K. I Egbuchulem
Division of Paediatric Surgery, Department of Surgery, University College Hospital, Ibadan.
The aim of this editorial is to highlight the necessary information needed; and the basic steps in estimating the minimum sample size for common study designs among resident doctors and clinicians alike.
Concepts such as sample size determination, sample size justification, sample size adjustment and reestimation will be elucidated.
The statistical theory for sample size estimation is based on certain assumptions such as:1
- The population from which the sample is drawn is infinitely large hence it will be cumbersome to study such a population.
- The sample is selected by a simple random sampling method using a design effect.
It is noteworthy that, too few subjects make estimates unreliable and imprecise, and a study with such is poorly powered to detect the desired difference or effect.2
On the other hand, too many subjects amount to waste of resources with increasing risk of type I error.2
It is true that most clinical and hospital-based studies are quantitative research and not qualitative studies hence that will be the focus in sample size estimation.
Sample size estimation is a compromise between statistical requirements (power) and what is feasible. Such samples must be selected to obtain information which is reliable, precise, with narrow confidence interval and from which valid conclusions about the larger population can be drawn.3
Before setting out to estimate sample size, this research question must be answered. How many subjects do I really need to study?
To answer this question the researcher must first answer other questions that provide information about what they expect to achieve from the study.
That is, the researcher provides the parameters like the objectives and primary outcome measures that will be used for determining the sample size.