Abstract:
This paper is concerned with Stratification as a component of design effect (Deff). There
is a need to determine the pattern and magnitude of this component when the sample
allocation is proportional, and to see how such component behaves when parameters like
strata means, strata variances, strata proportions and sample size of each stratum vary.
Factorial analysis is carried out, to determine the effects of those various factors on the
design effect and to form some idea about its expected magnitude when faced with a
design of a specific given structure. Also, an attempt will be made to arrive at the
expressions, which are believed original, for the design effect in the case of a population
consisting of two strata. It is shown that the different combinations of strata means and
strata variances affect significantly the design effect. The Deff values increase with the
increase in the levels of variances of strata, and as differences in strata means become
less. But decrease as strata sizes tend to equality. The precision of proportionate
stratified sampling is investigated and examined, and consequently, it is confirmed that it
is the same as that of simple random sampling (SRS) when strata means are equal.
