admin October 21, 2023 No Comments

Kruskal-Wallis Test in SPSS

Kruskal-Wallis Test in SPSS

Kruskal-Wallis test is used to determine whether there are any statistically significant differences among the means of three or more independent groups. The Kruskal-Wallis test is an alternative test for one-way ANOVA.

The Kruskal-Wallis test has the following hypotheses:

Null hypothesis: M1 = M2 = M3 (All medians are equal).

Alternative hypothesis: Not all medians are equal.

Assumption

  1. Dependent variable should be continuous.

  2. Independent variable should be of three or more categories.

  3. Independence of observations is necessary.

Systolic Blood Pressure is considered as the dependent variable and Body Mass Index is considered as the independent variable. Healthy weight is labelled as 1, Underweight is labelled as 2, Overweight is labelled as 3 and Obesity is labelled as 3.

In SPSS, the Kruskal-Wallis test is found in Analyze > Nonparametric Tests > Legacy Dialogs > K Independent Samples.

We will get “Tests for Several Independent Samples” dialog box.

Add the dependent variable (Systolic Blood Pressure) in ‘Test Variable List’ box and the independent variable (Body Mass Index) in the ‘Grouping Variable’ box. Then, click on ‘Kruskal-Wallis H’ box under the ‘Test Type’ group.

In the ‘Define Ranges’ option, enter ‘1’ in the ‘Minimum’ box and enter 4 in the ‘Maximum’ box, because Healthy weight is labelled as 1, Underweight is labelled as 2, Overweight is labelled as 3 and Obesity is labelled as 4. Then, click on ‘Continue’.

In the options box, click on ‘Descriptive’ box under the ‘Statistics’ group.

Click on ‘Continue’ and then on ‘Ok’.

Output of Kruskal-Wallis Test

First, we will get a ‘Descriptive Statistics’ table which gives mean, standard deviation, minimum value and maximum value for the dependent and independent variables.

Then, we will get a ‘Ranks’ table, which presents the mean rank for BMI groups. A high rank value is for obese people. Thus, it could be inferred that systolic blood pressure will be more for people with obesity.

Then, we will get a ‘Test Statistics’ table. It could be observed that the p value (Asymp. Sig.) is 0.000, which is less than 0.05. Therefore, significant evidence is available to reject the null hypothesis.

It could be concluded that BMI category of obesity (N=27) has a larger mean rank value (130.76) than the other three BMI categories (Healthy weight [N=89, rank=71.34], Underweight [N=14, rank=88.07], Overweight [N=60, rank=117.20]). A statistically significant difference was found in systolic blood pressure based on body mass index (Chi square, p<0.001).

Data: Kruskal_Wallis_Test_Data.sav

Write a comment

Your email address will not be published. Required fields are marked *