How To Interpret Multiple Regression Output in Spss
How To Interpret Multiple Regression Output in Spss
Company* Price (BDT) (x1) Sales (in units) (y) Brand Image (10)
(x2)
A 12 2,430 7
B 15 1,870 6
C 10 3,284 9
D 18 1,012 4
E 15 1,930 6
F 12 2,136 8
The table above shows the data on which multiple regressions will be run.
The dependent variable here is Sales, and the two independent variables are Price (X1) and Brand
Image (X2). We want to find out the relationship between Price and Brand Image to Sales.
So after running the regression in SPSS we get the following Output (this will be given in your exam
question):
Model Summary
a
ANOVA
Total 2776955.333 5
I will ask you people to interpret the results. So here I am giving you a sample interpretation:
Interpretation
Model Summary
Here we will focus only on the R square value. For an r square of 0.906, we can say that the model
explains 90.6% of the variations in real life and so the model is a good model.
a
ANOVA
Total 2776955.333 5
This table shows us the ANOVA results of the two independent variables X1 and X2. Remember
that the initial null hypothesis of ANOVA tells us that X1=X2. And the alternate hypothesis tells us
that X1 is not equal to X2.
In this table we look into two values F and the Sig (highlighted in yellow above). A high value of F
means that there are more chance of the Null Hypothesis being rejected and alternate accepted,
which means that X1 and X2 are different. Here it is 14.5, which means that the value is pretty high
and that X1 and X2 will be different. On the other hand, the significant tells us the confidence level
(1- Sig) of accepting the alternate hypothesis. Here the Sig is 0.029, which means that you are (1-
0.029 = 0.971) 97.1% confident that the alternate hypothesis is accepted, and that X1 is not equal to
X2.
So you can say from both the F value and the Sig value that the two variables are indeed different
from each other and that they affect the Sales (Y) in a different manner.
In case if you cannot confidently reject null or accept alternate, then you must say that the variables
X1 and X2 are indeed same and that multiple regression will not be possible in this case
a
Coefficients
The important things to note in the coefficients table are highlighted in yellow. These tell us the
structure of the model. The constant is the C, and then X1 is Price and X2 is Brand Image. So we can
write :
So price has a negative relationship with sales and that increase in 1 taka leads to a decrease in
sales by 288.8 units. On the other hand, Band Image has a positive relationship with sales, and that
1 increase in Brand Image leads to 376 increase in sales.