Running Head: ONE-WAY ANOVA

One Way ANOVA

Name of Student

Name of institution
 2
ONE WAY ANOVA
Part 1

In the practical field, there was a test run for analyzing whether there is a significant

difference between the salaries received by the employees of finance, marketing and human

resource departments. This problem will be dealt with the help of ANOVA because there are

more than two population means to be compared. The expected outcome is that there will be no

significant difference between the salaries received by these three departments.

Section 1

The context of this data set is that it shows the grades of students throughout a particular

semester. These grades are calculated by a large number of incidents like quiz 1 , final etc. The

current analysis will use the analysis of variance because the sections variable has 3 distinct

groups and we will have to see the difference between these groups based on quiz 3. The section

variable is categorical in nature and has 3 categories which are independent of each other. Even

within each group, the individual data points are independent of each other. The variable quiz 3

will be ordinal in nature because numbers are seen as being from worst to best or vice versa. The

sample size for this study is 105.

Section 2

The first assumption of one-way Anova is that there is no dependence of one data point

on other data points. This will mean that the student in any section will be independent of any

student in that section and the whole class.

The normality of the dependent variable is second assumption of one-way Anova. This

can be assessed by the histogram for the data of quiz 3.

 3
ONE WAY ANOVA

The above diagram shows histogram for quiz 3 scores of students and there is a bar which

is very high in the histogram, this may mean that there are outliers in the data. There are almost

equal values on either side of the center of the graph which means that there is a low amount of

skewness in the data but there will be a high amount of kurtosis because peak of this data will be

much higher.
 4
ONE WAY ANOVA

Descriptive Statistics

N Minimum Maximum Mean Std. Skewness Kurtosis

Deviation

Statistic Statistic Statistic Statistic Statistic Statistic Std. Statistic Std.

Error Error

quiz3 105 1 10 7.15 1.849 -.256 .236 .039 .467

Valid N 105
(listwise)

In the above table, we see that there is a relatively low figure on both skewness and

kurtosis for the variable quiz 3. The skewness figure is negative because there are more data

points on the left of mean than on the right side of it. The kurtosis value is also lower because

there is only one high bar in the histogram.

Tests of Normality

Kolmogorov-Smirnova Shapiro-Wilk

Statistic df Sig. Statistic df Sig.

quiz3 .171 105 .000 .938 105 .000

The above table shows the normality analysis of the variable quiz 3 with the help of

Shapiro Wilk test. The significance level of test is 0.05 and the significance value in the above

table is 0.00 which means that we can reject the null hypothesis of normality of the variable.

The next assumption of the analysis is homogeneity of variance in terms of the dependent

variable for each sample. This assumption is to be satisfied only in case there is a major

difference in sample sizes. Following table will show levene’s test for the variable quiz 3.
 5
ONE WAY ANOVA

Test of Homogeneity of Variances

quiz3

Levene Statistic df1 df2 Sig.

1.291 2 102 .280

The above table shows levene’s teat for the quiz 3 variable. The null hypothesis will be
the equivalence of variances for all sections of the class. The significance level will be 0.05 and
the significance in the above table is 0.28 which means that we can accept the null hypothesis of
equivalence of variances.
The assumptions of the test have been partially met because the variable quiz 3 is found
to be non-normal but the three sections of the class have almost equal number of data points so
we can ignore the normality assumption in this case. The variances of all groups have been found
to be equal.
Section 3

The research question for this analysis will be:

There is no significant difference between the sections with respect to scores in

quiz 3.

Null Hypothesis: µ1=µ2=µ3

Alternate Hypothesis: µ1≠µ2≠µ3

The level of significance will be 0.05.

 6
ONE WAY ANOVA

Section 4
 7
ONE WAY ANOVA
 8
ONE WAY ANOVA

The above three graphs show the value of means for quiz 3 for the three sections

of this study. We see that there is apparently very little difference of values in the means of the

three sections as shown by the sizes of bars in the graphs.

 9
ONE WAY ANOVA

Descriptive Statistics

section N Minimum Maximum Mean Std. Skewness Kurtosis

Deviation

Statistic Statistic Statistic Statistic Statistic Statistic Std. Statistic Std.

Error Error

quiz3 33 1 9 6.21 1.833 -.593 .409 .561 .798

section
Valid N 33
1
(listwise)
quiz3 39 5 10 8.33 1.528 -.367 .378 -1.145 .741
section
Valid N 39
2
(listwise)
quiz3 33 4 10 6.70 1.468 .438 .409 .302 .798
section
Valid N 33
3
(listwise)

The above table shows values of means and standard deviations for three sections with

respect to the variable quiz 3. The means of section 1 and 3 are close together whereas the mean

of section 2 is a little higher. We will check through the Anova if there is a significant difference

between the mean scores of sections. The standard deviation of section 1 is a bit higher than the

other two sections bit we have seen through the levene test that the null hypothesis for

equivalence of variances was accepted. Since we have accepted the null hypothesis of

equivalence, we do not have to implement any post-hoc test.

 10
ONE WAY ANOVA

ANOVA Table

Sum of df Mean Square F Sig.

Squares

Between Groups (Combined) 5.176 7 .739 1.179 .0322

section * quiz3 Within Groups 60.824 97 .627

Total 66.000 104

Eta Eta Squared

section * quiz3 .280 .078

The above table shows the anova for the three sections of the class. The significance statistic is

higher than 0.05 which was selected as the level of significance for this test. This means that the mean

scores of the three populations are not equal. The eta squared value is 0.078 which shows a medium level

of effect size in this scenario. The degrees of freedom are calculated by subtracting 3 from the total

sample size of 105. This shows the pieces of independent information which went into calculating the

estimates. This also means that there is a significant difference between test scores on quiz 3 and students

of various sections.

Section 5

The ANOVA has been applied to analyze if there is a significant difference between

scores of students in quiz three across all sections of a class. The analysis reveals that there is a

significant difference between the means of three sections for quiz 3. There is a medium level of

effect size in this analysis and most of the assumptions have been fulfilled.

The main advantage of ANOVA is its straight forward calculation and interpretation. The

main disadvantage is that it can only be applied with the normality assumption. This is a robust
 11
ONE WAY ANOVA
design and increases the power of a statistical assessment. If the null hypothesis is rejected, we

may not know which group is different from the other groups.