Kruskal-Wallis H test is a nonparametric method that is employed in testing the presence of statistically significant differences between three or more independent groups. The test has been broadly used in final year research in fields like education, business administration, economics, accounting, sociology, psychology, nursing, public administration, marketing, agriculture, and many other fields of social and management sciences. Kruskal-Wallis H test is chosen by researchers when the conditions for carrying out parametric tests like one way analysis of variance (ANOVA) cannot be fulfilled. The test is applicable especially when data is obtained on an ordinal scale, data is obtained from a small sample size or when data is non-normally distributed. Many studies at the undergraduate level usually employ questionnaire surveys that are answered on a Likert scale..

The KruskalWallis H-test is a statistical tool used after the researcher has collected, coded, and cleaned the data. At this point, all responses have to be thoroughly checked for missing values, data entry errors, duplicate records, and inconsistencies that could impact the analysis results negatively. After the dataset is set up in an efficient way using statistical software such as SPSS Stata R, Jamovi, or Python, the researcher should first look at the characteristics of the variables and figure out whether the requirements for parametric methods are broken. If the dependent variable is of ordinal type or if normality tests show that the data is not normally distributed, the KruskalWallis H-test is a suitable method for the statistical analysis.

Among the common instances where the Kruskal-Wallis H test ought to be used include occasions where the researcher wishes to conduct analysis of three or more independent groups. For instance, it may be necessary to compare satisfaction levels among customers of three different banks, levels of motivation among employees of federal, state and private institutions, or students' academic performance among three different institutions. Since the participants belong to different groups and each individual participant gives just one observation, the Kruskal-Wallis H test becomes the best way to examine whether any significant differences exist among the groups, without the need for normal distribution.

The Kruskal-Wallis H test becomes particularly appropriate when the researcher is dealing with studies that involve the collection of questionnaire responses from respondents, with answers given on the Likert scale. In most research projects conducted during the final years, respondents give their responses based on scales such as "Strongly Agree," "Agree," "Neutral," "Disagree" and "Strongly Disagree." Such responses constitute ordinal data rather than interval measures. Even though some researchers use parametric tests for the analysis of Likert-scale data, the Kruskal-Wallis H test offers a more conservative choice..

When it comes to violations of normality assumptions, researchers must not forget to also use the Kruskal-Wallis H-test method. The use of statistical tools like the Shapiro-Wilk or Kolmogorov-Smirnov tests can tell us that the distribution of a dependent variable dramatically deviates from the normal distribution. Usually, continuing to use one-way ANOVA might very well lead to erroneous and misleading conclusions. The Kruskal-Wallis H-test is a great solution as it does not analyze the raw scores, but instead processes observations by ranking them, Because of this making it not only less affected by extreme values but also less sensitive to departures from normality.

Besides, the Kruskal-Wallis H-test comes very handy when there are unequal sample sizes in the groups that are being compared. Usually, this is the case in a majority of research projects e.g. one group might have more respondents than the other First and foremost due to factors such as response rates and accessibility. Really the test is based on ranks of the data and not on strict distribution assumptions explains why it is still valid even when groups are unbalance. The feature is what makes it very appropriate for undergraduate field surveys and research using questionnaires.

One other major use of the Kruskal Wallis H test is hypothesis testing. In almost all final year projects, a null hypothesis is made, stating that there is no significant difference between the groups in question, while an alternative hypothesis states that there is a significant difference among the groups. The results of the analysis are then used by the researcher to interpret the value of H in conjunction with the value of the p-value. When the value of p is smaller than the specified significance level (normally 0.05), the null hypothesis is rejected, meaning that there is a statistically significant difference between the groups. When the value of p is larger than 0.05, the researcher does not reject the null hypothesis, meaning that there is no statistical difference between the groups.

When there is a statistically significant result obtained using the Kruskal Wallis H test, the researcher may decide to conduct further analysis in order to establish which groups are statistically different from each other. In the absence of this further analysis, all the researcher knows is that there is at least one significant difference among the groups..

While the Kruskal-Wallis H-test is very versatile, there are aspects that one needs to take into consideration when using this test. First of all, it will indicate whether there are any differences between groups; however, it will not provide information about the magnitude of these differences or the direction. In addition, it can be used only in cases where we have independent groups and cannot be applied to situations when we have repeated measurements. For repeated measurements, there are other statistical techniques available that deal with non-parametric data.

The use of statistical software makes the use of the Kruskal-Wallis H-test relatively easy. The H-value, degrees of freedom, significance value and other descriptive statistics are provided automatically by the software like SPSS, Stata, R, Jamovi, and Python. However, researchers need to know the basic rules of how the Kruskal-Wallis H-test works even when using a computer. Understanding the statistics correctly is necessary while writing Chapter Four of the final year project since examiners expect students to interpret their results in connection with research objectives, questions and hypotheses..

Finally, the Kruskal-Wallis H-test must always be used when conducting a final year research project involving three or more independent groups, which will either contain ordinal data or continuous data that do not meet the requirements of the one-way ANOVA test. This test is especially applicable when all the data collection and cleaning process is completed, and the results prove that a non-parametric test is needed. Unlike ANOVA, the Kruskal-Wallis H-test ranks the observations without having to assume that the data follows a normal distribution. Thus, the Kruskal-Wallis H-test is an effective way to make conclusions and conduct hypothesis testing about independent groups.