3. 5.Sample size and
selection techniques
The study must consider a sample size that is within the cost
constraint but should provide the ability to detect an independent variable
effect (Christensen, 1991: 372). Williamson (1982:113) comments on the sample
size as being a phase of research, which is crucial because of its major impact
on time and money that must go into data collection.
While selecting the sample size of the study; the researcher
purposely five respondents particularly; representing all staff members. Data
is collected from selected respondents from the study population due to the
reasons of necessary and convenience.
3. 5. 1. Stratified
sampling
The study population was stratified into two strata, one
stratum was comprised of staff members of selected small and medium enterprise
and the one was comprised of the population around that enterprise. From these
two strata, the researcher used both simple random and purposive sampling
techniques as these enabled her to select respondents who could provide her
with the information needed for the study.
3. 5. 2. Simple random
sampling
Simple random sampling was used to select enterprise. Simple
random sampling as defined by Baker (1988:148) refers to the situation whereby
each individual case in the population theoretically has a chance of being
selected for the sample. The simple random sampling technique is used to select
enterprise.
3. 5.3. Purposive sampling
Bailey (1978:83) explains purposive sampling technique as a
technique whereby the researcher uses her own judgment about which respondents
to choose and picks only those who can best meet the purposes of the study.
The formula of Alain Bouchard as cited by SABITI Fred (2004:
46)
 where, No = t2(p)(1-p)/d2
n is the sample size, N is the size of the population,
No is the sample size of a defined population, d is the error term
that is estimated 5%, p is the estimated frequency of the sample with size n,
while t is the figure obtained from the t-student's table.
Therefore, basing on the above formula, the researcher decided
to use 95% as the confidence level of which Alain Bouchard says is more
reliable. Thus, p=0.5, N= 45,000, d= 5% = 0.05, t=0.55
No=
(0.55)2(0.5)(0.5)/(0.05)2=31 thus n =31/1+31/45000 =
30
It was on this ground that the researcher selected 5 staff
members and from the enterprise 30stackeholders of it. The researcher's
intention was to ensure that the sample includes the elements that are directly
relevant to the problem being investigated /studied.
| 
