Part 9 · Module 1
Principles of Experimental Research
Explain how investigator-assigned interventions, counterfactual comparisons and protocol control distinguish experimental research.
Learning outcomes
After completing this lecture, you should be able to:
- Define an experiment and its counterfactual comparison.
- Distinguish intervention assignment from observation of routine exposure.
- Explain control, replication, randomisation and standardisation.
- Identify ethical and practical limits of experimentation.
- Align the experimental question with an intervention contrast and outcome.
Overview
Experimental research deliberately assigns an intervention or strategy and observes outcomes under controlled comparison. The investigator specifies what is delivered, to whom, when and against which alternative. Randomisation is a powerful assignment method, but experimentation is broader than randomised trials and always requires a clearly defined intervention contrast.
The study question, design, participant pathway, measurements and analysis must be specified as one coherent system. A recognised design label is useful only when the operational details preserve the comparison or prediction that the study is intended to support.[1–3]
Why this matters
Health interventions are complex events rather than labels. An apparent effect may reflect co-interventions, expectations, differential follow-up or changes in usual care. Experimental design makes these mechanisms visible by prespecifying the intervention, comparator, assignment, outcome schedule and participant pathway.
Core concepts
Intervention assignment defines the experiment
In an experiment, exposure to a study strategy is determined by a protocol rather than solely by participant or clinician choice. The intervention can be a medicine, procedure, behavioural programme, diagnostic strategy, policy or system change.
The causal question compares potential outcomes
For each participant, only the outcome under the assigned strategy is observed. The unobserved outcome under the alternative is counterfactual. A concurrent control group supplies an estimate of what would have occurred under the alternative, subject to design and adherence assumptions.
Control means more than a control group
Control includes standardised eligibility, delivery, measurement, follow-up and analysis. Excessive control can reduce applicability, while insufficient control can make it impossible to attribute differences to the intended intervention.
Randomisation protects assignment
Random allocation makes treatment assignment unpredictable and, on average, balances baseline causes of outcome. It does not guarantee perfect balance in one sample, prevent post-randomisation bias or compensate for poor outcome measurement.
Experiments require intervention equipoise and feasibility
The comparison must be ethically acceptable, deliverable and relevant. Interventions with established major benefit should not be withheld merely to create a control. Harms, burden, rescue treatment and stopping processes belong in the design.
When to use this approach
Is an experimental approach justified?
- Can the intervention and comparator be assigned ethically?
- Is there genuine uncertainty about their relative effects?
- Can delivery be standardised and monitored?
- Can contamination and co-interventions be measured?
- Will the experimental setting still inform the intended practice decision?
Working through the key design decisions
Decision 1: Can the intervention and comparator be assigned ethically?. This point should be resolved before recruitment, extraction or assignment begins. For principles of experimental research, the protocol should state the operational rule, who applies it, when it is applied and what source documents demonstrate compliance. The investigators should also consider how a different choice would change the eligible population, comparison, timing, measurement or interpretation. A pilot review of realistic records or participant pathways can reveal ambiguities before they become irreversible exclusions or post hoc analytical choices.
Decision 2: Is there genuine uncertainty about their relative effects?. This point should be resolved before recruitment, extraction or assignment begins. For principles of experimental research, the protocol should state the operational rule, who applies it, when it is applied and what source documents demonstrate compliance. The investigators should also consider how a different choice would change the eligible population, comparison, timing, measurement or interpretation. A pilot review of realistic records or participant pathways can reveal ambiguities before they become irreversible exclusions or post hoc analytical choices.
Decision 3: Can delivery be standardised and monitored?. This point should be resolved before recruitment, extraction or assignment begins. For principles of experimental research, the protocol should state the operational rule, who applies it, when it is applied and what source documents demonstrate compliance. The investigators should also consider how a different choice would change the eligible population, comparison, timing, measurement or interpretation. A pilot review of realistic records or participant pathways can reveal ambiguities before they become irreversible exclusions or post hoc analytical choices.
Decision 4: Can contamination and co-interventions be measured?. This point should be resolved before recruitment, extraction or assignment begins. For principles of experimental research, the protocol should state the operational rule, who applies it, when it is applied and what source documents demonstrate compliance. The investigators should also consider how a different choice would change the eligible population, comparison, timing, measurement or interpretation. A pilot review of realistic records or participant pathways can reveal ambiguities before they become irreversible exclusions or post hoc analytical choices.
Decision 5: Will the experimental setting still inform the intended practice decision?. This point should be resolved before recruitment, extraction or assignment begins. For principles of experimental research, the protocol should state the operational rule, who applies it, when it is applied and what source documents demonstrate compliance. The investigators should also consider how a different choice would change the eligible population, comparison, timing, measurement or interpretation. A pilot review of realistic records or participant pathways can reveal ambiguities before they become irreversible exclusions or post hoc analytical choices.
The approach should be chosen because its structure answers the research question, not because the available dataset happens to resemble a familiar table. The investigator should state the target population, time origin, comparison, outcome and intended measure before collecting or extracting data.
Design and participant pathway
Experimental research
Core structure: Assign one or more intervention strategies according to a protocol and compare subsequent outcomes with a concurrent or structured counterfactual.
Best suited for: Estimating effects of feasible medicines, procedures, behavioural programmes, service models and policies.
Principal threats: Predictable assignment, intervention contamination, unequal co-interventions, differential outcome ascertainment, non-adherence and attrition.
Experiments may randomise individuals, clusters, sequences or time periods. They may test superiority, non-inferiority or equivalence and can be explanatory or pragmatic. The design should be selected after defining the decision and estimand, not because a named design is fashionable. Intervention fidelity and adaptation should be documented so that the tested strategy is knowable.
Protocol specification
Decisions to prespecify
- Define the intervention and comparator in replicable detail.
- Specify the unit and method of assignment.
- Define eligibility, time zero and follow-up.
- Select primary benefit and harm outcomes.
- Plan adherence, contamination and co-intervention measurement.
- Establish safety oversight and analysis populations.
A protocol should make these decisions before outcome-dependent choices become tempting. It should also distinguish the design population from the analysis population, document exclusions after enrolment or sampling, and explain how missing or indeterminate observations will be handled. Any deviation should be dated, justified and retained in the study record.[2–4]
Measurements, timing and analysis interface
Baseline variables should be measured before assignment whenever possible. Outcome schedules should be equivalent across groups and assessors masked when feasible. Process measures can show whether the intervention was delivered, received and enacted, but they should not replace clinical or service outcomes. Fidelity thresholds and allowable adaptations require prespecification.
The analysis should reflect how participants or clusters entered the study, how observations are related, and how the effect or performance measure was defined. Statistical adjustment cannot repair a comparison that was invalid at the design stage. Conversely, a sound design can still be undermined by an analysis that ignores matching, clustering, censoring, sampling fractions or prespecified thresholds.
Worked example
Turning a programme into an experimental contrast
The intervention is specified as one counselling session before discharge, automated reminders at weeks 2 and 8 and pre-booking of a laboratory appointment. The comparator is current discharge practice without the three mandated components. The primary outcome is objectively verified glucose testing by six months. Both groups have the same record-linkage follow-up, and implementation fidelity is measured separately.
The example should be interpreted in the population and time frame created by the design. The numerical result does not by itself establish causation, transportability or clinical usefulness; those conclusions require additional assumptions and evidence.
Appraising methodological quality
Questions for critical appraisal
- Intervention: Can the strategy be reproduced?
- Comparator: Is the alternative explicit and clinically relevant?
- Assignment: Could enrolment or allocation be predicted or manipulated?
- Measurement: Are outcome schedules and definitions comparable?
- Ethics: Is uncertainty genuine and participant welfare protected?
A reporting guideline can improve transparency, but it is not a substitute for sound design. Appraisal should ask whether the reported methods could generate an unbiased, clinically meaningful answer—not merely whether every checklist item has been mentioned.[3–6]
Suggested reporting
Reporting should identify the design in the title or abstract when appropriate, describe the source and flow of participants, define all key variables, report missing and indeterminate observations, and present estimates with uncertainty. The direction and denominator of every comparison must be explicit.
Ethical and governance considerations
Explanatory and pragmatic intentionsMMed extension
Explanatory trials ask whether an intervention can work under selected conditions; pragmatic trials ask how strategies perform in usual care. These are dimensions rather than a binary classification. Eligibility, flexibility, expertise, follow-up and analysis should be calibrated to the intended decision.
Common mistakes
Knowledge check
What distinguishes an experiment from an observational study?
Answer and reasoning
The study protocol assigns the intervention strategy rather than only observing naturally chosen exposure.
Why is the comparator part of the intervention question?
Answer and reasoning
An effect is always relative to an alternative; different versions of usual care can produce different contrasts.
Does randomisation guarantee identical groups?
Answer and reasoning
No. It makes assignment independent by chance and supports valid inference, but chance imbalances can occur.
Why measure fidelity?
Answer and reasoning
To determine whether the intended intervention was delivered and to interpret a null or heterogeneous effect without redefining the main outcome.
Key take-home messages
Key take-home messages
- Define a replicable intervention and comparator.
- Organise the experiment around a causal contrast and time zero.
- Use assignment, measurement and follow-up safeguards together.
- Measure adherence and fidelity without conditioning the primary causal estimate improperly.
- Make ethics and safety integral to design.
References
- International Council for Harmonisation. ICH harmonised guideline E6(R3): Guideline for Good Clinical Practice. Geneva: ICH; 2025. https://database.ich.org/sites/default/files/ICH_E6%28R3%29_Step4_FinalGuideline_2025_0106.pdf
- Hopewell S, Chan AW, Collins GS, et al. CONSORT 2025 statement: updated guideline for reporting randomised trials. BMJ. 2025;389:e081123. https://doi.org/10.1136/bmj-2024-081123
- Butcher NJ, Monsour A, Mew EJ, et al. SPIRIT 2025 statement: updated guideline for protocols of randomised trials. BMJ. 2025;389:e081477. https://doi.org/10.1136/bmj-2024-081477
- Friedman LM, Furberg CD, DeMets DL, Reboussin DM, Granger CB. Fundamentals of clinical trials. 5th ed. Cham: Springer; 2015.
- Hoffmann TC, Glasziou PP, Boutron I, et al. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide. BMJ. 2014;348:g1687. https://doi.org/10.1136/bmj.g1687
- International Council for Harmonisation. ICH E9(R1) addendum on estimands and sensitivity analysis in clinical trials. Geneva: ICH; 2019. https://database.ich.org/sites/default/files/E9-R1_Step4_Guideline_2019_1203.pdf
These materials support learning and research planning. They require adaptation to the study context and do not replace methodological, statistical, ethical, or regulatory consultation.