Designing a Study You Can Defend
Multiple outcomes do not automatically require the same statistical adjustment, but studies making several confirmatory claims need a defensible multiplicity strategy. Learn when planning is necessary, how different outcome structures affect error control, and why these decisions should be made before data collection.
Read Guide →
Designing a Study You Can Defend
Multiple treatment arms can improve research efficiency, but they may also create additional opportunities for false-positive conclusions. Learn when multiplicity planning is necessary, how shared controls affect testing, and why the appropriate adjustment depends on the study's scientific claims.
Read Guide →
Designing a Study You Can Defend
Measuring numerous outcomes can make it easier to find an apparently significant result, particularly when researchers selectively report favorable findings. Learn how to control outcome proliferation without discouraging legitimate exploration or overlooking scientifically important results.
Read Guide →
Designing a Study You Can Defend
A broad research project may contain several legitimate questions that cannot all be answered convincingly within one design. Learn when separate studies improve methodological rigor, when integration is preferable, and how to avoid unnecessary fragmentation or duplicate publication.
Read Guide →
Designing a Study You Can Defend
Demographic characteristics may be important for assessing representation, equity, and variation in intervention effects, but they do not automatically justify separate hypothesis tests. Learn when demographic subgroup analyses are scientifically appropriate and how to interpret them without reinforcing unsupported assumptions.
Read Guide →
Designing a Study You Can Defend
Separate within-group tests cannot establish whether an intervention works differently across subgroups. Learn why interaction testing should generally be planned for effect-modification questions, how to specify the analysis, and when within-group estimates remain useful.
Read Guide →
Designing a Study You Can Defend
Oversampling underrepresented groups can improve the precision of subgroup estimates and support meaningful comparisons, but it is not automatically necessary or sufficient. Learn when disproportionate recruitment is justified, how to plan it, and how to preserve valid population-level inference.
Read Guide →
Designing a Study You Can Defend
An underpowered subgroup analysis may identify an interesting pattern without providing reliable evidence of differential effects. Learn when a separate, purpose-designed study offers stronger evidence and when existing subgroup analyses remain scientifically useful.
Read Guide →
Designing a Study You Can Defend
Representation asks whether relevant people or groups are present in research. Representativeness asks how adequately the sample reflects a defined target population for the inference being made.
Read Guide →
Designing a Study You Can Defend
A diverse sample can help researchers study variation that a narrower sample might miss, but diversity is not automatically the same as representativeness. What matters is which differences are relevant to the research question and the conclusions being drawn.
Read Guide →
Designing a Study You Can Defend
Sex and gender can both matter in research, but they do not necessarily represent the same constructs. Distinguishing them helps researchers measure the variable they actually need and interpret observed differences more carefully.
Read Guide →
Designing a Study You Can Defend
Analyzing results by sex or gender can reveal important differences that an overall average conceals, but not every dataset supports a meaningful subgroup comparison. The decision should follow from the research question, study design, sample size, and relevant scientific or reporting requirements.
Read Guide →
Designing a Study You Can Defend
Research participation becomes more accessible when researchers examine what the study requires of participants and remove barriers that are not necessary to answer the research question. Accessibility can involve communication, disability accommodations, scheduling, location, technology, language, and participant burden.
Read Guide →
Designing a Study You Can Defend
Adding participants from more groups can broaden whose experiences appear in a study, but diversity alone does not establish representativeness. Representativeness depends on the target population, selection process, and inference researchers want to make.
Read Guide →
Designing a Study You Can Defend
Inclusive research does not mean including everyone. Researchers need a population broad enough to capture relevant variation but sufficiently defined to answer the research question safely and meaningfully.
Read Guide →
Designing a Study You Can Defend
Eligibility criteria can protect participants and strengthen research, but overly restrictive requirements may exclude people who could safely and meaningfully participate. Learn how to distinguish necessary restrictions from avoidable exclusion and develop scientifically defensible criteria.
Read Guide →
Designing a Study You Can Defend
Requiring internet access can exclude otherwise eligible research participants and distort findings when digital access is related to the research question. Learn how to assess coverage gaps, distinguish internet access from digital capability, and design accessible participation alternatives.
Read Guide →
Designing a Study You Can Defend
Travel requirements can prevent otherwise eligible people from participating in research, particularly when transportation costs, distance, mobility needs, or lost working time create substantial burdens. Learn how to evaluate these barriers and reduce avoidable exclusion without compromising participant safety or data quality.
Read Guide →
Designing a Study You Can Defend
A study can follow rigorous methodological procedures while systematically excluding people most affected by the research problem. Learn why internal validity does not guarantee population relevance, how exclusion can distort interpretation, and what researchers can do to align scientific rigor with meaningful population coverage.
Read Guide →
Designing a Study You Can Defend
A data collection method may fit your research question conceptually but still be inappropriate if it places unreasonable demands on participants or cannot be implemented well. Learn how to balance evidence quality with burden and feasibility.
Read Guide →
Designing a Study You Can Defend
Participant burden must be justified by the scientific value and necessity of research procedures, not merely by participants' willingness to comply. Learn how to assess time, effort, financial costs, discomfort, and cumulative demands while protecting scientific validity and participant welfare.
Read Guide →
Designing a Study You Can Defend
The number of follow-up visits should reflect when meaningful changes are expected, which outcomes must be measured, and what monitoring is necessary. Learn how to design an efficient follow-up schedule without compromising scientific validity or imposing unnecessary participant burden.
Read Guide →
Designing a Study You Can Defend
Removing nonessential measurements can reduce participant fatigue, costs, and incomplete data, but eliminating important assessments may compromise validity or safety. Learn how to identify redundant procedures, evaluate their incremental scientific value, and simplify research protocols without losing necessary evidence.
Read Guide →
Designing a Study You Can Defend
Scheduling research around participant availability can reduce unnecessary barriers and improve opportunities for participation. Learn when flexibility is appropriate, how to establish defensible assessment windows, and how to protect scientific validity when research activities occur at different times.
Read Guide →
Designing a Study You Can Defend
Researchers may underestimate the time, expense, discomfort, and practical difficulties their procedures impose. Learn how participants and community partners can identify overlooked burdens, inform protocol improvements, and strengthen research feasibility without compromising scientific quality.
Read Guide →