Microsoft Word research method fm doc

6.7 Sample size determination 
• In planning any investigation we must decide how many people need to be studied in 
order to answer the study objectives. If the study is too small we may fail to detect 
important effects, or may estimate effects too imprecisely. If the study is too large then 
we will waste resources. 
• In general, it is much better to increase the accuracy of data collection (by improving 
the training of data collectors and data collection tools) than to increase the sample 
size after a certain point. 
 
• The eventual sample size is usually a compromise between what is desirable and 
what is feasible. 
• The feasible sample size is determined by the availability of resources. It is also 
important to remember that resources are not only needed to collect the information, 
but also to analyze it. 
 
In order to calculate the required sample size, you need to know the following facts: 
a) The reasonable estimate of the key proportion to be studied. If you 
cannot guess the proportion, take it as 50%. 
b) The degree of accuracy required. That is, the allowed deviation from 
the true proportion in the population as a whole. It can be within 1% 
or 5%, etc. 
c) The confidence level required, usually specified as 95%. 
d) The size of the population that the sample is to represent. If it is 
more than 10,000 the precise magnitude is not likely to be very 

---

Research methodology 
48
important; but if the population is less than 10,000 then a smaller 
sample size may be required. 
e) The difference between the two sub-groups and the value of the 
likelihood or the power that helps in finding a statistically significant 
difference. 
• Note that ‘e’ is required when there are two population groups and the 
interest is to compare between two means or proportions.