Which Research Methods Dominate the Existing Literature?

How Has the Field Usually Tried to Produce Its Evidence?

When reading a large body of literature, it is easy to focus on what researchers found and pay less attention to how they were able to find it. Yet methods shape the kinds of observations researchers make, the comparisons they can support, the biases they must confront, and ultimately the claims the literature can reasonably sustain.

A field may rely heavily on cross-sectional surveys. Another may be dominated by experiments, interviews, administrative datasets, case studies, longitudinal cohorts, or secondary analyses. Even within the same broad design, researchers may repeatedly use similar instruments, sampling strategies, data sources, analytical procedures, or time frames.

Identifying these patterns tells you how the field has constructed its evidence. The important task is not to declare the most common method good or bad, but to determine what its dominance allows researchers to understand and what questions may remain difficult to answer because other methodological approaches are uncommon.

Map Study Designs and Methodological Features Across the Evidence Base

How to Map the Methodological Structure of a Literature

Do Not Reduce Method to “Quantitative” or “Qualitative”

A useful methods map needs more resolution than broad methodological families. Two quantitative studies may differ fundamentally if one is a randomized experiment and the other is a cross-sectional observational survey. Two qualitative studies may use different forms of data collection, sampling, analytical traditions, and temporal designs.

Depending on the research area, useful characteristics to code may include:

• overall methodological approach;
• study design and important design features;
• cross-sectional, longitudinal, retrospective, or prospective structure;
• experimental or observational allocation;
• sampling and recruitment strategy;
• data source and data-collection method;
• measurement instruments;
• unit of analysis;
• analytical approach;
• follow-up duration; and
• whether multiple forms of evidence are integrated.

The appropriate categories depend on the substantive question and discipline. Do not force studies into a generic taxonomy merely because it is convenient to count.

Study Labels Are Useful, but Design Features Matter More

Researchers often classify studies using labels such as randomized trial, cohort study, case-control study, survey, ethnography, or case study. These labels are useful for organizing a literature, but they can conceal substantial variation.

Cochrane's guidance on non-randomized intervention studies explicitly recommends attention to specific design features rather than relying exclusively on study-design labels, partly because labels can be used inconsistently. Features such as how groups were formed, when measurements occurred, and how confounding was addressed may be more informative about what a design can establish.

Your methods map should therefore record enough detail to distinguish studies that share a label but differ in evidential capability.

Count Methods at More Than One Level

Methodological dominance can occur at several levels. A field may appear diverse at one level while being highly concentrated at another.

For example, a literature might contain both quantitative and qualitative studies yet rely overwhelmingly on self-reported data in both. Calling the field methodologically diverse without noticing that shared measurement pattern would miss an important concentration.

Use Counts to Reveal Patterns, Not to Rank Methods

Once studies are coded, calculate frequencies or proportions where appropriate. You might find that 70% of eligible studies are cross-sectional, that most use self-report questionnaires, or that only a small minority include follow-up observations.

These descriptive patterns are valuable, but frequency does not establish methodological superiority. The most common method may dominate because it is especially appropriate, historically established, inexpensive, accessible, ethically feasible, or simply conventional.

The first can be established by mapping the literature. The second cannot be inferred from popularity.

Ask Which Questions the Dominant Methods Are Good at Answering

This is where methods mapping becomes substantive.

Cross-sectional surveys can efficiently describe variables and associations at a particular point or period, but they generally provide less leverage for establishing temporal sequences than longitudinal designs. Randomized experiments can be particularly informative for estimating causal effects of interventions under appropriate conditions, while observational designs may be necessary or valuable where randomization is infeasible, unethical, or unable to address the relevant setting, exposure, population, or outcome.

Cochrane notes that some intervention questions cannot be answered by randomized trials alone and that non-randomized studies may provide evidence about issues such as long-term or rare outcomes, different populations, settings, or forms of intervention delivery. At the same time, non-randomized intervention studies require careful attention to confounding and other sources of bias.

The implication is not that a literature needs equal numbers of every design. It needs methods capable of addressing the questions researchers want to answer.

Look for Method-Population Combinations

Methodological patterns can be obscured when methods and populations are examined separately.

Perhaps the literature includes both experiments and qualitative studies, but experiments are conducted almost exclusively with university students while qualitative research involves practicing professionals. Perhaps longitudinal studies exist, but only in one country. Perhaps clinical populations appear primarily in observational research while experimental evidence comes from healthy volunteers.

Cross-referencing methods with populations that appear repeatedly can reveal which groups support particular types of claims.

Look for Method-Outcome Combinations Too

The same principle applies to outcomes. Certain methods may repeatedly be paired with particular outcomes because those outcomes are easier to measure using that design.

For example, short questionnaires may dominate studies of attitudes while behavioral outcomes appear mostly in experiments. Long-term consequences may be rare because the literature relies on designs with brief observation periods.

Mapping which outcomes researchers measure repeatedly alongside methodological patterns can show whether the evidence base systematically privileges certain kinds of findings.

Repeated Use of One Method Can Produce Strong Evidence

Methodological repetition is not inherently redundant. Repeated well-designed studies can replicate findings, improve precision, test robustness, or examine variation across contexts.

A mature experimental literature may appropriately contain many randomized studies. A qualitative research program may develop increasingly rich understanding through repeated work in different contexts. A long-running cohort may produce valuable longitudinal evidence unavailable through other designs.

The useful question is whether repeated use of the method continues to generate informative evidence.

Methodological Concentration Can Also Create Blind Spots

Problems arise when the dominant methods systematically cannot address questions the field nevertheless wants to answer.

A literature composed mainly of one-time surveys may repeatedly establish correlations while remaining uncertain about temporal processes. A field relying heavily on self-reports may know much about reported perceptions but less about observed behavior. Experimental studies may estimate intervention effects while providing limited understanding of implementation experiences if those processes are never examined.

Mixed methods can sometimes address questions requiring complementary forms of evidence by intentionally integrating quantitative and qualitative approaches. NIH's Office of Behavioral and Social Sciences Research describes mixed methods research as integrating both forms of data to provide a more comprehensive understanding of a research problem when appropriate. This does not make mixed methods automatically superior; integration should serve a research purpose.

Methodological Dominance Helps Explain Why Some Questions Remain Unresolved

Once you have mapped dominant methods, compare them with questions the literature has only partly answered.

The connection may be revealing. Perhaps mechanisms remain uncertain because studies repeatedly estimate outcomes without examining processes. Perhaps long-term effects remain unknown because follow-up periods are short. Perhaps causal claims remain uncertain because the available designs provide limited leverage on confounding or temporal ordering.

The gap, in such cases, is not merely “Method X has not been used.” It is that the current methodological structure leaves a substantive question difficult to answer.

Mapping a Literature Dominated by Cross-Sectional Surveys

Common Mistakes When Identifying Dominant Methods

Ask What the Dominant Methods Make Visible and What They Leave Difficult to See

Once methodological concentration is visible, resist the temptation to propose the opposite method simply for novelty. Instead, connect the methods map to unresolved substantive questions.

A strong methodological map should ultimately help you explain why the literature knows some things well and struggles to answer others.

Before Concluding That a Method Dominates the Literature

Questions About Dominant Research Methods

Dominant Methods Define Much of What a Field Can See

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.