DBA 1015

Statistical Tools and

Techniques for
Research

Week 6 Lecture
One-Way ANOVA
One-Way Analysis of Variance
Purpose of ANOVA

1. Determines whether the differences between the means of the groups are due to chance (sampling error).
2. Can be used with both experimental and ex post facto designs.
Experimental Research Designs

 Researcher manipulates levels of Independent Variable to determine its effect on a Dependent Variable

Example

 Dr. Sophie studies the effect of different dosages of a new drug on impulsivity among children at-risk of
becoming delinquent.
Example of an Experimental Research Design Using ANOVA -- continued

1. Independent Variable
A. Different dosages of new drug
a. 0 mg (placebo)
b. 100 mg
c. 200 mg
d. Measure impulsivity in each group, compare groups
Ex Post Facto Research Designs

 Researcher investigates effects of pre-existing levels of an Independent Variable on a Dependent Variable.

Example

 Dr. Horace wants to determine whether political party affiliation influences attitudes toward the death
penalty using a scale assessing attitudes.
Example of an Ex Post Facto Research Design Using ANOVA --
continued

1. Independent Variable
A. Political Party Affiliation
1. Democrat
2. Independent
3. Republican
4. Measure attitudes toward the death penalty in each group
5. No manipulation
Null and Alternative
Hypothesis in ANOVA

1. No differences among the group means.

2. Alternative: At least one group differs from at
least one other group.
Example of Pairwise Comparisons

1. Dr. Mildred wants to determine whether birth order influences the number of self-reported delinquent
acts.

2. Independent Variable
A. Birth Order
1. First Born (or only child)
2. Middle Born (if three or more children)
3. Last Born
Example of Pairwise Comparisons -- continued

1. Dependent Variable
1. Number of self-reported delinquent acts
2. Possible pairwise comparisons
1. FB ≠ MB
2. FB ≠ LB
3. MB ≠ LB
3. It is possible for this analysis that:
1. Any one of the pairwise comparisons could be statistically significant
2. Any two of the pairwise comparisons could be statistically significant
3. All three of the pairwise comparisons could be statistically significant
Types of ANOVA

One-Way ANOVA
1. One Independent Variable
2. Groups are independent

Repeated-Measures ANOVA
1. Groups are dependent
2. Measure the dependent variable at more than two points in time
One-Way ANOVA

 The one-way analysis of variance is used to test the claim that three or more population means are
equal.
 This is an extension of the two independent samples t-test.
One-Way ANOVA

 The response variable is the variable you’re comparing.

 The factor variable is the categorical variable being used to define the groups.
 We will assume k samples (groups)
 The one-way is because each value is classified in exactly one way.
 Examples include comparisons by gender, race, political party, color, etc.
One-Way ANOVA

 Conditions or Assumptions
 The data are randomly sampled.
 The variances of each sample are assumed equal.
 The residuals are normally distributed.
One-Way ANOVA

H :        
0 1 2 3 k
 The null hypothesis is that the means are all equal

 The alternative hypothesis is that at least one of the means is different

 Think about the Sesame Street® game where three of these things are kind of the same, but one of these
things is not like the other. They don’t all have to be different, just one of them.
The Logic of ANOVA

 Total variability of the dependent variable (DV) can be analyzed by dividing it into its component parts.

 This is done between groups (BG) and its aim is to measure the overall differences between treatment
conditions (groups, samples).

 The variable between the groups is called the Treatment Effect (TE).
Within-Group (WG) Variability

1. Individual Differences (ID).

Example

 For race, there is more within group variability than between group variability (more genetic variation
among white, or Asians, etc, than between the races).
 ANOVA Partitions Total Variation

Total variation

Variation due to Variation due to

treatment random sampling

 Sum of squares among

 Sum of squares within
 Sum of squares between
 Sum of squares error
 Sum of squares model
 Within groups variation
 Among groups variation
ANOVA One Way Example
One-Way ANOVA

 The statistics classroom is divided into three rows: front, middle, and back.
 The instructor noticed that the further the students were from him, the more likely they were to miss
class or use an instant messenger during class.
 He wanted to see if the students further away did worse on the exams.
One-Way ANOVA

H :   
0 F M B
 The ANOVA doesn’t test that one mean is less than another, only whether they’re all equal or at
least one is different.
One-Way ANOVA

 A random sample of the students in each row was taken

 The score for those students on the second exam was recorded
 Front: 82, 83, 97, 93, 55, 67, 53
 Middle: 83, 78, 68, 61, 77, 54, 69, 51, 63
 Back: 38, 59, 55, 66, 45, 52, 52, 61
One-Way ANOVA
 The summary statistics for the grades of each row are shown in the table below

Row Front Middle Back

Sample size 7 9 8
Mean 75.71 67.11 53.50
St. Dev 17.63 10.95 8.96
Variance 310.90 119.86 80.29
One-Way ANOVA

 Variation

 Variation is the sum of the squares of the deviations between a value and the mean of the value.
 Sum of Squares is abbreviated by SS and often followed by a variable in parentheses such as SS(B) or
SS(W) so we know which sum of squares we’re talking about.
One-Way ANOVA

 Are all the values identical?

 No, so there is some variation in the data
 This is called the total variation
 Denoted SS(Total) for the total Sum of Squares (variation)
 Sum of Squares is another name for variation
One-Way ANOVA

 Are all of the sample means identical?

 No, so there is some variation between the groups
 This is called the between group variation
 Sometimes called the variation due to the factor
 Denoted SS(B) for Sum of Squares (variation) between the groups
One-Way ANOVA

 Are each of the values within each group identical?

 No, there is some variation within the groups
 This is called the within group variation
 Sometimes called the error variation
 Denoted SS(W) for Sum of Squares (variation) within the groups
One-Way ANOVA

 There are two sources of variation

 The variation between the groups, SS(B), or the variation due to the factor.
 The variation within the groups, SS(W), or the variation that can’t be explained by the factor so it’s called the error
variation.
One-Way ANOVA
 Here is the basic one-way ANOVA table

Source SS df MS F p

Between

Within

Total
One-Way ANOVA

 Grand Mean
 The grand mean is the average of all the values when the factor is ignored.
 It is a weighted average of the individual sample means.

 Between Group Variation, SS(B)

 The between group variation is the variation between each sample mean and the grand mean.
 Each individual variation is weighted by the sample size.
 One-Way ANOVA

 Between Group Variation, SS(B)

 The between group variation is the variation between each sample mean and the grand mean.
 Each individual variation is weighted by the sample size.

SS  B   n  x  x   n  x  x     n  x  x 
2 2 2

1 1 2 2 k k

SS  B    n  x  x 
k 2

i i
i 1
One-Way ANOVA
 After filling in the sum of squares, we have …

Source SS df MS F p

Between 1902

Within 3386

Total 5288
One-Way ANOVA

 Degrees of Freedom, df
 A degree of freedom occurs for each value that can vary before the rest of the values are predetermined.
 For example, if you had six numbers that had an average of 40, you would know that the total had to be
240. Five of the six numbers could be anything, but once the first five are known, the last one is fixed so
the sum is 240. The df would be 6-1=5.
 The df is often one less than the number of values.
One-Way ANOVA

 The between group df is one less than the number of groups

 We have three groups, so df(B) = 2
 The within group df is the sum of the individual df’s of each group
 The sample sizes are 7, 9, and 8
 df(W) = 6 + 8 + 7 = 21
 The total df is one less than the sample size
 df(Total) = 24 – 1 = 23
One-Way ANOVA
 Filling in the degrees of freedom gives this …

Source SS df MS F p

Between 1902 2

Within 3386 21

Total 5288 23
One-Way ANOVA

 Variances
 The variances are also called the Mean of the Squares and abbreviated by MS, often with an accompanying variable
MS(B) or MS(W).
 They are an average squared deviation from the mean and are found by dividing the variation by the degrees of
freedom
 MS = SS / df.

Variation
Variance 
df
One-Way ANOVA

 MS(B) = 1902 / 2 = 951.0

 MS(W) = 3386 / 21 = 161.2
 MS(T) = 5288 / 23 = 229.9
 Notice that the MS(Total) is NOT the sum of MS(Between) and MS(Within).
 This works for the sum of squares SS(Total), but not the mean square MS(Total)
 The MS(Total) isn’t usually shown
One-Way ANOVA
 Completing the MS gives …

Source SS df MS F p

Between 1902 2 951.0

Within 3386 21 161.2

Total 5288 23 229.9

One-Way ANOVA

 Special Variances
 The MS(Within) is also known as the pooled estimate of the variance since it is a
weighted average of the individual variances
 Sometimes abbreviated

 The MS(Total) is the variance2 of the response variable.

s p table, but useful none the less
 Not technically part of ANOVA
One-Way ANOVA

 F test statistic
 An F test statistic is the ratio of two sample variances
 The MS(B) and MS(W) are two sample variances and that’s what we divide to find F.
 F = MS(B) / MS(W)
 For our data, F = 951.0 / 161.2 = 5.9
One-Way ANOVA
 Adding F to the table …

Source SS df MS F p

Between 1902 2 951.0 5.9

Within 3386 21 161.2

Total 5288 23 229.9

One-Way ANOVA

 The F test is a right tail test

 The F test statistic has an F distribution with df(B) numerator df and df(W) denominator df
 The p-value is the area to the right of the test statistic
 P(F 2,21 > 5.9) = 0.009
One-Way ANOVA
 Completing the table with the p-value

Source SS df MS F p

Between 1902 2 951.0 5.9 0.009

Within 3386 21 161.2

Total 5288 23 229.9

One-Way ANOVA

 The p-value is 0.009, which is less than the significance level of 0.05, so we reject the null
hypothesis.
 The null hypothesis is that the means of the three rows in class were the same, but we reject that,
so at least one row has a different mean.
One-Way ANOVA

 There is enough evidence to support the claim that there is a difference in the mean scores of the
front, middle, and back rows in class.
 The ANOVA doesn’t tell which row is different, you would need to look at confidence intervals or run
post hoc tests to determine that.