Microsoft Word research method fm doc

Some of the reasons for stratifying the population may be: 
ƒ
Different sampling schemes may be used in different strata, e.g. Urban 
and rural
ƒ
Conditions 
may 
suggest 
that prevalence rates will vary between strata: the 
overall estimate for the whole population will be more precise if 
stratification is used. 
ƒ
Administrative reasons may make it easier to carry out the survey through 
an organization with a regional structure. 
5. Cluster sampling: When a list of groupings of study units is available (e.g. 
villages, etc.) or can be easily compiled, a number of these groupings can be 
randomly selected. The selection of groups of study units (clusters) instead of 
the selection of study units individually is called cluster sampling. Clusters are 
often geographic units (e.g. districts, villages) or organizational units (e.g. 
clinics). 

---

Research methodology 
44
6. Multi-Stage Sampling: This method is appropriate when the population is large 
and widely scattered. The number of stages of sampling is the number of times 
a sampling procedure is carried out. 
ƒ
The primary sampling unit (PSU) is the sampling unit (or unit of 
selection in the sampling procedure) in the first sampling stage; 
• The secondary sampling unit (SSU) is the sampling unit in the second sampling stage, 
etc. 
e.g.
After selection of a sample of clusters (e.g. household), further sampling of 
individuals may be carried out within each household selected. This constitutes two-
stage sampling, with the PSU being households and the SSU being individuals. 
 
Advantages
: less costly, we only need to draw up a list of individuals in the clusters 
actually selected, and we can do that when we arrive there. 
Disadvantage
: less precise than SRS. 
When we take a sample, our results will not exactly equal the correct results for the whole 
population. That is, our results will be subject to errors. This error has two components: 
sampling and non-sampling errors.