ALGUNOS CONCEPTOS KANTORIANOS
2.6. La degradación de la fábula
The sample size was 595. This comprised 560 parents and 35 headteachers. A description of the sampling process is contained in the subsequent sections.
4.5.1 Headteachers
All 35 headteachers were selected. The headteachers were purposefully used for the study because they provided perspective about the nature of PI. These were based upon the MOE directives and included how parents are to be involved in their children's education. Also, they provided information on their perceptions of how parents are actually involved in IE.
4.5.2 Parents
Cohen, Manion and Morrison (2004) suggest that with a population of approximately 15,000, the ideal sample size is 375. Cohen et al. (2004) recommend that, due to the non-return of some of the questionnaires, the sample size can be increased to about half of the required number. For this reason, 560 parents were used for the study instead of the proposed 375. Based on the literature, this was a good representation of the sample and also accommodated questionnaires that were not returned. For representation purposes, the sample of parents was selected from all 35 schools.
The convenience sampling technique was considered initially to select parents for the study. Using this method meant that parents who came to the school to consult teachers or to discuss the progress of their child or for any other reason would be invited to be part of the sample. The use of this sampling technique was abandoned as it was likely to affect the ability to generalise results. Using the convenience sampling method also implies that every member of the population will not be given an equal chance of being selected. This was likely to introduce subjectivity into the study. This is because the method itself is biased. Secondly, the researcher may be biased and end up using their personal judgment to decide who will or will not be part of the sample.
The systematic sampling procedure was chosen to enhance the opportunity to generalise the results and lessen the possibility of introducing any form of subjectivity into the study.
This technique involves the selection of ‘‘subjects from a population list in a systematic rather than a random fashion’’ (Cohen et al., 2004, p.100). The first respondent is usually randomly selected from the list and thereafter every nth person is selected (Denscombe, 2010; Robson, 2002). It is recorded that the systematic sampling procedure does not give an equal chance for every member of the population to be selected, (Cohen, Manion and Morrison, 2007). The random selection of the first respondent, however, lessens this effect (Cohen et al., 2007)
The sample was weighted per region depending on the total population of parents who had children attending the respective schools (see Tables 4-3, 4-4, 4-5 and 4- 6). The systematic sampling technique was used on the list of pupils enrolled in the schools. The parents were contacted once the pupils had been identified. The parents became part of the sample for the study.
Parents with more than one child in the school were substituted after they had been selected once. Also, I ensured that my sample comprised equal numbers of fathers and mothers. This selection was done randomly until equal numbers of fathers and mothers were selected.
In order to get a representative sample of parents based on the sampling procedure described above, the following steps were taken. The first step was to count the total number of parents who had their children in schools which were part of the first phase of the pilot inclusive programme in each region. This was derived approximately from the total number of pupils in the schools. The approximated number of parents was based on the premise of one parent per child.
Table 4-3: Distribution of parents in the regions
Region Number of parents Number of schools
Greater Accra 3,865 10
Central 5,309 11
Eastern 5,422 14
Total 14,596 35
Based on Table 4-3, the proportion of each region was calculated (see summary on Table 4-4) and distributed accordingly (see Table 4-5).
Table 4-4: Sample proportion
Region Number of parents Proportion %
Greater Accra 3,865 27
Central 5,039 36
Eastern 5,422 37
Total 14,596 100
Table 4-5: Distribution of sample size among the regions
Region Percentage Sample per region
Greater Accra 27 151
Central 36 202
Eastern 37 207
Having derived the distribution of sample per region, the next step was to determine the sample size per school (see Table 4-6).
Table 4-6: Distribution of sample per school in each region
Region Number of schools Sample per region sample per school
Greater Accra 10 150 15
Central 11 198 18
Eastern 14 210 15
Total 35 558
From the calculation, the total sample of parents was 558. This fell short of two people as the calculation was based on a sample size of 560. Two parents were randomly selected from the Central region to make up for this shortfall. This region is considered to be the cradle of education in Ghana (where the Government focused on establishing model schools during the colonial era and after independence in 1957). It therefore seemed appropriate to select the extra two respondents from this region. It must be noted that the disproportionate sampling technique was used to reflect the uneven population of parents per region (see Table 4-6).
4.5.3 Interview with parents
Out of the 560 parents that were sampled for the study, 20 parents were purposively selected and interviewed to help provide more information and explanation for research questions two and three. Each interview lasted between 15 and 30 minutes. In purposively selecting the parents for the interviews, I first grouped them based on regions as shown in Figure 4-1 and then selected parents who fulfilled the criteria set till I had the required number per region.
Six parents each from the three regions were interviewed with an additional two from the Central region (see section 4.5.2 for the reason given for selecting additional sample from the Central region). In selecting the interviewees for each region, I ensured that there were equal numbers of fathers and mothers and at least 2 parents who had children with SEN (see Figure 4-1). The rationale for the distribution of the sample was that I wanted the opinion of parents from all the three regions, as well as that of fathers and mothers and also parents with and without children with SEN.
Male = 3 Females= 3 6 parents SEN= 2 No SEN = 4 Males = 4 Females = 4 8 parents SEN = 3 No SEN = 5 Males = 3 6 parents Female = 3 SEN = 2 No SEN = 4 Figure 4-1: Distribution of the parents interviewed
The sample of parents interviewed was relatively small because interviewing a large number would have resulted in data that may have been too large to handle.