11. POPULATIONS AND SAMPLES

 

University of Botswana

Department of Media Studies

BMS 328 COMMUNICATION RESEARCH METHODS

HANDOUT 10: POPULATION AND SAMPLES

Definitions

A population is a group or a class of subjects. For example, the population of Botswana includes all the people who live in the country; the population of Gaborone is all the people who live in the city and the population of UB Gaborone Campus is all the people who live on the campus. A population does not have to be a group of people; it can also be a group of things. For example, all the newspapers published in Botswana in 2012 could be a population.

The process of examining every member of a population is called a census. In most cases however it would be impractical to question an entire population because of the time and costs involved in doing this. To get over this problem a sample of the population is selected instead.

A sample is a smaller group that is taken to be representative of the entire population. It is important that the sample is representative of the entire population because if it is not representative the results gathered from the sample cannot be guaranteed. If the sample is representative of the entire population the results gathered from the sample can be generalized to the population. This means that if you did a census of the entire population you would get the same results that you got from only using the sample.  

However, because a sample does not provide the exact data that using the whole population does there is a likelihood that some error could creep into the results and therefore a sample would not give a true picture of the population as a whole.

Random and non-random sampling

A random sample is where each person or thing in a population has a known chance of being selected. An example is the selection of numbers in a lottery. Each number selected has an equal chance with any other number of being selected.

When we talk of random sampling we are including varying degrees of randomness.

These include:

Ø  Pure random sampling where everyone has an equal chance of being selected (e.g. a lottery).

Ø  Systematic sampling where every nth number is elected. For example, if you are conducting research into student attitudes to living on Campus you could select a sample by going round the dorms and interviewing the occupants of every tenth room.

Ø  Cluster sampling where samples are taken in clusters of the population. For example students at Campus might be studied by taking a sample from each academic faculty.

Ø  Stratified cluster sampling is a form of cluster sampling but it allows for varying sizes of the clusters. For example, if one faculty is twice as large as another the sample taken from it would also be twice as large.

If a sample is not random it is called a purposive sample.  Here are some types of purposive sample

Ø  Strategic informant sample which is where you select the people whom you think can give you the most information for your research. It may make sense to seek information about the content of Botswana’s newspapers by asking editors or senior journalists.

Ø  Snowballing (as a snowball rolls down a hill it gathers more snow along the way) could be seen as a kind of strategic informant sampling. Here you ask someone selected in the strategic informant sample (above) who else you ought to be talking to get useful information and then you do the same with each of the new people you are told about. In this way the number of informants you have ‘snowballs’.

Ø  Haphazard samples are samples that are readily available. A UB student researching into what young adults watch on television may use his or her own classmates as subjects.

Ø  Volunteer samples consist of people such as students, readers or television viewers who volunteer to participate in a sample.

How large should a sample be?

The size of a sample required for a study depends on at least one or more of the following:

• The type of project
• The purpose of the project
• Project complexity
• Amount of error you are prepared to risk
• Time constraints
• Financial constraints
• Previous research in the area

Research designed as a preliminary investigation (to search for general pointers before doing the main investigation) generally does not require a large sample. However, projects designed to answer significant questions (those that involve large sums of money or decisions that may affect people’s lives) require large samples.

Generally speaking, the larger the sample used the better. However, a large unrepresentative sample is as meaningless as a small unrepresentative sample, so researchers should not consider numbers alone. Quality is always more important in sample selection than size.

SOURCES

Du Plooy, G. M. (2001). Communication Research. Techniques, Methods and Applications. Juta: Landsdowne. Pp.99-116

Marshall, P. (1997). Research Methods How to Design and Conduct a Successful Project. How To Books: Plymouth. Pp.56-64.

Wimmer, R. D. and J. R. Dominick. (1994). Mass Media Research an introduction, (4th ed.). Belmont: Wadsworth. Pp.63-82.