Sociology - Research Methods

Positivism

Believe it is possible to study the whole social world using the methods of natural sciences. They prefer quantitative methods, ie. structured interviews, questionnaires. They look at social facts, correlations and cause and effect relationships. They believe that the behaviour of groups and individuals can be objectively measured. They use empirical (based on observed and measured phenomena and derives knowledge from actual experience rather than from theory or belief) and objective methods - eg. experiments, surveys & content analysis.

Interpretivism

They prefer qualitative methods - eg. observation, unstructured interviews. They believe that individuals and social groups construct the social world. They are interested in the meanings which people attach to their own actions & the actions of others. They take a more naturalistic approach and believe that human beings have feelings, attitudes and opinions so cannot be studied using methods of natural sciences. We cannot predict human behaviour so we need to explore internal meanings which direct our behaviour. Methods they favour - focus groups, content analysis (semiotics).

Feminism - Feminists argue that most sociological research reflects patriarchal society and benefits men. Hiedensohn - said research focused on males and their activities in society. Aspects of social life directly related to women are ignored. Research completely ignored women's experiences. It used to be argued that female researchers should use qualitative research methods. However, it is now said that sociologists of any gender should use the most appropriate methods for that piece of research rather than arguing about whether it is more or less suited to patriarchy or feminism (Skeggs). Pawson argues that sociology has always been like this and feminists use the usual range of methods.

Realism - Tries to bridge the gap between positivism and interpretivism. Influenced by natural sciences and also how peoples behaviour creates and transforms society. It was often claimed that Marx was a realist; he believed that positivists only looked at the surface of social life. So, you need to statistically observe but also gain a deeper understanding of social life at the same time. Realists therefore favour mixed research methods to break down the division between quantitative and qualitative data. Churton says much scientific research is not reliable - is rarely re-tested so is no more reliable than qualitative data. Realists researchers decide on a problem, identify the most likely theories to explain it and use a range of methods to compare the theories with each other.

Practical;

• Time it takes to complete the research
• Money taken to employ trained interviewers etc.
• Funding/lack of funding from Universities (Economic and Social Research Council), charities (the Joseph Rowntree Faoundation), protest groups (Green Peace), NHS, Businesses/companies, thinktank, Government etc.
• Difficulty getting access - may need a gatekeeper - sometimes these are hard to access themselves.
• Facilities, eg. paper for questionnaires, computers.
• Social characteristics of the researcher - important in observation - rapport is essential. If a female researcher was posing as a tough male gang member it obviously wouldn't work - have to have appropriate characteristics.

Ethical Issues;

• Gaining informed consent from participants (Laud Humphries did not do this in his Tearoom Trade experiment) regarding what will be measured, how it will be published, what will be done with it etc after already having informed them of all this information.
• Confidentiality/privacy - have the right to know that their information wont be shared with anyone else.
• Avoiding harm or danger to the participants and the researcher - it can be extremely dangerous to put yourself at risk as a researcher, especially whilst researching deviant groups. Sometimes this is even relevant whilst not researching deviant groups, ie. if you were told to measure the extent of exercise amongst elderly people they may feel pressured to do more exercise and may harm themselves.
• The right to withdraw - they should have the right to withdraw at any point
• Competence 
• Deception - ie. not telling the truth, lying.

Theoretical Issues;

Includes reliability, validity, reflexivity, representativeness, generalisability, positivism, interpretivism, realism, feminism.

• Generalisability - the extent to which it is safe to apply the findings from the research to the target population.
• Representativeness - the extent to which the sample being researched are a fair reflection of the target population and are typical of those in the target population.
• Reliability - the extent to which the research can be replicated by the same sociologist at another time or that the same study will be carried out and the same results will be collected.
• Validity - the extent to which the data is a true picture of the social reality of those being studied so that the research does what it set out to do.
• Reflexivity - where researchers are more open & honest with their interviewees and also think about their own experiences whilst conducting the research.

Hypothesis; a claim which can be proved or disproved

Paradigm; accepted ideas about something

Value freedom; sociology which is not tainted by people's beliefs, sociology which is objective. Is sociology value free?

YES - just because you choose to study a topic, it doesn't mean you have a particular opinion on it, if carried out in a scientific (positivist) way then it can be value free. Quantitative data can be value free, also if the research is covert it can be value free. Also, structured interviews (positivist method). Weber claims that values are a central feature of all societies and are at the heart of sociological study. Researchers are inevitably influenced by the society around them and their own values.

NO - Sociology is not value free because even if they are doing it subcosciously, researchers views, opinions and feelings affects the results gained & the data collected which makes it subjective. Topics you choose and the way you approach the research can be influenced by yout norms and values. If you are a career academic, you are therefore biasing the research because you need to get it published - if it is funded be a certain group it needs to follow their guidelines. Where the method is subjective, it is less likely to be value free.

Sampling; the selection of a group to study from the target, or survey, population.

The ten-yearly census carried out for the government by the ONS is based on the UK population as a whole. It is carried out in the first year of each decade.

Sampling frames; a list of the target population, eg. The Electoral Register - a list of all eligible to vote in the UK, The Postcode Address File (PAF) - this includes all UK households in each postcode, Schools or university registers or a Payroll list - includes all employees of a company. There can be many problems with sampling frames, including;

• They may be out of date. For eg, when people move it may take them a while to update their address details.
• They may not include everyone in the target population. For eg, traveller children, home-schooled children and long-term hospitalised children are not on a school register.
• They may not identify gender, ethnicity or age.
• Many files may not be available for confidentiality reasons.
• There are no lists of groups such as the homeless.

Sometimes a gatekeeper is needed to access sampling frames, ie. a headteacher of a school.

Random Sampling Techniques: (when everyone in the population has the same change of being selected) - is objective and scientific - favoured by positivists who value representativeness.

Simple random sampling: involves selecting a certain number of entries from a list, usually one that is computer generated, completely randomly. S - everyone has an equal change of being chosen - representative, and advanced auxiliary information on the elements in the population is not required. Such information is required for other probability sampling procedures, such as stratified sampling. W -  Respondents may be widely dispersed; hence, data collection costs might be higher than those for other probability sample designs such as cluster sampling, a sampling frame of elements in the target population is required. An appropriate sampling frame may not exist for the population that is targeted, and it may not be feasible or practical to construct one. 

Systematic sampling: where every nth names (for example, every tenth name) on a list is chosen. S - If the selection process is manual, systematic sampling is easier, simpler, less time-consuming, and more economical than simple random sampling, systematic sampling ensures that the sample is more spread across the population. W - whereas in simple random sampling every combination of n elements has an equal chance of selection, this is not the case for systematic sampling, principle of independence is violated, for the selection of the first element determines the selection of all the others