CATEGORIES:

# Sampling methods

There are several techniques of sampling:

1 Random sampling

This method aims to ensure that each member of the survey population has an equal chance of being selected for the sample. It requires an up-to-date 'sampling frame' to be accurate - a sampling frame is a list of people included in the survey population. This is not the same as the total population of the whole country. It would be very rare for a firm to wish to sample from the entire population of a country - from one day old to over a 100 years old! It refers instead to everyone in the target survey population, so, for example, a manufacturer of camping equipment would only be interested in those who go camping. From this survey population a list of names is drawn up and a sample is selected randomly from the list. This can either be done by selecting every nth item in the list (systematic random sampling) or electronically, by means of computer generated random numbers. The interviewer will then call on the people selected - possibly trying on more than one occasion if they are not at home. In Business Studies project assignments many students undertake market research and they interview shoppers at the high street bus station at a certain time of day. This is often described as being 'random sampling' - it is not. The sample selected in this way will not be representative because car drivers will not be included and, if conducted during the working day, full-time workers will be underrepresented. A random sample could instead be obtained by using the methods described from a list of all voters, the electoral roll, or even a comprehensive street directory.

2 Stratified sampling

This method involves dividing the population into sub-groups and only sampling from those sub-groups that are likely to be interested in the product in question. For example, a new magazine aimed at 12-16 year old girls. It would be pointless and time consuming to select samples from groups other than this particular sub-group. Once the group has been identified, then the sample can be selected by using either random of quota methods.

3 Quota sampling

By this method interviewees are selected according to the different proportions that certain consumer groups make of the whole survey population. For instance, if it is already known that, out of all consumers of denim jeans:

male 65%

female 35%

Age

14-20 35%

21-30 35%

31-40 20%

over 41 10%

Then the sample selected would conform to the same proportions. Therefore, if there was a sample of 200 people, 130 would be male, 70 female, 70 between 14 and 20 years old, and so on. The interviewer could then either obtain the quotas by questioning the right number of people in the high street or, if a sampling frame was available, the respondents could be selected randomly up to the quota for each group.

4 Cluster sampling

When a full sampling frame list is not available or when the product is mainly likely to appeal to specified groups of consumers, for example town or regional newspapers, then cluster sampling will take a sample from just this group - not the whole population. Random methods can then be used to select the sample from this group.

5 Systematic sampling

This is where a set numerical formulae is used to choose the sample, for example every tenth or twentieth person is selected. Again, if the population consists of distinct strata, then it is not really a very appropriate method.

6 Convenience sampling

In this case, the interviewer selects any convenient place, where large groups of people might congregate, as the location for conducting the interviews. Obviously, this makes the sample quicker and cheaper to conduct, but bias may be introduced by the fact that everyone is from the same location. An example of this might be interviews carried out in large shopping centres or in queues for the cinema.

Date: 2015-02-16; view: 1645

 <== previous page | next page ==> Quantitative research techniques | Methods of scientific cognition
doclecture.net - lectures - 2014-2024 year. Copyright infringement or personal data (0.007 sec.)