Method type. The method type is how one chooses to gather evidence based on previous research and the data available, e.g. researchers trying to decipher Egyptian hieroglyphics could only guess what the symbols meant until archaeologists discovered the Rossetti Stone, a tablet with side-by-side Greek and Egyptian versions of the same text. In contrast, physicists investigating the atom could carry out experiments to discover its structure. In general, research methods are grouped according to whether they discover, describe, correlate or causally link constructs as follows:
- Exploratory. Exploratory methods use data to discover initially unknown constructs, e.g. grounded theory, a method used to study new cultures.
- Descriptive. Descriptive methods gather data to better define constructs in complex areas where the constructs are unclear, e.g. case studies or political focus groups.
- Correlational. Correlational methods show connections between constructs but not what causes what, e.g. surveys or longitudinal studies in psychology.
- Experimental. Experimental methods establish cause and effect by manipulating a independent variable to see its effect on a dependent variable, e.g. laboratory studies.
For example, historians can describe history and sometimes correlate events but cannot experiment on what is past. None of the above methods are better or worse, just more or less suited for the research situation. To use exploratory research for a well-established topic may be as misplaced as to do an experiment in a poorly understood area. Choose a research method that suits your topic area.
Realism vs. control. Research methods differ in how realistic the data is and how much control the researcher has over it. In general, control increases in the order exploratory, descriptive, correlational and experimental, while realism decreases in the same order, e. g. case studies are more realistic than experiments but offer less control. Exploratory research is the most widely applicable but experiments are the most demanding. Choose a research method that is as realistic as possible while still being feasible.
Data types. Data can be qualitative or quantitative. The research method chooses the type of data collected, which can be:
- Qualitative. The data is semantic, e.g. a subject statement.
- Quantitative. The data is numeric, e.g. a response count.
- Mixed Method. A combination of both the above.
For example, to investigate web site usage one can gather the number of views to give quantitative data, or interview people who use it to give the qualitative opinions. The properties of good research, like validity, reliability, repeatability and generalizability, apply equally to qualitative and quantitative data. Each has different benefits, so while qualitative studies struggle for precision, numeric studies struggle for meaning. In practice, qualitative and quantitative research are complementary, so consider a mixed-method approach that gathers both qualitative and quantitative data to get the best of both worlds, like the Repository Grid or Delphi method. Interviewing subjects after an experiment can be very helpful to decide what the results mean. Gather both qualitative and quantitative data for the best results.
Method or analysis theory. If the method is uncommon or new, briefly present the theory behind it, e. g. Delphi method, factor analysis theory, interpretive method theory or conjoint analysis theory. Note that the method theory is not the research theory, as the latter is central to your research but the former is just an aspect of it. So keep it brief – readers who are already familiar with factor analysis don’t want to read pages of factor analysis theory copied from a textbook! Research theory needs extended coverage but method theory does not, so refer readers elsewhere for details. Briefly present the method theory if it is uncommon or new.