Search across guide titles, topics, keywords, and content throughout the Researcher's Field Guide.
Showing 101–125 of 2,119 results
Poor research design is not merely a technical problem when people bear burdens or risks for knowledge the study is unlikely to produce. Methodological quality can therefore become part of the ethical justification for research.
Read Guide →Vulnerability is not always something participants bring into a study. Recruitment methods, incentives, privacy practices, authority relationships, and other design choices can create or intensify vulnerabilities that might otherwise have been absent.
Read Guide →A study can have well-written research questions, objectives, hypotheses, and methods that do not actually fit together. Learn how generative AI can help identify conceptual and methodological inconsistencies before they compromise the evidence your study can produce.
Read Guide →Generative AI can help researchers explore and compare study designs, but its recommendations may overlook methodological assumptions, validity threats, and practical constraints. Learn how to use AI to evaluate design options while retaining responsibility for scientific decisions.
Read Guide →Generative AI can explain differences among research designs, but it may confuse methodological approaches, assignment procedures, and data collection methods. Learn how to classify designs accurately by examining what researchers actually do rather than relying on labels.
Read Guide →Generative AI can help researchers plan mixed-methods studies by comparing design structures, organizing quantitative and qualitative components, and identifying opportunities for integration. Learn why collecting two types of data is not enough and how to evaluate whether mixed methods genuinely serves the research question.
Read Guide →The strongest study on paper is not always the strongest study you should conduct. Learn how to distinguish necessary rigor from unnecessary ambition and design research that is both credible and realistically completable.
Read Guide →Community involvement may need to begin before a population-specific study is designed when community knowledge is necessary to define the problem, choose meaningful outcomes, develop feasible procedures, or avoid harmful assumptions. Not every study requires co-design, but consultation that arrives after all consequential decisions have been made may arrive too late.
Read Guide →A worthwhile research question does not automatically justify the first study you design around it. Compare credible alternatives and ask which approach would reduce the important uncertainty most effectively.
Read Guide →A stronger design is not simply a more complicated or prestigious design. It is one that addresses the weaknesses preventing existing evidence from supporting the specific inference your research question requires.
Read Guide →A research design should be chosen according to what the study must establish, not simply which method is familiar or available. Learn how to work backward from your intended conclusion to the evidence, measurements, comparisons, and design capable of supporting it.
Read Guide →The same broad research topic can often be studied cross-sectionally or longitudinally, but the two designs do not answer exactly the same question. Changing the temporal structure changes what evidence is available and therefore what conclusions the study can support.
Read Guide →An additional time point matters when it changes the temporal pattern or competing explanation a study can examine. The methodological value comes not from having more observations, but from what the new measurement allows researchers to distinguish.
Read Guide →Two research groups do not need to be identical to support a meaningful comparison. What matters is whether their differences threaten the intended inference. Learn how to evaluate comparability, distinguish descriptive from causal requirements, and address selection, confounding, and limited overlap without assuming that statistical adjustment can solve every problem.
Read Guide →A statistically convincing association is not automatically a causal effect. Learn how to recognize when your design supports a relationship between variables but cannot credibly establish what would happen if one were changed.
Read Guide →Selection bias often begins before researchers open their statistical software. Learn how to audit sampling frames, eligibility rules, recruitment, consent, follow-up, and analytical inclusion before choosing adjustment methods, and distinguish problems that statistical analysis can address from those requiring better design.
Read Guide →Electronic health records and administrative databases can support valuable research without requiring new data collection. However, records created for clinical or operational purposes may contain missing information, changing definitions, selection effects, and measurement errors. Learn how to determine whether these systems can provide defensible evidence for a research question.
Read Guide →Generative AI may recommend sophisticated research designs that appear rigorous but exceed available time, resources, expertise, or institutional access. Learn how to distinguish justified methodological complexity from impractical recommendations and select a design that can be implemented without compromising scientific integrity.
Read Guide →Generative AI may recommend research designs that appear methodologically sound but overlook the requirements of particular disciplines, institutions, or research settings. Learn how to evaluate contextual constraints before adopting AI-generated methodological recommendations.
Read Guide →Research is more than searching for information. It is a systematic and disciplined process of investigating questions, examining evidence, and developing knowledge that can withstand scrutiny.
Read Guide →Looking up reviews before buying a laptop and conducting a scientific study both involve seeking answers, but they operate under different standards. The key differences concern how questions are investigated, evidence is evaluated, and conclusions are justified.
Read Guide →Research is generally expected to contribute to knowledge, but that does not mean every study must discover something completely unprecedented. Novelty can arise through new evidence, contexts, explanations, tests, methods, interpretations, or clearer knowledge about the limits of existing claims.
Read Guide →Research does not have to pursue a completely new claim. Confirmatory and replication studies can make important contributions by testing whether existing findings withstand additional scrutiny, hold under new conditions, or require qualification.
Read Guide →Research does not have to collect new data. Existing datasets, records, archives, documents, and other previously created materials can support rigorous and original research when they are appropriate to a new research question.
Read Guide →Research does not have to involve an experiment or test a hypothesis. Experiments and hypothesis testing are powerful approaches for particular questions, but observational, exploratory, descriptive, qualitative, historical, and other forms of research may require different designs.
Read Guide →