3.1.6 Target Population, Sampling and Sampling
Procedures
Moonie, (2000) defines population as the entire group of
persons having the same characteristics that are of interest to the researcher.
Ringrose (1986) postulates that population denotes all the potential
participants from which the sample is drawn. This study targets a population of
four hundred and fifty (450) respondents. According to Best and Khan (1993:13)
a sample refers to a small proportion of a population selected for observation
and analysis. In this study the researcher will sample 150 participants for
detailed study.
In this research study, the sampling processes will be carried
out at two levels. Firstly, the identification of Gasabo District, Donor
organizations and boundary partners will be done purposively. In Gasabo
participants will come from 30 beneficiary cooperatives and 15
imirenge (sectors) coordinators representing the 15 sectors in the
district. Selected donor communities include UNDP, USAID, DFID, Belgian
Technical Cooperation (BTC) and WORLD VISION. Boundary partners purposively
selected will include His Excellency the President of the Republic of Rwanda or
his spokesperson, and officials (preferably permanent secretaries) from the
Infrastructure, Education, Health, Commerce and Trade, Finance and Economic
Development, Local Government, Gender and Foreign Affairs Ministries. All these
are selected purposively as key informants. Sullivan et al., (1990) say that
purposive or judgmental sampling specifically excludes certain types of people
because their presence might confuse the research findings. Bailey, (1987)
credits this technique with the advantage of being more representative and has
the ability to be generalized over a larger population. However, because this
technique does not give equal opportunity for all elements to be included
(being non-probability sampling technique) other sampling techniques need to be
employed to cater for the limitation.
Simple random sampling (SRS) technique is the natural
alternative. Aldridge and Levine (2001) define it as a design in which the
cases that will make up the sample are chosen in a single process of selection
from the sampling frame that covers the entire target population. In this
regard each element has an equal chance of being chosen. The questionnaire will
be administered on randomly selected members of the target community. The total
number of members of the target community will constitute the sampling frame.
Systematic random sampling will then be employed to ensure that all the 15
sectors of the Gasabo district will be represented. Sullivan et al. (1990)
credits this technique with the exceptional advantage of treating the target
population as a unitary whole. In this regard, its attempt to guarantee an
equal opportunity may in a way minimize bias and prejudice. Nonetheless some of
the key district officials will be purposively chosen for in-depth interviews.
Below is a summary of sampling frame, subgroups and the final sample size.
Table 3.1: Sampling systems
|
STRATUM
|
SUB-GROUP SIZE
|
SAMPLE SIZE
|
|
Rwandan Population
|
1 representative (head of state), 8
officials from 8 ministries
|
9
|
|
Donor Community
|
2 officials from the 5 chosen
organizations
|
10
|
|
Gasabo District
|
2 district officials (mayor & town clerk), 15
imirenge Coordinators, 10 members from each of the 15 cooperatives or
associations
|
131
|
|
TOTAL
|
|
150
|
(Source: own design)
Below is an overview of sample size in relation to the study
population and data gathering tool used.
|
Data collection method
|
Study population
|
No. of respondents
|
|
Questionnaires
|
120
|
40
|
|
Interviews
|
27
|
9
|
|
Focus group discussion
|
180
|
60 (5 sessions of 12 each)
|
|
Observational
|
123
|
41 (4 sessions of 10 each)
|
|
Total
|
450
|
150
|
N.B The researcher took heed of recommendations
by Frankel and Wallen (1996) that a representative sample should be at least a
third of the entire study population.
| 
