Social Research in Brief

            Social Research in Brief


Written by Norman Blaikie in Analyzing Quantitative Data From Description to Explanation. London: SAGE. 2003. PP.34-36.


- Social research must start with a research problem, an intellectual puzzle or a practical problem.

Image:www.digital-mr.com

- Social research is about answering three types of research questions: ‘what’,‘why’ and ‘how’ questions.

- Social research pursues a range of objectives: exploration, description, explanation, understanding, prediction, intervention, evaluation and impact assessment. The objectives of explanation and understanding are expressed as ‘why’ questions and the objective of intervention as ‘how’ questions. The remaining objectives are mostly related to ‘what’ questions.

- Research objectives are frequently pursued in a logical sequence, the most common of which is description, explanation/understanding and intervention.

- Theoretical hypotheses provide possible answers to ‘why’ research questions.

- Statistical hypotheses are used to establish whether patterns found in a random sample are present in its population. This is their only role in social research.

- Data are produced by the use of the human senses, mainly sight and hearing,and through the use of instruments that extend and systematize their use. This requires agreement about rules and criteria. Such procedures do not guarantee objectivity, only comparability between times, places and researchers.

- All forms of measurement in the social sciences are socially constructed by experts, the data they produce, and the results that follow, have to be understood in terms of the assumptions and procedures adopted.

- These assumptions are both ontological and epistemological and, while they are usually taken for granted, they can be understood with reference to one of the major philosophies of social science: positivism, critical rationalism, scientific realism and interpretivism.

- There are three types of social science data: primary, secondary and tertiary. Each type has its advantages and disadvantages and varies in terms of the distance it creates between the researcher and the social reality being studied.

- Social science data can be either qualitative or quantitative, in either words or numbers. Transformations between words and numbers, or in the reverse direction, can occur at various stages in a research project.

- Quantitative data are expressed in the form of variables that are produced by operationalizing the key concepts in research questions and theoretical hypotheses.

- Concepts can be measured at four different levels. From lowest to highest,these are nominal, ordinal, interval and ratio. The first two produce categorical variables, because objects, events or people are placed into one of a set of mutually exclusive categories. The second two produce metric variables,as objects, events or people are mapped onto an established measuring scale.

- Metric variables can be either discrete or continuous. The former consist only of whole numbers, while the latter have an unlimited number of possible values between the whole numbers.

- Data can be transformed from metric to categorical. While this means some loss of information, and entails the use of less sophisticated forms of analysis, it may allow for a better understanding of the characteristics or relationships being examined.

- There are four main types of data analysis: univariate descriptive, bivariate descriptive, explanatory and inferential. The first two are concerned with characteristics and patterns in data, the third with influence between variables and the fourth with generalizing from samples to populations. Explanatory analysis is the ultimate objective in social research and is also the most complex.

- Explanation is usually associated with the idea of causation. However, this is a highly contested notion and has to be reduced to simpler ideas to be Social research and data analysis useful in social research. One way of doing this is in terms of the influence between independent (predictor) and dependent (outcome) variables.

- Different views of causation are associated with the major logics of enquiry: inductive, deductive, retroductive and abductive. These logics also constitute different research strategies.