 # Tutorial on Introduction to biostatistics

sampling

Once the hypothesis, level of significance and power of the study have been fixed, the researcher can proceed to determine the statistical processes for the proper conduct and analysis of the study.

As discussed earlier, the researcher usually draws conclusions about the population from a small part of it – the sample. The information collected from the sample is known as sample statistics which is used to estimate the characteristics of the unknown population i.e. population parameters.

We know that the sample taken from the population should accurately represent the population under study. To get a representative sample, the most important intervention is to select a sample large enough to adequately represent the population. Sadly, researchers have to strike a balance between striving for maximal validity while keeping the cost of the study at a level they can afford...!

From what has been stated so far, it can be deduced that the sampling process involves two important aspects. One is deciding the

Method of sampling

Method of sampling involves selection of samples from the given population. There are two basic methods in sampling.

a)       Probability sampling

b)    Non-probability  sampling

a)Probability Sampling

(i)                  Simple Random Sampling

(ii)                Stratified Random Sampling

(iii)               Systematic sampling

(iv)              Cluster sampling

b)Non Probability Sampling

(i)                  Judgement sampling

(ii)                Convenience sampling

Determining the sample size is based of a number of issues such as:

• Type of study
• Nature of study i.e. whether estimating parameters or comparing parameters
• Type of sampling method
• Type of analysis used in the study
• Power of the study
• Effect size
• Study budget
• Time factor

It can be readily appreciated that sample size calculation is rather complex. It is always best to consult a statistician for determination of sample size and other challenging biostatistical issues if embarking on a research project.