Leadership behaviors of school principals are of extreme importance because it helps to improve learner outcomes and job satisfaction of staff members. The article titled “The Leadership Behaviour of the School principal: An Exploratory Study in Five Special Schools in Kwazulu-Natal” by Naidoo, Muthukrishna, and Hobden (2012) aims at describing teachers’ experiences in terms of leadership behavior of the school principals. The researchers used a Likert-scale survey questionnaire to gather data from 50 participants from five special schools in Kwazulu-Natal to answer their research question. The present paper aims at analyzing the article by assessing all the constituencies.

## Research Problem

In the reviewed study, Naidoo et al. (2012) reported that despite the introduction of numerous policies that aimed at reforming the schools, low morale was prevalent among teachers. The central reason for the matter was that “most stakeholders in schools do not understand or poorly interpret the philosophical change and the practical implications of policy change” (Naidoo et al., 2012, p. 4886). Such a misunderstanding was a result of poor leadership behavior practices by principals in schools (Naidoo et al., 2012).

Even though the problem was well-researched in school settings, there was little information about principals’ leadership behaviors in special schools (Naidoo et al., 2012). Therefore, the central problem the research aimed at addressing was the gap in knowledge concerning principals’ leadership behaviors in special schools in Kwazulu-Natal.

The research problem is identified, and its significance was explicitly discussed. The problem was important and relevant because, if addressed, it could have informed policymakers about the changes needed in special schools to improve learner outcomes and teachers’ morale. The background of the problem was extensively discussed, and the variables were identified.

Additionally, the researchers clearly indicated that the study was explanatory, which added to the understanding of its design. In summary, the problem statement is one of the definite strengths of the article.

## Review of Literature

Even though a review of the literature was extensive, Naidoo et al. (2012) did not dedicate a separate section to it. The reference list included 50 references, the majority of which were up-to-date and relevant to the research problem. While the findings of the literature review were synthesized directly to the problem, the article did not include a critical appraisal of evidence. The review was not logically organized, and the findings were not related to the current body of knowledge. However, Naidoo et al. (2012) identified the gap in the literature and established a theoretical framework for data collection and analysis. In short, while the review of literature includes all the important studies, it can be improved by restructuring and providing critical analysis of the level of evidence.

### Selection of Participants

The population under analysis was 64 special schools in Kwazulu-Natal. The total sample size was 50 randomly selected participants from five special schools. Naidoo et al. (2012) selected ten participants from each school, among which 11 were male, and 39 were female. While the number of participants is sufficient, it does not support the generalizability of results, as it appears too small. At the same time, there is evidence of adequate assurance that the rights of human participants were protected. Naidoo et al. (2012) stated that “anonymity was assured, and participation by the teachers was voluntary” (p. 4887). Informed consent was obtained from the Department of Education and the participants. Even though the selection of participants was adequate, its generalization abilities are questionable.

### Instrumentation

Naidoo et al. (2012) utilized two questionnaires to obtain the data for analysis. The first questionnaire included personal information needed for the analysis, such as “gender, age, qualifications, number of completed years of teaching, post level as well as the classification of the school” (p. 4887). The second survey included 37 items created using the Principal Leadership Behaviors Questionnaire (Naidoo et al., 2012).

All the instruments were explicitly described in terms of purpose and content, and they were adequate for measuring the identified variables. Internal consistency was 0.959, which is very high, and the reliability coefficient was 0.70, which is acceptable (Naidoo et al., 2012). The questionnaires were created specifically for the study; however, they were based on findings from the extensive literature review (Naidoo et al., 2012). In brief, the instrumentation is a definite strength of the research.

### Design and Procedure

The research design seems to be appropriate for answering the research question. Even though the procedures were not described in much detail, the research can be replicated to improve the generalizability of findings. ANOVA is used to determine the differences in leadership behaviors of five schools, which is appropriate for the purpose. However, it should be noted that control procedures are not described, and the authors do not account for any potentially confounding variables.

A pilot study was not conducted, which can be considered, as pilot studies are useful for testing research protocols, data collection instruments, and sample recruitment strategies (Hassan, Schattner, & Mazza, 2006). However, even though the procedures had not been tested, they proved to be adequate for addressing the stated problem.

## Research Questions and Hypotheses

The study by Naidoo et al. (2012) aimed at answering one research question, which was: “What are teachers‘ experiences of the leadership behaviour of the school principal in the context of their own special schools?” (p. 4886). The researchers divided the central question into five sub-questions and used various methods for answering them. The research questions were as follows:

- RQ1: Is there a difference in teachers’ perceptions of the leadership behaviors of the school principal, depending on school?
- RQ2: Is there a difference in teachers’ perceptions about the leadership behaviors of the school principal, depending on the post level?
- RQ3: Is there a difference in teachers’ perceptions of the leadership behaviors of the school principal, depending on gender?
- RQ4: Is there a difference in teachers’ perceptions of the leadership behaviors of the school principal depending on experience?
- RQ5: What are teacher ratings of leadership behavior of the principal?

All the stated research questions followed the literature review. Even though the authors did not state the hypotheses, they can be formulated from the research questions. All the hypotheses were testable using ANOVA or other methods of statistical analysis. Further analysis of research questions and hypotheses will focus on research questions that were answered using ANOVA, which are RQ2 and RQ4.

### Research Question 2

- RQ2. Is there a difference in teachers’ perceptions about the leadership behaviors of the school principal, depending on the post level?
- Null hypothesis. There is no difference in teachers’ perceptions about the leadership behaviors of the school principal, depending on the post level. (H20: µlevel1 = µlevel2 = µlevel3)
- Alternative hypothesis. There is a difference in teachers’ perceptions about the leadership behaviors of the school principal, depending on the post level. (H21: µlevel1 ≠ µlevel2 ≠ µlevel3)
- Results. The researchers provided descriptive statistics of teachers’ perceptions about the leadership behaviors of the school principal by post level in a table and discussed these differences. The observation revealed that there were differences in the perception of leadership behaviors; however, none of them was statistically significant (Naidoo et al., 2012). The researchers used five ANOVA analyses (one test per factor) with a significance of 5% (Naidoo et al., 2012). Even though the p-values are not presented, it is stated that all of them are above 0.05, which implies that the null hypothesis could not be rejected (Naidoo et al., 2012). In other words, it can be stated that there was less than 95% certainty that the difference in means of teachers’ perception of leadership skills was attributed to post level rather than variation within groups. The ANOVA table can be reproduced using the summative statistics presented by the researchers. For instance, for the factor of collegial supportive relationships, one can use the data for level 1 (N=41; mean=-0.2466; SD=0.57401), level 2 (N=7, mean=0.0635, SD=0.80917), and level 3 (N=2, mean= -0.6389, SD= 0.03928) to calculate p-value and F-score using SPSS or any other statistical software. The results for this analysis would be F= 1.2945, p= 0.2836, variance within groups = 0.3640, variance between groups = 0.4713, effect size = 0.055. The effect size was calculated using the formula provided by Yockey (2018), which is the sum of squares between groups (0.9425) divided by the sum of squares within groups (17.1096) (p. 99). The results demonstrate that since the p-value (0.2836) is greater than the significance level, the null hypothesis could not be rejected.

### Research Question 4

- RQ4: Is there a difference in teachers’ perceptions of the leadership behaviors of the school principal depending on experience?
- Null hypothesis. There is no difference in teachers’ perception of the leadership behaviors of the school principal depending on experience. (H40: µ0-5 years ≠ µ6-10 ≠ µ11-15 ≠ µ16-20 ≠ µmore than 20)
- Alternative hypothesis. There is a difference in teachers’ perceptions about the leadership behaviors of the school principal, depending on the post level. (H41: µ0-5 years = µ6-10 = µ11-15= µ16-20= µmore than 20)
- Results. For reporting the results for RQ4, the researchers used a table of descriptive statistics that included the number of observations (N), means, and standard deviations. Similar to RQ2, the researchers did not provide ANOVA tables; however, they stated that all five ANOVA analyses could be reproduced using the descriptive statistics. For instance, one can input data for collegial supportive relationships for group 1 (N=3; mean=-0.0741; SD=0.65105), group 2 (N=5; mean=-0.1556; SD=0.84747), group 3 (N=11; mean=-0.1970; SD=0.59957), group 4 (N=13; mean=-0.3376; SD=0.49403), and group 5 (N=41; mean=-0.1883; SD=0.66507) into statistical software and calculate the results. The results for this analysis would be p=0.9463, F=0.1825, variance within groups = 0.3947, variance between groups = 0.0721, effect size = 0.016. Since the p-value is above the significance level identified by Naidoo et al. (2012), the null hypothesis should not be rejected. The authors report that the results of the ANOVA analysis were all above the significance level, which implied that there was less than 95% certainty that the difference in means of teachers’ perception of leadership skills was attributed to post level rather than variation within groups.

## Further Considerations

There is no evidence that the researcher avoided violating the required assumptions for parametric tests. According to Yockey (2018), there are three assumptions of ANOVA tests, which are the observations are to be independent, the dependent variables are normally distributed, and the variances are equal (p. 100). In particular, the researchers did not consider if the variables were normally distributed, which could mean that the results are biased. At the same time, there is no attempt to explain why the results were negative.

In general, it should be said that all the hypotheses are tested, and research questions are answered. The results are presented clearly in well-formatted tables and charts, which helps to understand the results. The results are complete and can be reproduced using the data provided in the article. However, it should be stated that the results are not discussed against the findings of previous research, which makes it difficult to understand the gap the research is trying to address. However, all the acquired results seem reliable, and the absence of a clear description of the testing procedures may be explained by the desire not to overload the reader with information.

## Conclusion

In summary, the research is an ambitious research process that examines multiple variables and relationships between them. However, the researcher answered all the research questions. All generalizations are consistent with the results, implications for practice are discussed, and recommendations for further research are outlined. At the same time, the assumptions and limitations of the study are not clearly stated, which is a significant omission. The major strength of the article is the research design, while the most significant flaw is the sample characteristics. The article will be helpful in understanding how ANOVA can be used in my future research and what mistakes should be avoided.

## References

Hassan, Z. A., Schattner, P., & Mazza, D. (2006). Doing a pilot study: Why is it essential? *Malaysian Family Physician: The Official Journal of the Academy of Family Physicians of Malaysia*, *1*(2-3), 70–73.

Naidoo, L., Muthukrishna, N., & Hobden, S. (2012). The leadership behaviour of the school principal: An exploratory study in five special schools in Kwazulu-Natal. *Gender and Behaviour*, *10*(2), 4883-4912.

Yockey, R.D. (2018). *SPSS demystified: A simple guide and reference* (3d ed.). Abington, UK: Routledge.