Sampling and ethical issues - Research methods

AQA A-level Psychology: Revision Made Easy - Jean-Marc Lawton 2017

Sampling and ethical issues
Research methods

Sampling techniques

A sample is a part of a population and should be as representative as possible, i.e. possess the same characteristics as the population from which it is drawn. Several sampling techniques exist.

Random sampling occurs where all members of a target population have an equal chance of being selected. Computer-generated random number lists can be used.


Image Selection for random sampling is unbiased and the sample should be fairly representative.


Image Sometimes random sampling is impractical, e.g. not all members of a population are available.

Image Samples can still be unrepresentative, for example all females are selected.

Opportunity sampling uses whoever is available.


Image Opportunity samples are easy to obtain.

Image It is the only sampling type available with natural experiments.


Image Opportunity sampling is often unrepresentative, for example all students are selected.

Image As participants can decline to take part, it can turn into self-selected sampling.

Self-selected sampling involves using volunteers, usually responding to advertisements.


Image Self-selected sampling involves minimal effort to obtain participants.

Image There is less chance of the ’screw you phenomenon’ (where participants deliberately sabotage the study).


Image Self-selected sampling often provides biased samples, as volunteers can be a certain ’type’ of person and so eager to please that demand characteristics occur.


Fig 7.3 Self-selected sampling involves participants volunteering to take part in a study

Systematic sampling involves taking every nth person from a list of the target population. This includes calculating the size of the population and assessing how big the sample needs to be to work out what the sampling interval should be (how big n is).


Image Systematic sampling involves no bias in selection, so the sample should be fairly representative.


Image Samples can still be unrepresentative if n coincides with a frequency trait, e.g. n = every fifth house in a street and every fifth house is flats occupied by young people.

Stratified sampling is a small-scale reproduction of a population and involves dividing a population into its strata (sub-parts) and then random sampling from each strata. If one strata has 15 per cent of the population, then 15 per cent of the sample is drawn from that strata, and so on.


Image Sampling is unbiased, and as selection occurs from representative sub-groups, the sample should be fairly representative.


Image It is time consuming.

Image A detailed knowledge of strata is required.

Ethical issues

To protect the dignity and safety of participants, as well as the integrity of psychology, research should be conducted in an ethical manner. Full details of research should be submitted to the appropriate ethical committee for approval before commencing. The British Psychology Society publishes a code of ethics that researchers should follow:

1 Informed consent — participants should be fully informed of the objectives and details of research to make a considered decision as to whether to participate. Parental consent is obtained for under 16s.

2 Deception — misleading of participants and withholding of information should be avoided.

3 Protection of participants — participants should not be put at risk of harm and should leave a study in the same state they entered it.

4 Right to withdraw — participants should be aware that they can leave at any time, including withdrawing their data in the future.

5 Confidentiality — participants’ data should not be disclosed to anyone, unless agreed in advance.

6 Anonymity — participants are referred to by numbers, not names, so that data cannot be traced back to them.

7 Inducements to take part — participants should not be encouraged to participate through offers of financial gain or other gratification.

8 Observational research — observations should occur only in environments where people would expect to be observed.

9 Cost—benefit analysis — only if the benefits of research, in terms of knowledge gained etc., outweigh the costs, in terms of possible harm to participants etc., should the research be undertaken.

If deception is unavoidable there are measures that can be taken:

1 Presumptive consent — people of a similar nature are given full details of a study and asked whether they would have been willing to participate. If so it is presumed the real participants would not object.

2 Prior general consent — participants agree to be deceived but without knowing how it will occur.

3 Debriefing — immediately after a study finishes participants should be given full details and the right to withdraw their data. This applies to all studies, not just those involving deception, and also helps to alleviate possible psychological harm, so that participants leave in the same state they entered.