F1 Multiple-Regression

Multiple Regression
1. To predict the value of one numerical dependent variable (Y) using more than one independent
variables (X1, X2…,Xn). To determine if the set of independent variables affects/influences your
dependent variable. In the analysis, we can also determine which independent variables can
significantly influence/ affect your dependent variable. Note: your independent variable can be
categorical.
2. Multiple regression equation with two independent variables:

Yî = b0 + b1X1i + b2X2i
Yî = predicted Y-value
X1i and X2i values of the independent variables
b1 and b2 are the regression coefficients/slopes
b1 is the slope of Y with variable X1 holding constant the effects of X2
b2 is the slope of Y with variable X2 holding constant the effects of X1
b0 is the y-intercept (estimated value of Y when all the independent variables are 0]
3. The regression coefficients in multiple regression are called net regression coefficients and
measure the mean change in Y per unit change in a particular X, holding constant the effect of the
other X variables.
4. Interpreting the regression coefficients (slopes) and the y-intercept.
î = 5837.52 - 53.2173X1i + 3.613X2i where :

Suppose: Y
Yî = predicted monthly number of Vitamax tablets sold for store i
X1i = price of Vitamax tablet (in cents) for store i
X2i = monthly in-store promotional expenditures (in pesos) for store i.
Interpretations:
i. y-intercept: 5837.52 vitamax tablets will be sold in a month if the price is P0.00 and the monthly
in-store promotional expenditures is P0.00.
ii. b1 = - 53.2173; for a given amount of monthly in-store promotional expenditure, the mean
number of vitamax tablets that can be sold in a month decreases by 53.2173 tablets per one
cent increase in the price of vitamax. You may think of it this way: You have two stores that
spend same amount for promotions, one increased the price by one cent while the other did
not. The store that increased its price by one cent will at the average have 53.2173 fewer sales
of vitamax tablets than the store that did not increase its price.
iii. b2 = +3.613; For a given price, the mean number of vitamax tablets that can be sold in a month
increases by 3.613 tablets per one peso increase in monthly in-store promotional expenditures.
You may interpret it this way also: You have two stores that sell vitamax tablets at the same
price, one increased its promotional expenditure by one peso while the other did not. The store
that increased its promotional expenditure by one peso will at the average have 3.613 more
sales of vitamax tables than the other store.
5. Coefficient of multiple determination

It represents the proportion of the variation in Y that is explained (attributable) by the set of
independent variables.
SSR
r2 = SSR = regression sum of squares
SST
SST = total sum of squares
6. Testing the significance of the overall multiple regression model
We use the overall F test to test for the significance of the overall multiple regression model.
This test determines whether there is a significant relationship between the dependent variable and
the entire set of independent variables.
Ho: No linear relationship between the dependent variable and the independent variables
Ha: Linear relationship between the dependent variable and at least one of the independent
variables.
MSR
F=
MSE
ANOVA table
Source Degrees of freedom Sum of squares Mean Square F

Regression k SSR MSR= SSR/k MSR/MSE
Error n-k-1 SSE MSE = SSE/n-k-1
Total n-1 SST
k= number of independent variables
7. Testing the significance of the regression coefficients (slopes) [ here we will be able to
determine which independent variable is related to your dependent variable (can
affect/influence your dependent variable).
b j− B j
tcomputed =
sbj
bj = slope of variable j with Y holding constant the effects of all other independent variables
sbj =standard error of the regression coefficient b j
t = test statistic for a t-distribution with n-k-1 degrees of freedom
k = number of independent variables in the regression equation
Bj = hypothesized value of the population slope for variable j, holding constant the effects of
all other independent variables. [this is equal to zero if we hypothesized that there is no
significant relationship between your dependent variable and you j independent
variable(Ho)].
b1 (slope of sales on price) = -53.217

Ho: Price does not affect sales.
Ha: Price affects sales.
Alpha= 0.05
t-test (two-tailed; df= n-k-1)
b j− B j −53.217 −0
tcomputed = = = -7.7664
sbj 6.8522
ttabular = 2.042
Alpha = .05; two tailed; df = n-k-1 = 34-2-1= 31
7.7664 > 2.042 reject Ho

Price affects sales. For every one peso increase in price, the sales decreases by 53.217 units
holding constant the effect of promotional expense.
b2 (slope of sales on promotional expenditure) =3.613
Ho: Promotional expense does not affect sales.

Ha: Promotional expense affects sales.
Alpha= 0.05
t-test (two-tailed; df= n-k-1)
b j− B j 3.613− 0
tcomputed = = = 5.273
sbj .6852
ttabular = 2.042
Alpha = .05; two tailed; df = n-k-1 = 34-2-1= 31
5.273 > 2.042 reject Ho
Conclusion: Promotional expense affects sales. That is, for every one peso increase in
promotional expense, the sales increases by 3.613 units holding constant the effect of price.

F1 Multiple-Regression

Uploaded by

Copyright:

Available Formats

F1 Multiple-Regression

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

F1 Multiple-Regression

Uploaded by

Copyright:

Available Formats

Multiple Regression

2. Multiple regression equation with two independent variables:

4. Interpreting the regression coefficients (slopes) and the y-intercept.

^i = 5837.52 - 53.2173X1i + 3.613X2i where :

5. Coefficient of multiple determination

Source Degrees of freedom Sum of squares Mean Square F

b1 (slope of sales on price) = -53.217

7.7664 > 2.042 reject Ho

b2 (slope of sales on promotional expenditure) =3.613

Ho: Promotional expense does not affect sales.

You might also like