EXPLANATION
Contents:
Data collection:
Often the problem that you are investigating is such that there is no original data available. In this case, you would have to use collection techniques such as interviews or questionnaires to extract the data from a group of respondents. This is known as collecting primary data.
At other times, though, you may be able to use data that has already been collected. It may be useful to you in its original form, or you may have to change its format to fit your needs. Access to this kind of secondary data is increasingly becoming available in electronic form. TimeWeb is an example of this data availability.
With data that is available in electronic form, you have the opportunity to download the raw material directly into an Excel (or other spreadsheet or statistical package) workbook.
[Top]
Sampling Methods:
Imagine that you are studying political attitudes in the UK and want to know how the electorate is feeling about the political parties at one point in the year, say for instance in the Autumn. If your method of sampling consisted of standing outside the venues of the main political parties' annual conferences, asking questions of passers-by, it is likely that any conclusions that you make are invalid. This is because it is likely that the views of the respective political parties will be more loudly voiced at or around their own conferences. Many of the people in the immediate area of the conference are likely either to be sympathetic to the respective party or they may be journalists - neither of which outcomes is promising in terms of achieving a reliable result to your survey!
-
Purposive Sampling
This means choosing the sample with an underlying purpose or theme. Purposive sampling can be an obvious source of bias. Unscrupulous groups or individuals may wish to make a particular point and may choose their sample with this purpose in mind. But if, for example, a tea producer wants to find out consumer attitudes to a new blend, the producer should construct a sample out of known tea drinkers, otherwise the survey would be a waste of time. In general it is wise to beware of the results of surveys when the nature of the sample is not reported.
-
Random Sampling
Random sampling aims to overcome in-built bias by making sure that any member of the population is as likely to be chosen in the sample as any other.
-
Multi-Stage Sampling
Sampling that is done in several stages in order to minimise cost and effort.
-
Stratified Sampling
Used by researchers to eliminate bias in the sample, stratified sampling involves dividing the target population into strata or groups that are linked by some characteristic, such as socio-economic features.
-
Quota Sampling
Often found in street-based market research surveys, quota sampling involves interviewers questioning members of the public up to a given quota. The quota is generally stratified so that the correct proportions from the various sections of the population are represented. The interviewing locations are usually chosen by the multi-stage method and the selection of the sample is up to the interviewer's judgement - although the interviewer must keep to their quota of the required number of males, females, appropriate age profiles, and so on.
-
Cluster Sampling
Used when the population is not known, cluster sampling involves a mixture of random and multi-stage sampling techniques. To carry out a survey of rail commuters, for example, a number of starting points is chosen, with the total area divided into regions and a random sample of these chosen to survey. Next, a random number of trains to survey is selected. Then, a random sample of train carriages is chosen and then, every fifth passenger in the selected carriages is interviewed (for instance).
-
Survey Types
Four main types are used: postal, telephone, face-to-face and electronic surveys.
Postal surveys are a cheap way of reaching a large number of people. In principle, the whole population can be reached by post, but in practice not many people respond unless there is a strong incentive for them to act. For this reason postal surveys are prone to bias.
Telephone surveys are also comparatively cheap to operate. Interviewers can reach a large area very quickly, without the need to travel long distances.The main drawback is that despite the near universal coverage of the phone network and phone ownership, many people cannot be reached easily over the phone. Also, increasing numbers of phone customers are choosing not to list their numbers. There is a strong possibility that without checks to correct for these points, the phone survey may prove to be unrepresentative.
Face-to-face surveys allow for a more intensive questioning and response method to be practiced. Surveys aimed at delving into the reasons for certain attitudes developing will often use this method. However, it can be a very costly method, if a large geographical area is being sampled from.
Electronic surveys are increasingly common and this can be expected to continue as greater online coverage of the population is achieved. Typically, the findings of electronic surveys are raw (giving, for example, numbers of people watching a TV programme, or number of 'hits' to a web site) and these are augmented by telephone surveys to find out the socio-economic background of the sample population.
There is more detail available on sampling and survey types in the 'Illustration' section
[Top]
