Last updated on January 13, 2023
The methods section of a research paper is kind of like the instructions you get with a piece of furniture from IKEA. It tells you how the furniture was or will be put together but not why it works, why we care, and it’s not the thing itself. The methods section should be seen as the means for which the reader can critically examine a study to understand the breadth and limits of the conclusions. Methods sections can also be useful as a guide for how to do research yourself, especially with newer or more advanced research methods.
How you organize your methods section will depend on whether the research you’ve carried out is qualitative or quantitative and depending on the requirements of the journal (or school/prof if for a thesis or term paper). Nonetheless, in general, but not always, methods sections are subdivided into sections for: Data, Sample, Measures, Method, and Ethics. If submitting to a journal, read the instructions for authors for general requirements and then review other articles they’ve published for more ideas on what specifically to include. I’ve listed some general suggestions below, though, to get you started.
Any research study should detail what data was used. Sometimes this is as simple as naming a survey, such as the General Social Survey (2018). Other times, when the researchers collect data themselves, they’ll need to offer more details as to what the data include. Important pieces of information readers will want to know might include: where the data were collected; what time frame the data were collected; how the data were collected (e.g. phone interviews, online interviews, in-person interviews); if there changes made to a pre-existing data set or only parts of the data were used; how the questionnaire/interview guide/coding schema was developed; and so on.
For ethnographic field work, you might discuss how you took notes while you were out in the field and what constituted the “data” of your study. Did you include thick descriptions of your observations every night after returning from the field?
For interview data, you might include a discussion of how the interviews were transcribed and what you did with the data once collected (did you put it in excel or another software package like NVivo or Atlas.ti?)
The sample is “who” you included in your study. If you interviewed people, for example, this section should discuss who these people are. Are they of a certain age? From a certain region?
You also need to discuss your sampling method. Did you use a random sample or a convenience sample? In other words, how did you find these people? Did you put out an ad in the newspaper? Did you ask your friends to participate? Did you have a list of people from which you drew a random selection?
Sometimes, qualitative research will offer some basic statistics here on the composition of the sample, although in quantitative research this is usually saved for the discussion section.
If doing quantitative inferential statistics you’ll want to also include a discussion of the population your sample is supposed to represent. In large surveys, this is typically included in the supplemental materials that come with data. For example, many Statistics Canada data sets are sampled to represent all Canadians over the age of 18 who are not in the military and do not live in the Northern Territories, on First Nations Reservations, or in Institutions; and only if you use the proper weights in the data. (Note that limitations or restrictions in how the data were sampled pose restrictions in the conclusions you can draw at the end of the paper. If the data do not include people under the age of 18, you cannot conclude that your study shows anything about young people. This isn’t bad, per se, just something to pay attention to.)
You also should identify the overall sample size and how many cases were lost from the original sample for whatever reason (e.g. I didn’t record the first phone interview I ever did in a qualitative study so I mentioned that; in quant papers this usually means how many cases were dropped b/c the respondent didn’t answer one or more questions).
Measures often are not included in qualitative or other inductive types of research. The goal of such projects is not to start with specific concepts or theories that you’ll test, but rather to explore the data and see what emerges. However, for quantitative or other deductive studies you will need to be very explicit as to what your measures are. The measures should match precisely with your hypotheses which should match your theory and review of the literature. For example, if you want to assess whether poverty leads to poor health, you need to have a measure of poverty and a measure of health. In your measures section, you will describe what variables in your data you used to measure poverty (e.g. income measured to the penny or a cut-point or a relative poverty line) and what you used to measure health (e.g. a self-reported five-point scale from 1 for I feel like crap all the time to 5 for I feel amazing; or perhaps a count of how many times the respondent was prescribed antibiotics in the previous year).
The methods is, well, the method that you used to analyse the data. For quantitative studies using secondary data, this will be the specific statistical tests you used, such as descriptive statistics with t-tests to assess whether there are statistically significant differences between the two groups. More more advanced methods, often times the statistical equations will be included to make it clearer for number crunchers to understand what it was that you did exactly (and to flex about how smart you are).
For qualitative studies, this is a place to discuss more the particular steps you took to analyse your data after they were collected. You may include details as to what software you used such as Atlas.ti or NVivo to identify key themes.
For content analyses, you might also discuss inter-rater reliability or other means you used to ensure that your conclusions were valid.
Sometimes this is not included for content analyses or survey research if the author(s) used secondary data that did not require ethics approval. Sometimes, a kind of pat phrase will be included like “This research was carried out in accordance with the ethical principles of the institutions and national laws of the authors.”
If the research involved an original survey, an experiment, qualitative interviews, ethnographic field work, or anything involving human subjects, there will need to be a clear description of the ethics procedure and often refer to the specific ethics approval code given by the Research Ethics Board (REB in Canada)/ Institutional Review Board (IRB in the US). Journals typically will not publish research that does not clearly articulate that the researchers engaged in ethical research and received REB/IRB approval.
Methods sections often begin by referring to the hypotheses or arguments but should not include them or state them, leave that in the lit review/theory section
Methods sections should usually not include any analysis (aside from qualitative studies that present some statistics on the sample sometimes)