Professional Documents
Culture Documents
sums of squares
within groups
We can then extend those principles to a wide range of
applications.
sums of squares
between groups
sums of squares
within groups
degrees of freedom
We can then extend those principles to a wide range of
applications.
sums of squares
between groups
sums of squares
within groups
degrees of freedom
means square
We can then extend those principles to a wide range of
applications.
sums of squares
between groups
sums of squares
within groups
degrees of freedom
means square
sums of squares
within groups
degrees of freedom
means square
hypothesis testing
We can then extend those principles to a wide range of
applications.
sums of squares
between groups
sums of squares
within groups
degrees of freedom
one-way
means square
ANOVA
hypothesis testing
We can then extend those principles to a wide range of
applications.
sums of squares factorial
between groups
ANOVA
sums of squares
within groups
degrees of freedom
one-way
means square
ANOVA
hypothesis testing
We can then extend those principles to a wide range of
applications.
sums of squares factorial
between groups
ANOVA
sums of squares
within groups split plot
ANOVA
degrees of freedom
one-way
means square
ANOVA
hypothesis testing
We can then extend those principles to a wide range of
applications.
sums of squares factorial
between groups
ANOVA
sums of squares
within groups split plot
ANOVA
degrees of freedom
one-way repeated measures
ANOVA ANOVA
means square
hypothesis testing
We can then extend those principles to a wide range of
applications.
sums of squares factorial
between groups
ANOVA
sums of squares
within groups split plot
ANOVA
degrees of freedom
one-way repeated measures
ANOVA ANOVA
means square
ANCOVA
F ratio & F critical
hypothesis testing
We can then extend those principles to a wide range of
applications.
sums of squares factorial
between groups
ANOVA
sums of squares
within groups split plot
ANOVA
degrees of freedom
one-way repeated measures
ANOVA ANOVA
means square
ANCOVA
F ratio & F critical
Level 1:
Football Player
Thus far we have only considered one dependent
variable and one independent variable that was
categorized into several levels
One dependent variable
Level 1: Level 2:
Football Player Basketball Player
Thus far we have only considered one dependent
variable and one independent variable that was
categorized into several levels
One dependent variable
For example:
We can consider the effect of multiple independent
variables on a single dependent variable.
For example:
First Independent Variable: Athletes
For example:
First Independent Variable: Athletes
For example:
First Independent Variable: Athletes
Level 1: Level 2:
Adults Teenagers
We can consider the effect of multiple independent
variables on a single dependent variable.
degrees of freedom
one-way
ANOVA
means square
hypothesis testing
Factorial ANOVA partitions the total sums of squares
into the unexplained variance and the variance
explained by the main effects of each of the
independent variables and the interaction of the
independent variables.
Factorial ANOVA partitions the total sums of squares
into the unexplained variance and the variance
explained by the main effects of each of the
independent variables and the interaction of the
independent variables.
Adults
eat 13 slices on average Teenagers
eat 11 slices on average
Let’s begin with the main effect for Age Group
Adults
eat 13 slices on average Teenagers
eat 11 slices on average
Basketball Players
eat 10 slices on average
Soccer Players
eat 7slices on average
Now let’s look at main effect for Type of Athlete
Football Players
eat 15.5 slices on average
Basketball Players
eat 10 slices on average
Soccer Players
eat 7slices on average
Basketball Players
eat 10 slices on average
Soccer Players
eat 7slices on average
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
Then we square the deviations
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev.
Then we square the deviations
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev. 1.93
Then we square the deviations
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev. 1.93 1.93
Then multiply each squared deviation by the number of persons (9). This is
called weighting the squared deviations. The more person, the heavier the
weighting, or larger the weighted squared deviation values.
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev. 1.93 1.93
Then multiply each squared deviation by the number of persons (9). This is
called weighting the squared deviations. The more person, the heavier the
weighting, or larger the weighted squared deviation values.
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev. 1.93 1.93
wt. sq. dev.
Then multiply each squared deviation by the number of persons (9). This is
called weighting the squared deviations. The more person, the heavier the
weighting, or larger the weighted squared deviation values.
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev. 1.93 1.93
wt. sq. dev. 17.36
Then multiply each squared deviation by the number of persons (9). This is
called weighting the squared deviations. The more person, the heavier the
weighting, or larger the weighted squared deviation values.
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev. 1.93 1.93
wt. sq. dev. 17.36 17.36
Finally, sum up the weighted squared deviations to get
the sums of squares for age group.
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev. 1.93 1.93
wt. sq. dev. 17.36 17.36
Finally, sum up the weighted squared deviations to get
the sums of squares for age group.
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev. 1.93 1.93
wt. sq. dev. 17.36 17.36
Finally, sum up the weighted squared deviations to get
the sums of squares for age group.
Adults Teens
17 11
19 12
21 13
13 8
14 10
15 12
2 7
6 8
8 9
mean 12.78 10.00
grand mean 11.39 11.39
dev.score 1.39 - 1.39
sq.dev. 1.93 1.93
wt. sq. dev. 17.36 17.36 34.722
Note – this is the value from the ANOVA Table shown
previously:
Note – this is the value from the ANOVA Table shown
previously:
Tests of Between-Subjects Effects
Dependent Variab le: Pizza_Slices
Source Type III Sum of Squares df Mean Square F Sig.
Age_Group 34.722 1 34.722 10.25 0.01
Type of Player 237.444 2 118.722 35.03 0.00
Age_Group * Type of Player 73.444 2 36.722 10.84 0.00
Error 40.667 12 3.389
Total 386.278 17
Next we calculate the Type of Player Sums of Squares
Next we calculate the Type of Player Sums of Squares
– – – =
We simply sum up the total sums of squares and then
subtract it from the other sums of squares
386.278 – – – =
We simply sum up the total sums of squares and then
subtract it from the other sums of squares
386.278 – 34.722 – – =
We simply sum up the total sums of squares and then
subtract it from the other sums of squares
If the F ratio is greater than the F critical, we would reject the null hypothesis
and determine that the result is statistically significant.
By dividing the sums of squares by the degrees of
freedom we can compute a mean square from which
we can compute an F ratio which can be compared to
the F critical.
Tests of Between-Subjects Effects
Dependent Variab le: Pizza_Slices
Source Type III Sum of Squares df Mean Square F Sig.
Age_Group 34.722 1 34.722 10.25 0.01
Type of Player 237.444 2 118.722 35.03 0.00
Age_Group * Type of Player 73.444 2 36.722 10.84 0.00
Error 40.667 12 3.389
Total 386.278 17
If the F ratio is greater than the F critical, we would reject the null hypothesis
and determine that the result is statistically significant. If the F ratio is smaller
than the F critical then we would fail to reject the null hypothesis.
Most statistical packages report statistical significance.