When choosing an appropriate research method for a chosen topic, sociologists consider the following issues:
- Is it dangerous?
For example, studying gangs and crime authorities.
- How much will it cost?
- How much time will you have to spend on the research?
- How big will the sample be? Will it represent the population? Will it be easy to generalise the collected data?
- Is enough data available? Obviously, it's almost impossible to study FBI.
- Positivists tend to concentrate on the quantitative, 'hard', data like statistics. They are looking for causes and effects and prefer scientific approach. They will prefer questionnaires with close-ended questions and structured interviews.
- Interpretivists prefer qualitative data as it shows the meanings beyond the actions. They are more likely to use observation and unstructured interviews to get the data.
- Privacy. The respondents might regret about giving particular info during unstructured interview, for example.
- Confidentiality. Should the identity of the respondents be revealed? The Statement of Ethical Practise (1996) says that confidentiality must be honoured 'unless there are clear and overriding reasons to do otherwise'. For example, Homan (1991) says that the identity of people in powerful positions should be named if they misuse their power.
- Protection. The collected data might bring harm to someone. For example, Ditton's study (1977) of workers in a bread factoty revealed petty thefts - the publication his research might have affected them.
- Informed consent\deception. Should the respondents be informed that they are studied?
Many sociologists argue that the respondents should be informed about the research, its aims and process. However, awareness of the respondents might affect their behaviour thus provide inreliable or invalid results. This is known as the Hawthorne Effect.
On the other hand, there is deception - not informing respondents about a research or providing wrong information about the purpose of the study. This might be considered as unethical, but sometimes it's the only way to collect data. Cover observation includes deception. Or another example - Humphreys (1970) gathered info about gay men pretending that he was doing a health survey.