Abstract : One of the crucial questions in
researchers mind is how big my sample should
be. There are several methods of estimation of
sample size. We classify them in two categories.
The first one is the forecasting, estimation or
prediction and the second is the testing of
hypotheses or comparison. In each case the
results are reliable or precise if sampling is
proper and the size of sample is adequate. We
give the formulae to determine the size of
sample in each of the situations which will be
of immense use to the researchers.
Keywords : Sample, population, margin of
error, confidence coefficient, hypothesis, level
of significance, power of test, ANOVA, normal
distribution, control chart, degrees of
freedom,chi-square test, contingency table.
I. INTRODUCTION
Many researchers conduct survey using
sampling techniques. One of the important
questions they face is the size of the sample.
The answer to the question is not so simple or
unique. The complexity of the answer is mainly
due to the lack of knowledge of population or
universe (Population is the group of objects
under study). The researchers are always
criticized that the size of sample is inadequate.
There are two misunderstandings about the
sample size. The first one is, larger the sample
size better are the results. The second one is,
the sample size should be at least 10% of the
population. However the results are reliable if
the sample is proper representative of
population rather than it is unexpectedly larger
in size. Merely increasing the size of sample,
we do not achieve the precision. It does not
increase proportionately as the sample size
increases.