Classification by purpose
Social research can be classified by data-collection method or time dimension, but also by purpose. Since Selltiz et al. (1959), methods textbooks have generally distinguished exploratory, descriptive and explanatory research. The three are less mutually exclusive categories than stages in the growth of knowledge, and a single study often serves several purposes. The overall flow of a study is covered in the social work research process.
Exploratory research
Exploratory research is used when little is known about a topic, to grasp the problem and find directions and hypotheses for further study. It asks, "What is going on here?"
- Aims: define the problem, discover key variables, generate hypotheses, test the feasibility of a larger study
- Methods: literature reviews, expert or key-informant surveys, in-depth interviews with a few cases, focus groups, field observation, pilot studies
- Features: flexible procedures, small samples and much qualitative data; findings are hard to generalize.
- Example: interviewing counselors and young people to learn about the nature and causes of newly emerging online gambling among teenagers
Descriptive research
Descriptive research portrays the characteristics and distribution of a phenomenon and the relationships among variables accurately. It asks "how many, who and how."
- Aims: establish the facts, estimate frequencies and proportions, describe correlations
- Methods: surveys of representative samples, censuses, cross-sectional and longitudinal (trend, panel) surveys, analysis of existing statistics
- Features: sample representativeness and measurement accuracy are crucial, and what to measure must be clear in advance. It shows correlations but cannot establish causation.
- Example: a survey of the number of runaway youths nationwide, their age and sex distribution, length of time away and living conditions; a survey of older adults' use of local welfare services
Explanatory research
Explanatory research asks why a phenomenon occurs and whether variables are causally related. It asks "why?"
- Aims: test hypotheses, identify causes, predict outcomes, evaluate program effects
- Methods: experimental and quasi-experimental designs, causal models such as multiple regression and path analysis, longitudinal data analysis
- Conditions for causal inference: the cause must precede the effect (temporal order), the variables must covary, and alternative explanations (third variables) must be ruled out.
- Example: comparing experimental and control groups to see whether a mentoring program raises adolescents' self-esteem
Comparing and choosing
- State of knowledge: when almost nothing is known, explore; when the phenomenon is known but its extent is not, describe; when there is a theory about causes, explain.
- Degree of structure: exploratory designs are flexible, while descriptive and explanatory designs follow predetermined procedures and standardized measures, typically within quantitative research.
- Sequence: a typical progression is to form hypotheses through exploration, establish the facts through description and then test causes through explanation.
Some textbooks add evaluation research, which applies findings to improve policies or programs, as a separate type. Exam questions often present a scenario and ask which type it is; checking whether the research question is about "what," "how many" or "why" usually settles it.
References
- Selltiz, C., Jahoda, M., Deutsch, M., & Cook, S. W. (1959). Research Methods in Social Relations (rev. ed.). New York: Holt, Rinehart and Winston.
- Neuman, W. L. (2014). Social Research Methods: Qualitative and Quantitative Approaches (7th ed.). Harlow: Pearson.
- Rubin, A., & Babbie, E. R. (2017). Research Methods for Social Work (9th ed.). Boston: Cengage Learning.
- Babbie, E. R. (2021). The Practice of Social Research (15th ed.). Boston: Cengage.