Types and Sources of Data...
Quantitative data is information presented in numerical form.e.g graphs or stats.
Qualitative data is descriptive rather than numerical and is usually expressed in words to convey feelings and experiences.
Quantitative sources include questionnaires, formal interviews, and the use of official stats. Qualitative sources include in-depth interviews, participant observation and the use of diaries.
Primary data is information collected by sociologists themselves for their own sociological purposes e.g. to test a hypothesis. Secondary data is not collected by sociologists, but by others such as governments, charities or individuals who keep personal diaries.
Sociologists use the following criteria:
- Validity... Will the method measure what it sets out to measure, i.e. will it give a true picture?
- Reliability... Will the findings of the study be replicable, i.e. will they be the same if the study is repeated?
- Representativeness... Will the sampling method be representative of the wider population to allow generalisations to be made from the findings?
- Practical Considerations... How do factors such as time and money affect the type of method we can use?
- Subject matter... Will the chosen method be appropriate to the kind of people or activity being studied?
- Ethical Issues... Will the study adhere to ethical guidelines? How will participants in the study give 'informed consent'?
- Theoretical perspective... The sociologists preferred perspective may affect the choice of method, e.g. positivists may prefer quantitative methods while interactionists prefer qualitative methods.
Natural sciences make extensive use of laboratory experiments to establish cause and effect relationships between factors or variables. Experiments are carried out under controlled conditions to make sure that results are objective and not influenced by the presence of the researcher.
Laboratory experiments are seldom used in sociology. Most research is carried out in society 'as it is' and not in the artificial surroundings of a laboratory. Society is too large and complex to investigate in laboratory conditions and there are ethical objections to experiments on people.
Some sociologists use the comparative method to identify key variables by comparing 'naturally occurring differences'. Durkheim's study of suicides in protestant and catholic areas.
Check Yourself in Types and Sources of Data/Experi
1) Give three examples of quantitative sources of data.
2) Explain the difference between primary and secondary data.
3) List seven criteria sociologists use when they select a research method.
4) Suggest three reasons why laboratory experiments are seldom used in sociology.
Surveys involve asking people questions. Survey research makes use of questionnaires and interviews to obtain information about a sample of the population. Research begins with a choice of a topic to investigate and with formulating an aim or hypothesis. A hypothesis is an untested theory, usually expressed as a statement to be proved or disproved.
The next step is to conduct a pilot study, i.e. a trial-run, which allows potential problems to be identified and adjustments to be made before the survey begins.
Questions may be either closed-ended or open-ended. Closed-ended questions give the respondent a fixed list of possible answers to choose from (e.g. Yes/No/Don't know). Open-ended questions allow respondents to answer freely in their own words (e.g. 'What made you decide to leave school at 16?').
Surveys use a variety of sampling techniques, because it is not normally possible to include everyone from the population in a survey. Sampling techniques allow a smaller group of people - a sample - to be selected from the population. The sample is chosen from a sampling frame, which is a list of people in a given population, such as the electoral roll or a list of patients at a doctor's surgery. The main types of sampling techniques:
- Random Sampling... This is where the sample is chosen from the sampling frame literally 'at random' to ensure that it is representative of the survey population.
- Stratified Random Sampling... This is where seperate random samples are chosen from sub-groups within the survey population, e.g. men and women and/or different age groups. More representative.
- Quota sampling... This is where an interviewer is not told exactly who to interview but must find a certain number or quota of interviewees who fall into certain categories, e.g. 20 women aged 16-25.
- Snowball sampling... The researcher builds up a sample by asking respondents to refer him of her to others. This has been used e.g. to study drug-taking. It is usually less representative.
Accurate sampling and a high response rate will ensure that a study's findings are representative. A survey's conclusions can be undermined by a low response rate, therefore successful use of follow-up questionnaires or interviews can be important. Surveys are based on 'asking questions' involving the use of self-completed questionnaires (including postal questionnaires) and structured, semi-structured or unstructured interviews:
- Questionnaires... These are a relatively cheap and quick way to collect information from a large sample. They can be sent by post or completed on the spot. E.g. In a classroom or office. The questionnaires and answers are usually pre-coded and results are quantified. Supporters of this method claim the data collected is reliable, easy to quantify and to apply when testing a hypothesis and identifying trends. Questionnaires can reach huge numbers of people. All respondents answer the same set of questions and they are not influenced by the presence of an interviewer. Critics point to the low response rate of postal questionnaires. Non-response can undermine the representativeness of a sample. Interactionists argue that by drawing up questions in advance it prevents respondents expressing true feelings and undermines the validity.
- Structured or formal interviews overcome some of the problems but offer less scope for surveying large numbers. Structured interviews also use pre-coded questions and answers, but a trained interviewer is present to read out the questions and complete the interview schedule. Formal interviews are conducted face-to-face with respondents but they are standardised to ensure reliability. They make fairly efficient use of time and resources. Critics argue that formal interviews lack validity because they give little scope for respondents to express true feelings. Labov shows that respondents who are reticent in a formal interview can be forthcoming and imaginative when the interview is informal.
- Semi-structured interviews give more scope to the interviewer and respondent. An interviewer schedule is drawn up in advance but some of the questions are likely to be open-ended; they may be asked in any order and different topics may be explored. Dean and Taylor-Gooby approached their interviews with benefit claimants in this way.
- An unstructured interview is a 'guided conversation' where the interviewer has complete freedom to decide which questions to ask and how to ask them. Questions are not pre-determined and are usually open-ended, allowing the respondent to expand in his or her own way.
Supporters of semi- and unstructured interviews argue that they allow the interviewer more freedom and flexibility to explore ideas and develop a hypothesis when conducting research. More flexible as not standardised and controlled. More valid as reveals more about respondent's true feelings, beliefs and behaviour. Unstructured can achieve greater empathy as interviewer can understand respondents and see things from their point of view.
Critics argue that unstructured interviews lack reliabilty because no two interviews will be alike. Time consuming, so sample likely to be small, making it harder to generalise the findings. Interviewers also need lengthy training to acquire the necessary skills and knowledge of sociology.
Although semi- and unstructured interviews are widely used, they attract theoretical criticism both from positivists, who see them as lacking reliability, and from interactionalists, who argue that even unstructured interviews create an artificial situation.
Both structured and unstructured are criticised for the interviewer effect: the fact that respondents adjust to the person asking the questions. The interviewer effect stems from the fact that interviews are social processes and distorts the info obtained.
Raises the problem of validity because the answers given can be greatly influenced by the way the respondent perceives the interviewer. Thus an interviewer of a different gender, class or cultural background. Williams found African Americans responded differently to black and white interviewers.
Graham argues surveys often fail to produce valid explanations of women's lives and experiences. Formal methods remove women from the real life situations in which they experience gender inequality. The interview situation prevents respondents true feelings.
Positivists argue better training and refinement of the techniques can help to overcome the difficulties, but many interactionalists reject interview techniques in favour of an approach based on direct observation.
Check Yourself in Survey Methods
1) Explain why pilot studies are used.
2) What is quota sampling?
3) Explain why follow-up questionnaires or interviews are important.
4) Suggest two reasons why questionnaires are considered to be a reliable method.
5) True or false? 'Structured interviews are like a guided conversation.'
6) Suggest three potential weaknesses of postal questionnaires.
7) What type of interviews did Dean and Taylor-Gooby use in their study?
8) Explain why unstructured interviews are criticised for being unreliable.
9) True of false? 'Unstructured interviews restrict the scope for empathy between interviewer and respondent.'
10) Explain what is meant by the interviewer effect and give an example.
The main observation method used by sociologists is participant observation, which involves the researcher joining in with the group that he or she wishes to observe. The researcher will often enter the group with an open mind and allow the hypothesis to emerge from the research. Pyrce adopted this approach in his study 'Endless Pressure'.
The aim of the PO is to gain understanding or insight into the lives of the subjects of the study. Its supporters claim it produces uniquely valid data because the best way to gain insight is through personal involvement.
When conducting overt PO the researcher does not disguise his or her identity, but with covert PO the researcher conceals their true identity and purpose.
Covert PO can give access to aspects of social life that are secret but critics object on the following grounds:
- Practical problems... Routine aspects of conducting the study are likely to raise suspicion, such as asking questions and taking notes. Covert observers run the risk of their cover being blown, and possibly injury.
- Ethical problems... Covert PO involves the researcher getting information by deception and contravening the principle of 'informed consent'.
Humphreys is criticised for disregarding ethics, not only because his observation was covert. but also because he noted the car reg numbers of the homosexuals he observed so that he could trace and interview them later.
PO can be divided into three stages:
- Getting in... Initially the observer will contact the group and work at gaining their trust. This may mean establishing a relationship with a key informant as a first step to gaining access to the group. E.g. Whyte's study of 'Street Corner Society', his key informant was the leader, Doc. Some groups are easier to join than others. Differences of age, class, gender or ethnicity can be an obstacle. The research also needs to decide what role to adopt.
- Staying in... The researcher has to strike a balance between being an observer and being a participant. The researcher needs to remain detached to study the group sociologically. However, he or she also needs to become involved with the group to understand them. There is a danger of the observer's objectivity being undermined by 'going native', by being too involved with the group and seeing things only from their point of view. Whyte notes he began as a non-participant observer and ended as a non-observing participant.
- Getting out... This presents further challenges, especially with covert PO. It may be difficult for the sociologist to explain why he or she is leaving the group and to break the emotional attachments formed during the study. Whyte reports he had difficulty readjusting to life.
PO has been widely used for the study of crime and deviance where suspicion of outsiders makes the survey methods impractical. As Yablonski points out, deviant groups are distrustful of researchers who want answers. An advantage, is that PO provides evidence. Since observations are conducted in natural settings it is less likely that researchers impose their versions of reality on the people being studied.
However PO is usually small-scale research. Difficult to generalise the findings to wider population. Also lacks reliability.
Check Yourself in Observational Methods
1) Name a study where the researcher began with an open mind, allowing the hypothesis to emerge.
2) Explain the difference between covert and overt.
3) Suggest two practical problems with covert PO.
4) Explain why Humphrey's study has been criticised.
5) Explain what is meant by 'going native'.
6) Identify three stages of doing participant observation.
7)Write a paragraph about the advantages and disadvantages of PO using the following terms: natural settings; uniquely valid data; re-test reliability; small scale.
Official stats are a major source of secondary data. They provide evidence of social trends, such as changing family patterns, educational achievement and crime rates. They can also be used to test a hypothesis as in Durkheim's study of suicide.
Official stats are collected by the state either by registration or by surveys. Unofficial stats are collected by agencies other than the state, such as pressure groups, trade unions, businesses and churches. Wilson used church attendance stats in his study of secularisation.
The practical advantage of stats is that they are a free source of huge quantities of data. However there may be omissions: Stats may not be available. Durkheim found in his study of suicide that the victims religion was not recorded on the death certificate. As a result he had to compare suicide rates of protestant and catholic areas.Definitions may be sociological.
Depending on their theoretical perspective, sociologists hold widely different views about the usefulness of stats:
- Positivists see them as a useful source of data that covers large numbers. The census covers the whole UK population. Positivists believe that, provided they are accurate, stats are a reliable and valid source of data.
- Interactionists are extremely critical of official stats. E.g. they argue that crime stats are socially constructed and probably tell us more about the priorities of the police than about patterns of crime. In this view, official stats are not a valid measure of the nature and distribution of crime. Atkinson puts forward this view in his study of coroner's suicide verdicts.
- Marxists argue that the state serves the interests of the ruling class and therefore official stats give a distorted impression that serves capitalism. Definitions of unemployment have regularly been changed and almost always with the result that the numbers officially counted as unemployed have been reduced.
Documents are a secondary source of qualitative data: usually written texts such as personal diaries, government reports, medical records, novels, newspapers and letters. Documents can also be electronic texts, works of art, music or styles of dress.
The classic example of the use of documents in sociological research is Thomas and Znaniecki's study of the polish peasant in Europe and America, which used 764 letters, biographies and autobiographies and public documents to explore the experience of Polish migration to the USA in the early 20th century. However, the sources used are impressionistic and there is no way of telling whether they are representative. The data is open to a variety of interpretations so it is difficult to prove its validity.
Content analysis is a way of dealing systematically with documents and it is widely used in studies of the mass media. It often involves careful recording of the number of images or references contained in the output of the mass media, such as the number of references made to strikes in TV news or the way men and women are portrayed in magazines. An example is 'Bad News', a study carried out by the Glasgow Uni Media Group. Content analysis can be relatively cheap and effective.
Check Yourself in Secondary Sources
1) Identify the two main methods used for collecting government stats.
2) Read the following statements and identify the perspective associated with it:
A ' Unemployment stats are a prime example of how the state manipulates info.'
B ' Crime stats are socially constructed. They tell us more about police procedures.'
C ' Provided they are accurate, official stats are a reliable and objective source of data.'
3) True of false? 'Sociologists use newspapers as a source of primary data.'
4) What types of documents did Thomas and Znaniecki use in their study of Polish peasant in Europe and America?
5) Give an example of a study based on content analysis.
Case studies, Longitudinal studies and Triangulati
A case study is a close examination of a single case or example. Case studies can be based on survey research or observation or on any other method and can be primary or secondary sources or both.
The main drawback of case studies is that we cannot always easily generalise about their results because we cannot be sure that a single case is a typical case. However, case studies have practical advantages. Case studies are less expensive than national surveys and, although they are smaller, can nevertheless provide detailed evidence for testing a hypothesis or comparing social groups, processes and events.
A longitudinal study is one that follows the same sample or group over an extended period of time. The best examples are surveys such as Douglas' study, which followed all the children born in England and Wales in the same week in 1946 and the National Child Development Survey based on a sample born in 1958. Some PO studies are longitudinal. E.g. Lacey spent four years studying pupils. The advantage of them is that they trace development over a period of time, although there can be problems keeping track of an original group.
Triangulation is a technique that involves more than one kind of method or source when carrying out research. E.g. It might combine observation with informal interviews and also use official stats. It means an approach that looks at something from more than one point of view.
Check Yourself in Case Studies etc
1) True or false? 'Case studies are based on large national samples.'
2) Explain what is meant by a longitudinal study and give two examples.
3) True or false? 'Triangulation is where three different research methods are used in a study.'