Scribd
Upload
91 views

Variance Lecture

This document provides an overview of analysis of variance (ANOVA) and multiple regression analysis. It discusses the underlying assumptions of ANOVA, including that populations must be normally distributed with equal variances. The ANOVA procedure and formulas are described. An example is provided to illustrate how to test for differences in three population means. Multiple regression analysis is introduced, including how to calculate regression coefficients and evaluate the regression equation. Assumptions, ANOVA tables, and tests for individual variables are covered. An example demonstrates how to develop a regression model and test hypotheses. Residual analysis techniques are also summarized.

Uploaded by

Lennard Pang
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
91 views

Variance Lecture

This document provides an overview of analysis of variance (ANOVA) and multiple regression analysis. It discusses the underlying assumptions of ANOVA, including that populations must be normally distributed with equal variances. The ANOVA procedure and formulas are described. An example is provided to illustrate how to test for differences in three population means. Multiple regression analysis is introduced, including how to calculate regression coefficients and evaluate the regression equation. Assumptions, ANOVA tables, and tests for individual variables are covered. An example demonstrates how to develop a regression model and test hypotheses. Residual analysis techniques are also summarized.

Uploaded by

Lennard Pang
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 14

5/29/09

1-1

Statistics 2 Supplementary lecture 2

Analysis of variance and


Multiple regression

10-7

Underlying Assumptions for ANOVA


The F distribution is also used for testing the
equality of more than two means using a
technique called analysis of variance (ANOVA).
ANOVA requires the following conditions:
The populations being sampled are normally
distributed.
The populations have equal standard
deviations.
The samples are randomly selected and are
independent.

5/29/09

10-8

Analysis of Variance Procedure


The Null Hypothesis: the population means are the
same.
The Alternative Hypothesis: at least one of the
means is different.
The Test Statistic: F=(between sample variance)/
(within sample variance).
Decision rule: For a given significance level ,
reject the null hypothesis if F (computed) is greater
than F (table) with numerator and denominator
degrees of freedom.

10-9

NOTE
If there are k populations sampled, then
the df (numerator) = k-1
If there are n sample points, df (denominator) =
n-k
The test statistic is : F=[(SST)/(k-1)]/[(SSE)/(n-k)].
SST is the treatment sum of squares.
SSE is the error sum of squares.
TC is the column totals.
nc is the number of observations in each column.
X is the sum of the observations.

5/29/09

10-10

Formulas

10-11

EXAMPLE 1
Rosenbaum Restaurants specialize in meals for
senior citizens and families. Katy Polsby,
President, recently developed a new meat loaf
dinner. Before making it a part of the regular menu
she decides to test it in several of her restaurants.
She would like to know if there is a difference in the
mean number of dinners sold per day at the
Sylvania, Perrysburg, and Point Place restaurants
for sample of five days. At the .05 significance level
can Katy conclude that there is a difference in the
mean number of meat loaf dinners sold per day at
the three restaurants?

5/29/09

10-12

EXAMPLE 1

continued

EXAMPLE 1

continued

10-13

Step 1:
Not all the means
are the same
Step 2: The significance level is .05.
Step 3: H0 is rejected if F>4.10
Step 4: From the table Katy determines SST=76.25,
SSE=9.75, and the test statistic: F=[76.25/2]/
[9.75/10]=39.10
Step 5: H0 is rejected. There is a difference in the mean
number of dinners sold

5/29/09

12-3

Multiple Regression Analysis


For two independent variables, the general
form of the multiple regression equation is:
X1 and X2 are the independent variables.
a is the Y-intercept.
b1 is the net change in Y for each unit change
in X1
holding X2 constant. It is called a partial
regression coefficient, a net regression
coefficient, or just a regression coefficient.

12-4

Multiple Regression Analysis


The general multiple regression with k independent
variables is given by:
The least squares criterion is used to develop this
equation.
Calculating b1, b2, etc. is very tedious. There are
many computer software packages that can be used
to estimate these parameters.

5/29/09

12-5

Multiple Standard Error of Estimate


The multiple standard error of estimate is a
measure of the effectiveness of the regression
equation.
It is measured in the same units as the dependent
variable.
It is difficult to determine what is a large value and
what is a small value of the standard error.

12-6

Multiple Standard Error of Estimate


The formula is:

5/29/09

12-7

Multiple Regression and Correlation


(Assumptions)
The independent variables and the dependent
variable have a linear relationship.
The dependent variable must be continuous and at
least interval-scale.
The variation in (Y-Y) or residual must be the same
for all values of Y. When this is the case, we say the
difference exhibits homoscedasticity.
The residuals should be normally distributed with
mean 0.
Successive values of the dependent variable must be
uncorrelated.

12-8

The ANOVA Table


The ANOVA table gives the variation in the
dependent variable (of both that which is and is
not explained by the regression equation).

5/29/09

12-9

Correlation Matrix
A correlation matrix is used to show all possible
simple correlation coefficients between all
variables.
The matrix is useful for locating correlated
independent variables.
How strongly each independent variable is
correlated to the dependent variable is shown
in the matrix.

12-10

Global Test
The global test is used to investigate whether any
of the independent variables have significant
coefficients. The hypotheses are:

5/29/09

12-11

Global Test

continued

The test statistic is the F distribution with k(number


of independent variables) and
n-(k+1) degrees of freedom, where n is the sample
size.

12-12

Test for Individual Variables


This test is used to determine which independent
variables have nonzero regression coefficients.
The variables that have zero regression coefficients
are usually dropped from the analysis.
The test statistic is the t distribution with
n-(k+1)degrees of freedom.

5/29/09

12-13

EXAMPLE 2
A market researcher for Super Dollar super markets
is studying the yearly amount families of four or
more spend on food. Three independent variables
are thought to be related to food expenditures.
Those variables are: total family income, size of
family, and whether the family has children in
college.

12-14

EXAMPLE 2

continued

10

5/29/09

12-15

EXAMPLE 2

continued

Use a computer software package, such as SPSS


or Excel, to develop a correlation matrix.
From the analysis provided by the computer, write
out the regression equation:
What food expenditure would you estimate for a
family of 4, with no college students, and an income
of $50,000?

12-16

EXAMPLE 2

continued

Y=954 + 10.9(50) + 748(4) + 565 (0) = 4491.


Conduct a global test of hypothesis to determine
if any of the regression coefficients are not zero.
H0 is rejected if F>4.07
From the output, the computed test statistic value
is 10.94
Decision: Since F=10.94>4.07, H0 is rejected.
Thus not all the regression coefficients are zero

11

5/29/09

12-17

EXAMPLE 2

continued

Conduct an individual test to determine which


coefficients are not zero.
From the output, the only significant variable is
FSIZE (family size) using the p-values. The
other variables can be omitted from the model.
Thus,
Using the 5% level of significance, reject H0 if
the
p-value<.05

12-18

EXAMPLE 2

continued

Since p-value =.039<.05, reject H0 and conclude


that
. That is, the size of the family and
the amount spent on food are significantly related.

12

5/29/09

12-19

Qualitative Variables & Stepwise


Regression
Qualitative variables are nonnumeric and are also
called dummy variables.
For a qualitative variable, there are only two conditions
possible.

Stepwise Regression leads to the most efficient


regression equation.
Only independent variables with significant regression
coefficients are entered into the analysis. Variables are
entered in the order in which they increase R^2 the fastest.

12-20

Analysis of Residuals
A residual is the difference between the actual value
of Y and the predicted value Y.
Residuals should be approximately normally
distributed. Histograms and stem-and-leaf charts
are useful in checking this requirement.
A plot of the residuals and their corresponding Y
values is used for showing that there are no trends
or patterns in the residuals.

13

5/29/09

12-21

Residual Plot

Residuals

1000
500
0
-500
4500

6000

7500

12-22

Frequency

Histograms of Residuals
8
7
6
5
4
3
2
1
0
-600

-200

200

600

1000

Residuals

14

You might also like