221130P - NON-RANDOM (NON-SCIENTIFIC) SAMPLING

Presentation prepared by Professor Omar Hasan Kasule Sr. MB ChB (MUK), MPH (Harvard), DrPH (Harvard) Research Consultant Almaarefa University.

INTRODUCTION:

• When random sampling is not feasible, we resort to non-random sampling methods that are logistically easier but are more prone to sampling bias. These methods are becoming more prominent in the internet era when researchers distribute and receive back questionnaires online.

CONVENIENCE SAMPLING:

• Convenience sampling also known as opportunity sampling, casual sampling, haphazard sampling, accidental sampling or availability sampling is purely subjective. It is according to the whims or inclinations of the investigator who will recruit research participants depending on the ease or convenience of accessing them. Being subjective, convenience sampling has no special for objectivity or representativeness. The only requirement of convenient sampling is that the participants are easy to find and they are willing to participate. An example of convenience sampling is approaching any person found in the market and asking them to join the research. Convenient sampling is easy and cheap but has a lot of bias and its statistical power is lower than random sampling.

SNOW BALLING SAMPLING:

• This is based on selection of friends or acquaintances. It is also called chain sampling or referral sampling. The first participant to receive a questionnaire will send it to as many friends or acquaintances as possible. Each of these will repeat the process and so on. The networks of social connections enable recruitment of many participants in a short time (high response rate). This method of sampling however suffers from high and unknown levels of bias

PURPOSIVE SAMPLING:

• Purposive sampling is also celled judgmental sampling. The researcher uses his or her judgment to choose research participants. The judgment is based on personal expertise and judgment. The method can be use when studying a phenomenon that is rate or is confined to a specific and knowable population.

QUOTA SAMPLING:

• A quota sample is a form of stratified sampling. It attempts to represent the population studied. A fixed number to be selected from each population category is fixed in advance. Each interviewer is given instructions to interview a certain number of females or a certain number of teenagers. The number corresponds to the proportion in the general population. The method is very subjective because the interviewer chooses as he or she wishes. Snowballing has the advantage of reaching population groups that are hidden from general view and are not easily accessible. The method does not assure but tries to get a sample representative of the population. It also has a high level of bias.