Glossary

Action theory: social theory in which action, its purposive nature and its meaning to people, is taken to be of central importance. Action theory is often associated with the name of Max Weber, who developed the interpretive tradition in social science.

Analysis of variance or ANOVA: A method for analysing the relationship between two or more variables where the dependent variable is interval-level and the independent variable(s) is or are nominal. It proceeds by testing the significance of any differences between the mean values of the dependent variable within the different groups described by the independent variable. ANOVA is useful where the independent variable has three or more categories, and can then be understood as an extension of the logic of the t-test.

Androcentrism: ideas or methods of research which prioritize men's views of the world, excluding the experience of women.

Anthropological strangeness: the art or mental trick of making a social setting and behaviour within it appear as if the observer is encountered it as a stranger. If applied to mundane 'taken-for-granted' events, this can lead to unusual and original insights.

Anti-essentialism: see essentialism.

Archives: repositories of a variety of materials, such as documents, photographs and films, often of an historical nature, catalogued and filed for the use of researchers, scholars and other investigators. See also data archives.

Association: see correlation

Bias: A realist approach to bias depicts this as consisting of any systematic error that obscures correct conclusions about the subject being studied. Typically, such bias may be caused by the researcher, or by procedures adopted for data gathering, including sampling. The concept makes little sense from a relativist standpoint, though provision of a reflexive account of the research process can help in addressing issues of trust that the concept of bias was intended to resolve.

Boolean searches: searches for material (such as references or segments of coded text) using combinations of keywords linked by operators such as 'and', 'or' or 'not.' Databases (for example, library catalogues) and qualitative analysis software (such as NVIVO) commonly support such searches.

Bracketing: used in semiotics to indicate the suspension of interest (for analytic purposes) in the relationship between signs and their referents. The term is also helpful in understanding the mental attitude required when doing discourse analysis or any analytic approach that treats text as a topic rather than a resource. Instead of considering the claims made in texts about reality outside the text, bracketing forces the analyst to consider the 'reality' the text constructs.

Career: used, primarily, by symbolic interactionists and ethnographers to describe a person's progress through a social setting, as where marijuana users progress through various stages in learning how to experience the drug, or mental patients pass through a series of institutional settings.

Case study: the study of a single 'case' - for example, a person, an institution, an event. How 'caseness' is defined depends on the logic of the particular research inquiry. For example, a nation might be thought of as a 'case' for certain purposes, even though a nation contains many people, each of which might be understood as a 'case' in some other inquiry.

Census: a count of the characteristics of every member of a given population (as opposed to a survey of a selected sample from that population).

Coding: this is done when observations, segments of text, visuaul images or responses to a questionnaire or interview are collected into groups which are like one another, and a symbol is assigned as a name for the group. Data may be `coded' as they are collected, as where respondents are forced to reply to fixed-choice questions. Alternatively, the coding of qualitative data can form a part of an interpretive, theory building approach.

Comparative method: the comparison of people's experiences of different types of social structure or social setting in terms of historical points in time, or across cultures at a single point in time. This is an approach which can shed light on the particular arrangements of both sides of the comparison. (see also constant comparison in index)

Connotation: used in semiotics to indicate the interpretive meanings of signs, which may be ideological. Thus a picture of a soldier saluting a flag connotes nationhood and patriotism as well as the more straightforward things such as `soldier' and `flag' that it denotes.

Constructionism: see social constructionism.

Content analysis: normally used in methods texts to refer to the quantitative analysis of texts or images, content analysis is in practice often combined with qualitative thematic analysis to produce a broadly interpretive approach in which quotations as well as numerical counts are used to summarise important facets of the raw materials analysed,

Contingency table: a table of numbers in which the relationship between two variables is shown. Contingency tables can usefully be broken down into rows and columns. Percentages placed in the cells of the table, giving the proportion which each cell contributes to the sum of particular rows or columns, are often helpful in detecting the strength and direction of relationships.

Correlation: in social statistics this term means the same as association, referring to a situation where two variables vary together. Amongst other things an association or correlation may be positive (in which case the two variables rise together) or negative (where one goes down the other goes up). Correlation coefficients (or tests of association) exist to indicate the strength and direction of linear relationships like this.

Cultural scripts or texts: terms used by those concerned to analyse cultural objects, such as pictures, films, sports events, fashions, food styles, to indicate that these can be viewed as containing messages in a manner comparable to a piece of written text.

Data: is the plural of datum, which refers to a record of an observation. Data can be numerical (and hence quantitative) or consist of words or images (hence qualitative). A distinction is sometimes made between naturally occurring data - such as tape recordings of conversations that would have occurred whether a researcher was present or not - and data generated in research settings, as in interviews or on questionnaires. Quantitative data are often arranged in a data matrix for ease of analysis.

Data archives: can be distinguished from the more general term 'archive' in that they contain quantitatively coded material from surveys, or qualitative material collected as part of social research studies, made available through the archive for secondary analysis.

Deconstruction: is an approach to social analysis that undermines claims to authority by exposing rhetorical strategies used by social actors, including the authors of research reports themselves. It has been promoted in particular by the post-modernist Derrida.

Deduction: see hypothetico-deduction.

Dependent variable: see variable.

Determinism: is the view that everything that happens is caused. When applied to human action, it suggests that our perception of having a free will is an illusion, and that the task of social research is to expose the true causes of action.

Discourse: has come to refer, under the influence of Foucault, to systems of knowledge and their associated practices. More narrowly, it is used by discourse analysts to refer to particular systems of language, with a characteristic terminology and underlying knowledge base, such as medical talk, psychological language, or the language of democratic politics.

Elaboration paradigm: a structured approach to the exploration of causal relationships between variables through the examination of contingency tables. By introducing third variables to bivariate tabulations, arguments about causal direction and spuriousness are tested. The logic of this approach underlies most multivariate statistical analysis.

Empiricism: the view that knowledge is derived from sensory experience, for example visual observation. More loosely, it has been used to describe research that contains little in the way of reflection or theory, preferring to report `facts' as they appear to be (as in the term 'abstracted empiricism').

Epistemology: refers to the philosophical theory of knowledge, consisting of attempts to answer questions about how we can know what we know, and whether this knowledge is reliable or not. Debates about the adequacy of empiricism, for example, are epistemological debates.

Essentialism: is now increasingly used in order to explain why anti-essentialism is preferable, though in more purely philosophical discussion the term has greater usefulness. Amongst social and cultural researchers, anti-essentialism involves the rejection of a scientific quest for universal essences, such as the discovery of a universal psychological makeup, or generally applicable sex differences, in preference for a view that human `nature' is a social construction.

Ethnocentrism: refers to the practice of judging a different society by the standards and values of one's own. This is seen, particularly by ethnographers, as inhibiting understanding of other ways of life.

Ethnomethodology: involves the examination of the ways in which people produce orderly social interaction on a routine, everyday basis. It provides the theoretical underpinning for conversation analysis.

External validity: see validity.

Facticity: is the process whereby certain perceptions or phenomena achieve the status of uncontroversial fact. Phenomenological analysis attempts to reduce facticity, as does the method of deconstruction, by exposing the social practices which generate it. Achieving facticity may involve both the objectification and the naturalisation of something as a fact.

Frequency distribution (or frequency count): a count of the number of times each value of a single variable occurs. Thus, the proportion of the population fitting into each of six categories of social class may be given as a frequency distribution. The distribution can be presented in a variety of ways, including for example a raw count, percentages or a pie chart.

Functionalism: is an approach to explaining social phenomena in terms of their contribution to a social totality. Thus, for example, crime is explained as necessary for marking the boundary of acceptable behaviour, reinforcing social order. Prominent functionalists include Durkheim and Parsons.

Genealogy: Foucauldian term involving the metaphor of a family tree to indicate an interest in the historical and social roots of ideas, systems of knowledge or discourses. Foucault suggested in a further metaphor that an archaeological approach towards the elucidation of these be adopted.

Grounded theory: a term coined by Glaser and Strauss to describe the type of theory produced by their methods of ethnographic data collection and analysis. The approach emphasizes the systematic discovery of theory from data, by using methods of constant comparison and theoretical sampling, so that theories remain grounded in observations of the social world, rather than being generated in the abstract. This they propose as an inductive alternative to hypothetico-deductive approaches.

Hypothetico-deduction: is the view that science proceeds by deriving hypotheses from theories, which are then tested for truth or falsity by observation and experimentation. It is the opposite of induction, which proposes that theories can be derived from observations.

Idealism: often opposed to realism, this term describes the view that the world exists only in people's minds.

Independent variable: see variable.

Internal validity: see validity.

Interpretive content analysis: see content analysis.

Interpretivism: refers to approaches emphasizing the meaningful nature of people's participation in social and cultural life. The methods of natural science are seen as inappropriate for such investigation. Researchers working within this tradition analyse the meanings people confer upon their own and others' actions.

Intersubjectivity: the common-sense, shared meanings constructed by people in their interactions with each other and used as an everyday resource to interpret the meaning of elements of social and cultural life.

Intertextuality: The explicit or implicit echo of one text in another text. This may take the form of explicit cross-references, or implicit, latent themes.

Linguistic repertoire: a term used in discourse analysis to refer to the resources (discourses, intersubjective meanings, etc.) on which people draw in order to construct accounts.

Logistic regression: see regression.

Malestream knowledge: see androcentrism

Marginality: Used to describe the typical position of the ethnographer, who exists on the margins of the social world being studied, in that he or she is neither a full participant nor a full observer. Also used to describe groups of people living outside mainstream culture.

Measures of central tendency: statistics such as the mean, median or mode which in various ways indicate the central point in a frequency distribution.

Methodology: concerns the theoretical, political and philosophical roots and implications of particular research methods or academic disciplines. Researchers may adopt particular methodological positions (for example, concerning epistemology or political values) which establish how they go about studying a phenomenon. Method, on the other hand, generally refers to matters of practical research technique.

Multiple regression: see regression.

Multivariate analysis: analysis of the relationships between three or more variables (as opposed to bivariate analysis, which involves two variables, or univariate analysis which involves one).

Naturalists: take the view that the methods of the natural sciences are appropriate to the study of the social and cultural world. This should be distinguished from another meaning of the term naturalism or naturalistic which is sometimes used to refer to the claim of ethnographers to collect naturally occurring data.

Naturalizing: is the process whereby matters which are in fact socially constructed and were once fluid and changeable come to be perceived as a part of the natural order and therefore fixed, inevitable and right. Social researchers often wish to denaturalize phenomena (such as sexual identity for example) by exposing the human processes whereby they are constructed.

Ontology: a branch of philosophy concerned with what can be said to exist. This can be distinguished from epistemology which concerns how we may know what exists.

Operationalisation: the process of developing indicators for concepts. Thus a concept such as 'alienation' might be indicated by questions on a questionnaire about powerlessness, isolation or moral deviance. The adequacy of operationalisation is an aspect of measurement validity, but can also be applied usefully to assess the adequacy of links made in qualitative research between ideas and examples.

Paradigms (Kuhnian): the overall conception and way of working shared by workers within a particular discipline or research area. In this view, paradigm shifts occur from time to time as scientific communities experience revolutions of thought.

Participant observation: used to describe the method most commonly adopted by ethnographers, whereby the researcher participates in the life of a community or group, while making observations of members' behaviour.

Path analysis: a procedure associated with multiple regression involving a diagram indicating the strength and direction of influences between several variables, enabling calculation of direct and indirect causal pathways.

Plagiarism: presentation of someone else's work as if it were your own. This can be done by direct copying without citation of the original work, or by summarising another person's ideas and presenting them as if they were your own. Self plagiarism occurs when a person presents or publishes the same piece of work more than once without indicating the first source. Plagiarism is generally used to indicate the idea of copying, although failing to acknowledge another person's contribution to a work that is being presented or published for the fist time may be considered plagiarism: it is certainly a dishonest practice. Other forms of cheating include fabrication of quotations, data and other results, although the concept of 'fabrication' depends on certain epistemological assumptions.

Positivism: in its looser sense has come to mean an approach to social enquiry that emphasizes the discovery of laws of society, often involving an empiricist commitment to naturalism and quantitative methods. The word has become almost a term of abuse amongst social and cultural researchers, losing its philosophical connotations where its meaning is both more complex and precise.

Postmodernism: a social movement or fashion amongst intellectuals centring around a rejection of modernist values of rationality, progress and a conception of social science as a search for over-arching explanations of human nature or the social and cultural world. By contrast, postmodernists celebrate the fall of such oppressive grand narratives, emphasizing the fragmented and dispersed nature of contemporary experience.

Poststructuralism: see structuralism.

Primary sources: see secondary sources.

Qualitative thematic analysis: analysis based on the identification of themes in qualitative material, often identified by means of a coding scheme. A widely used approach to qualitative analysis, generally treating accounts as a resource for finding out about the reality or experiences to which they refer, this is similar to interpretive content analysis.

Quasi experimental design: involves control of spurious variables by means of statistical operations at the analysis stage, rather than the design stage (as occurs in randomized controlled trials). The approach is often used to analyse survey data, or in situations where strict experimental designs may be impractical or unethical.

Randomized controlled trial: an experimental method whereby subjects are randomly allocated to either a group receiving a `treatment' or another which acts as a control', so that the effects of the treatment can be established. The method is effective in ruling out spurious causation.

Reactivity: the reactions of people being studied to the presence of an observer, seen by some to be a source of bias, in that behaviour may become artificial as a result.

Realism: is the view that a reality exists independently of our thoughts or beliefs. The language of research is seen to refer to this reality, rather than purely constructing it, though more subtle realists recognize constructive properties in language as well.
Reductionism: the identification of a basic explanation for a complex phenomenon. Thus sexual identity may be explained by reference to genetic determinants alone, or social life explained in terms of economic relations alone.

Reflexivity: in its broad meaning this is used to refer to the capacity of researchers to reflect upon their actions and values during research, whether in producing data or writing accounts. More narrowly, ethnomethodologists use the term to describe a property of language, which reflects upon actions to make them appear orderly.

Regression: a statistical technique for using the values of one variable to predict the values of another, based on information about their relationship, often given in a scattergram. Multiple regression involves the prediction of an interval-level variable from the values of two or more other variables. Logistic regression does this too, but predicts the values of nominal or ordinal variables.

Relativism: can be epistemological (or `conceptual'), cultural or moral. The first of these involves the rejection of absolute standards for judging truth. The second suggests that different cultures define phenomena in different ways, so that the perspective of one culture cannot be used to understand that of another. The third implies that perceptions of good and evil are matters of social agreement rather than having universal validity.

Reliability: the capacity of a measuring device, or indeed of a whole research study, to produce the same results if used on different occasions with the same object of study. Reliability enhances confidence in validity, but is insufficient on its own to show validity, since some measurement strategies can produce consistently wrong results. Establishing intercoder or interrater reliability may be important in some studies where unambiguous meanings for codes in a coding scheme are at stake, so that exercises in which the same material is coded by more than one person and the results compared for consistency may be carried out.

Replication: is closely linked with reliability, involving the repetition of a study to see if the same results are obtained on both occasions. (The term has a narrower meaning within the context of the elaboration paradigm.)

Rhetoric: the linguistic strategies used by speakers or authors of text to convey particular impressions or reinforce specific interpretations, most commonly in support of the authority of the text to speak the truth.

Sampling: the selection of units of analysis (for example, people or institutions) for study. Sampling can involve attempts to statistically represent a population, in which case a variety of random or probability methods are available. Alternatively, sampling can be opportunistic, or formed by emerging theoretical concerns of a researcher.

Secondary analysis: analysis of data by researchers unconnected with the original purposes of the data collection, as where academic researchers use data sets gathered as a part of government social surveys.

Secondary sources: analyses or restatements of primary sources (records of events as they are first described or original data) by other authors or researchers. Secondary sources might take the form of research reports, news articles, biographies, documentaries or history books) used to gain an understanding of a topic. Primary sources might be poems, raw tabulations of census data, video recordings or other records of observation. The use of secondary sources should be distinguished from secondary analysis of other researchers' original data (a primary source).

Social constructionism: the view that the phenomena of the social and cultural world and their meanings are created in human social interaction. Taken further, social constructionism can be applied to social research itself, prompting debates about whether social research and fiction differ. The approach often, though not exclusively, draws on idealist philosophical orientations.

Social facts: regularities of social life that appear to have an independent existence, acting to determine or constrain human behaviour. Norms of conduct or religious rules are examples. The concept is of particular importance in relation to functionalism and positivism.

Social structure: ordered interrelationships that are characteristic of particular societies, such as its class structure or system of economic or political relations.

Standpoint models: involve the assumption that different social positions produce different experiences and therefore lead to different types of knowledge. Because of this researchers often engage with the experiences of socially oppressed and marginalised groups. The knowledge derived from this is felt to provide a more valid account of the social world than adopting an apparently 'neutral' or 'objective' stance.

Statistical inference: the generalization of findings from a sample to the broader population from which the sample has been randomly drawn. A variety of statistical tests, such as the chi-square, help in estimating the level of probability that such inferences about the population are true, given the sample size. This is expressed as the statistical significance of the finding.

Structuralism: the view that behind the social and cultural realities we perceive, such as clothes or food fashions, kinship organization and even language itself, deep structures exist which, through combinations of their elements, produce the surface complexity of the relevant phenomena. Poststructuralism retains elements of structuralism (its interest in surface signs for example) but abandons the quest for deep structures.

Symbolic interactionism: a body of theory that emphasises the organisation of everyday social life around events and actions that act as symbols to which actors orient themselves. Interactionists frequently study this through observation of face to face interaction and a preferred method for doing this is ethnography.

Text: although this term includes the kind of thing we usually mean by 'text' (eg: a written document) under the conditions of the literary turn, structuralism and poststructuralism the term can be applied to almost any object in the world. Semioticians, for example, have considered items as diverse as wrestling matches and Coca Cola cans as 'texts', worthy of analysis for their cultural connotations

Theoretical sampling: choosing a sampling element (eg: a person, a social setting) on the basis of its likely contribution to a (grounded) theory emerging during the course of a study.

Thick description: a term adapted by the anthropologist Clifford Geertz to convey the essence of his semiotic approach to ethnography, based on intensive observation of social life from which interpretations of cultural signs can be generated, as for example in the many layered meanings of a Balinese cock fight.

Topicalising an account: refers to an analytic approach to texts (such as an interview) in which the focus of interest is the world constructed in the text and the methodsmby which the text achieves this. This involves bracketing the truth claims of the text, refusing (at least temporarily) to treat it as a resource for discovering truths about some reality outside the text.

Triangulation: a metaphor derived from surveying and navigation to indicate the convergence of two or more viewpoints on a single position or, in social research, truth. Triangulation expercise migh, for example, involve seeing whether the results of a questionnaire are repeated in observational data. Associated with a realist approach and, largely, with early qualitative discussions of validity, triangulation is treated with scepticism by non-realists who reject the view that revelation of a single truth is the object of a research account.

Validity: at its most simple this refers to the truth status of research reports. However, a great variety of techniques for establishing the validity of measuring devices and research designs has been established, both for quantitative and qualitative research. More broadly, the status of research as truth is the subject of considerable philosophical controversy, lying at the heart of the debate about post-modernism. A convenient way of categorising concerns about validity is to divide these into internal and external. The former refers to the internal design of a study (for example, can it prove causality?); the latter refers to the generalisability of a study (for example, does the sample represent a population adequately?)

Variables: qualities on which units of analysis vary. Thus, if a person is the unit of analysis in, say, a social survey, examples of variables might be their social class, gender, attitudes to politics, and so on. Variables can be measured at a variety of levels, according to which they can be subjected to specific mathematical operations. In considering relationships between variables it is important to define which is a causal (or independent) variable, and which is an effect (dependent) variable.

Verstehen: Max Weber used this word to describe the study of intersubjectivity, involving an attempt to understand the meaning of social action from the actor's viewpoint.