How to Complete a Regression Analysis in Minitab 18
Why We Use Regression Analysis
When you look at a scatter plot graph created from the plotting of dots from two different axes, you will find that the variables are inversely related, directly related, or show no relation at all. If the line you draw to estimate regression seems to be moving from left to the upper right corner of the scatter plot the data is said to be directly related and inversely related if moving from left to the lower right corner. When the data plots Are spread out evenly with no apparent direction there is no relationship. It‘s nice to have a quick view of the scatter plot, but a relationship between variables can be more accurately determined through regression analysis. In this tutorial a scattered plot with a regression line will be created. To follow along with this lesson download this Minitab file. Also, if you don't have the latest version of Minitab you can download a free trial for the new version here.
Add Data to Minitab
To add data to Minitab the data must be entered or pasted into the program from a spreadsheet. The data should come in the form of X,Y separately to be able to complete an analysis.
Set-up Scatter Plot With a Regression
The first graph that we need to bring up for a regression analysis is a scattered plot graph. To set-up this graph click Graph→Scatterplot. When the scatterplot window appears select the box labeled "With Regression" and click OK. When the scattered plot with regression window appears put the cursor in the first row of the Y-axis box, then double click on the Y variable to the left. Next, double click on the X variable and it will populate the X section. A default name will appear if you do not create a title so I'm going to create my own title by clicking on "Labels." Then click in the title text box and type "Scatter Plot Graph of Cars Sold VS. TV Ads" and click OK. Click the OK button again and the scatter plot diagram with regression will appear.
Click the Tab Graph and Select Scattered Plot
Select With Regression
Set-up Fit Regression Model
To set-up additional regression graphs click Data→Regression→Regression→Fit Regression Model. Now you must place your cursor in the "Responses" section and click on the cars sold header (Y variable) to the left. Click in the "Continuous predictors" section and then click on TV Ads header (X variable). Find the storage button and click on it. From the check boxes select Fits, Standardized residuals, and Coefficients. Click Okay. There one more task that will allow us to show multiple residual graphs. Click on Regression Graphs and choose the option "Four in one." Now click Okay. Then click OK again.
Click Data, Regression, Regression, Fit Regression Model
Scattered Plot, Residual Plots, and Regression Output Data
The scatter plot shows us that the number of cars sold is directly related to the amount of TV adverting. This can be visibly seen without having the regression display in the graph. The residual plots graphically display difference between the observed value of the dependent variable (y) and the predicted value (x). And finally, the output data shows a numerical analysis of the variance.
Boyer, K. & Verma, R. (2010). Operations & supply chain management for the 21st century. Mason, OH: South-Western.
© 2018 Joshua Crowder