Chapter 3 Methodology and Methods
3.3 The Two-Stage Cluster Sampling Strategy
It was obviously challenging to collect and analyse data from all schools in Kenya. To narrow down to manageable numbers, the study applied a two-stage cluster sampling. This process involved purposefully selecting a sample in at least two stages. In the first stage, I
conveniently sampled 3 counties out of the 47 counties. Secondary students’ achievement data from all schools in these 3 counties were quantitatively analysed (forming the
population). Second stage sampling drew from the analysed students’ achievement data. I stratified schools in clusters of achievement trends (C3-Thriving; C2-Oscillating; C1-
Dipping). C3 schools (thriving) are those whose achievement trends have progressively improved over time. C2 schools (Oscillating)) are those struggling to gain stable achievement progress; neither improving nor dipping. C1 schools (Dipping) are those whose achievement trends are regressively dipping over time. From these clusters, I purposefully sampled the final nine schools for in-depth qualitative study (3 schools from each cluster). The objective of this sampling process was to identify a manageable number of schools for qualitative data collection and analysis. The idea behind analysing achievement data was to help stratify and get categories of schools. The quantitative analysis, therefore, was used to categorise, and help determine schools for in-depth qualitative study. Scholars find this sampling strategy effective because it allows multiple criteria focused sampling process that builds rigor and credibility, increasing validity and reliability of the study (Teddlie and Yu, 2007; Agresti and Finlay, 2008). In this study, two-stage sampling was useful in narrowing down the sample; the population of schools from the three counties was too large. Secondly, within the
population, many schools lay in the outlined strata of C1, C2 and C3, yet only a small sample of the population was required for the study. The sampling strategy assisted in avoiding the use of all sample units in all selected clusters; important in avoiding the large sample, and perhaps unnecessary costs and time requirement associated with it.
3.3.1 Research Site
The study purposively identified 3 counties for secondary quantitative data analysis; Kakamega, Nakuru and Kajiado as shown in figure 3.1. Kakamega County is in a rural setting. Nakuru County is an urban setting. Kajiado country is a sub-urban but also a metropolitan setting bordering the capital city. The three counties’ location is significant in showing the variation or similarities in not only the practice but also, in identifying the various mechanisms specific to contexts of school leadership practice (Hammersley, 2005).
Figure 3.1: Research Sites KAKAMEGA COUNTY NAKURU COUNTY Sampled Counties KAJIADO COUNTY
3.3.2 Stage 1: Sample Clusters
The three Counties had 350 public schools, however, after data cleaning only achievement data from 300 schools qualified for quantitative analysis. The cleaning involved identifying schools with full seven-year continuous achievement data. Schools with missing data were excluded from the analysis. These schools were stratified along two identifiers; type (National, County and Sub-County) and achievement trends (C1-Dipping; C2-Oscillating; C3-Thriving). The latter identified after secondary analysis of students’ achievement data. Stage 1 sampling resulted in sample clusters shown in table 3.1 below.
Table 3.1: Sample Clusters (All school names are Pseudonyms)
Type of School C3 Schools (Thriving) C2 Schools (Oscillating) C1 Schools (Dipping)
National 1.Bakeko H.S. 1.Sideki H.S. 1.Sameki H.S.
2. Nabeko H. S 2.Makisia H.S. 2.Limuka H.S.
3.Kikuba H.S. 3.Wengeti H.S. 3. Bageno H. S
County 1.Mubindi H.S. 1.Koshere H.S. 1. Lidude H.S.
2.Kokoiko H.S. 2.Wiwa H.S. 2.Dosita H.S.
3. Mubari H.S. 3. Bagamu H.S. 3.Gegombe H.S.
Sub-County 1. Nabibo H.S. 1. Luguyo H.S. 1.Nodete H.S.
2.Sembe H.S. 2. Finyago H.S. 2. Bidobe H.S.
3.Shikuyo H.S. 3.Hutwesa H.S. 3.Temba H.S.
3.3.3 Stage 2: Final Study Schools
The final 9 schools sampled for in-depth qualitative data collection and analysis were drawn from sample clusters in table 3.1. The nine schools were conveniently selected and access to
schools sought. CEO facilitated the access to school principals prior to the start of data collection. Where the principal denied access, I dropped the sampled school and picked another from the sample cluster. The final 9 sampled schools from the sampling frame were reached after a written consent was provided by the school principal. Through this robust sampling procedure, 9 schools (3 from each cluster) were finally identified for in-depth qualitative study (See detailed secondary data analysis in chapter 4 section 4.2).
Table 3.2. Sampled schools for Qualitative Data Collection
C3 (Thriving) C2 (Oscillating) C1 (Dipping)
Nabeko H.S. Mubari H.S. Nabibo H.S. Sideki H.S. Bagamu H.S. Luguyo H.S. Bageno H.S. Lidude H.S BidobE H.S. 3.3.4 Research Participants
Having identified study schools, I purposively sampled research participants. Within schools, I sampled school Principals, Deputy Principals (Academic and Curriculum), Form Principals, Director of Studies (DOS), Strategic Leaders, Long-serving teachers (LST), New-teachers (NT), Board of Management and Parent Association (BOM and PA) chairpersons and Heads of Departments (HODs). In the local education authority (LEA), I sampled the Sub-County Education officer (SCEO) and the Sub-County quality assurance and standards officer (SCQASO). Thus, 9 schools, 9 principals, 92 teachers (holding senior, middle and junior leadership positions), 6 BOM/PA chairpersons, 5 LEA officers formed qualitative data sources. Table 3.3 (A, B, C, D) summarise descriptive details of research participants in this study. These leaders were most suitable for the study because they practice leadership within schools and LEA settings (Day et al, 2009), and their experiences importantly informed resultant study findings.
Form principals, strategic leader and DOS are non-official leadership position (Not recognised by TSC), however, created in schools by senior leadership to enhance system functionality.