number right in
front of it is a z-score and should be greater than 2 or less than –2. On the scatterplot, circle this
observation and write unusual
Minitab marks Observation 1 with an X, which denotes an observation whose X value gives it large influence. Circle this observation on the scatterplot and write “influential.”