One-way Analysis of Variance is a statistical technique for assessing if differences exist between two or more "groups". It analyzes comparisons of variance estimates tests to test whether the means of two or more independent groups are equal. Usually, though, comparing the means of two groups can be examined using an independent t-test so One-Way ANOVA is used to test for differences among three groups. The F-test and the t-test are equal and generate the same results when there are only two means to compare. It is one of the reasons that One-Way ANOVA is termed as an expansion of the independent t-test. Only quantitative data can be used for this test.
Statistical Assumptions of One-Way ANOVA
Following are the three assumptions of One-Way ANOVA.
Assumption of Normality: It is assumed that the dependent variable is distributed normally or is nearly distributed normally.
Homogeneity of Variance Assumption: The variance of dependent variable is assumed to be equal for all population.
Independence: It is assumed that the samples are independent.
Independent Variable and Dependent Variable
For the two hypotheses, we have established one dependent and one independent variable each.
Hypothesis:
Dependent Variable = Pregnant as a Result of Rape
Independent Variable = Condition of Health
Hypothesis
The One-Way ANOVA statistic tests the null hypothesis that samples in two or more groups are drawn from the same population.
Our Hypothesis is as follows:
Hypothesis
Null Hypothesis, H0
Alternate Hypothesis, H1
Hypothesis
Condition of health is different in Women who are pregnant due to rape.
Condition of health is not different in Women who are pregnant due to rape.
SPSS Output
Table 1.1: Test of Between-Subject Effects
Tests of Between-Subjects Effects
Dependent Variable: ABORTION IF WOMAN WANTS FOR ANY REASON
Source
Type III Sum of Squares
df
Mean Square
F
Sig.
Partial Eta Squared
Corrected Model
13.413a
1
13.413
62.169
.000
.119
Intercept
688.402
1
688.402
3190.770
.000
.874
ABHLTH
13.413
1
13.413
62.169
.000
.119
Error
99.460
461
.216
Total
1267.000
463
Corrected Total
112.873
462
a. R Squared = .119 (Adjusted R Squared = .117)
The dependent variable is Abortion if women want for any reason. Here Corrected Model, Intercept and independent variable (ABHLTH) have “Sig” values less than 0.05. The Partial ETA Squared is only 0.119.
Table 1.2: Estimated Marginal Means
WOMANS HEALTH SERIOUSLY ENDANGERED
Dependent Variable: ABORTION IF WOMAN WANTS FOR ANY REASON
WOMANS HEALTH SERIOUSLY ENDANGERED
Mean
Std. Error
95% Confidence Interval
Lower Bound
Upper Bound
YES
1.510
.023
1.464
1.556
NO
2.000
.058
1.887
2.113
Table 1.3: Levene's Test of Equality
Levene's Test of Equality of Error Variancesa
Dependent Variable: ABORTION IF WOMAN WANTS FOR ANY REASON
F
df1
df2
Sig.
160119.401
1
461
.000
Tests the null hypothesis that the error variance of the dependent variable is equal across groups.
a. Design: Intercept + ABHLTH
Table 2.1: Test of Between-Subjects Effects
Tests of Between-Subjects Effects
Dependent Variable: PREGNANT AS RESULT OF RAPE
Source
Type III Sum of Squares
df
Mean Square
F
Sig.
Partial Eta Squared
Corrected Model
.221a
3
.074
.401
.752
.003
Intercept
283.646
1
283.646
1543.954
.000
.768
HEALTH
.221
3
.074
.401
.752
.003
Error
85.611
466
.184
Total
809.000
470
Corrected Total
85.832
469
a. R Squared = .003 (Adjusted R Squared = -.004)
The dependent variable is “Pregnant as result of rape”. Here the “sig” of Corrected Model and Health is higher than 0.05.
Univariate Analysis of Variance
Between-Subjects Factors
Value Label
N
CONDITION OF HEALTH
1
EXCELLENT
130
2
GOOD
238
3
FAIR
86
4
POOR
16
Descriptive Statistics
Dependent Variable: PREGNANT AS RESULT OF RAPE
CONDITION OF HEALTH
Mean
Std. Deviation
N
EXCELLENT
1.25
.432
130
GOOD
1.23
.420
238
FAIR
1.28
.451
86
POOR
1.19
.403
16
Total
1.24
.428
470
Levene's Test of Equality of Error Variancesa
Dependent Variable: PREGNANT AS RESULT OF RAPE
F
df1
df2
Sig.
1.559
3
466
.199
Tests the null hypothesis that the error variance of the dependent variable is equal across groups.