Explanatory vs. Pragmatic Trial: What’s the Difference?

Are You Testing an Intervention Under Carefully Controlled Conditions or Evaluating It as It Would Be Used in Practice?

Suppose you develop an AI-assisted learning program intended to improve university students' research writing. You have two possible approaches to evaluating it.

In the first, you recruit students meeting narrowly defined eligibility criteria. Trained facilitators deliver the program according to a detailed protocol, participants receive extensive implementation support, and researchers monitor every instructional activity.

In the second, several universities introduce the program into their existing courses. Ordinary instructors deliver it, students use it alongside other academic resources, and institutions retain their usual instructional arrangements.

Both studies could be randomized trials. Both could produce credible evidence. Yet they may answer different questions about the same intervention.

This distinction lies at the heart of explanatory and pragmatic trials. Understanding it requires examining what each trial is designed to establish, not merely counting how many controls the researchers impose.

Explanatory Trials Emphasize Efficacy; Pragmatic Trials Emphasize Decisions in Usual Practice

Understanding the Explanatory–Pragmatic Continuum

Where Did the Distinction Come From?

The explanatory–pragmatic distinction is commonly associated with Daniel Schwartz and Joseph Lellouch, whose influential work distinguished trials designed to clarify treatment effects from trials intended to support practical therapeutic decisions.

The distinction concerns the purpose of experimentation and its implications for design.

An explanatory attitude emphasizes understanding whether an intervention can produce a particular effect under specified conditions. A pragmatic attitude emphasizes choosing between treatment strategies in circumstances relevant to actual decisions.

The original distinction has subsequently informed clinical trial methodology, health services research, and evaluations of complex interventions.

Although the terminology originated in therapeutic research, its underlying reasoning can also inform educational and organizational intervention studies.

What Is an Explanatory Trial?

An explanatory trial emphasizes estimating the effect of an intervention under conditions selected to clarify whether it can produce the intended outcome.

Researchers may use relatively restrictive eligibility criteria, specialized intervention providers, carefully specified treatment procedures, and close implementation monitoring.

For example, investigators might evaluate a digital feedback intervention among students with similar prior preparation, using specially trained instructors and a fixed instructional schedule.

Such choices may help clarify the intervention's efficacy under the selected conditions.

However, explanatory trials do not necessarily establish an intervention's mechanism merely because they use controlled procedures. Demonstrating a mechanism requires appropriate theoretical, measurement, and causal evidence.

Likewise, an explanatory trial can still produce findings relevant to practice, although the conditions under which those findings apply require consideration.

What Is a Pragmatic Trial?

A pragmatic trial emphasizes estimating the effects of treatment strategies under conditions relevant to routine decisions.

It may include a broad range of eligible participants, ordinary practitioners, existing service arrangements, flexible delivery, and outcomes meaningful to patients, students, institutions, or policymakers.

For example, a pragmatic educational trial might evaluate whether universities should adopt an AI-assisted writing program as part of ordinary postgraduate instruction.

Participating institutions could retain their usual academic support services while implementing the intervention within specified boundaries.

The resulting estimate may inform decisions about introducing the program under conditions resembling those studied.

However, a pragmatic trial is not automatically representative of every real-world setting. Its applicability depends on the participants, institutions, implementation arrangements, and contexts included.

What Are the Main Differences Between Explanatory and Pragmatic Trials?

These are tendencies rather than mandatory rules. A trial may use broad eligibility but tightly standardized intervention delivery, or employ ordinary practitioners while collecting specialized outcome measurements.

Are Explanatory and Pragmatic Trials Two Separate Research Designs?

Not in the sense of mutually exclusive categories with fixed methodological requirements.

The explanatory–pragmatic distinction describes the orientation of trial design choices toward different research purposes.

A randomized parallel-group trial can be designed with explanatory or pragmatic characteristics. The same applies to some cluster-randomized and other experimental arrangements.

Researchers should therefore distinguish the trial's allocation structure from its explanatory or pragmatic orientation.

Calling a study pragmatic does not specify how participants were randomized, how outcomes were measured, or how the treatment effect was estimated.

Those methodological details must still be reported explicitly.

What Is the Explanatory–Pragmatic Continuum?

The continuum recognizes that trial characteristics can vary in their alignment with explanatory and pragmatic objectives.

A trial need not occupy one position across every design dimension.

For example, investigators may recruit a broad population and use ordinary instructors but require a highly structured intervention protocol.

Such a trial may be relatively pragmatic in eligibility and setting while more explanatory in delivery flexibility.

It is therefore more informative to describe specific design choices than to rely exclusively on a single trial label.

The continuum also discourages the assumption that pragmatic trials are necessarily better than explanatory trials. Their value depends on whether the design answers the intended question.

What Is PRECIS-2?

PRECIS-2, the Pragmatic Explanatory Continuum Indicator Summary 2, is a tool developed to help trial designers assess how well design choices align with the trial's intended purpose.

It considers nine domains: eligibility, recruitment, setting, organization, flexibility of delivery, flexibility of adherence, follow-up, primary outcome, and primary analysis.

Each domain is scored on a five-point scale, with lower scores indicating a more explanatory approach and higher scores indicating a more pragmatic approach.

The scores are commonly represented using a wheel to reveal the profile of the trial across domains.

PRECIS-2 is a design aid, not a universal measure of trial quality or an instrument that proves external validity.

A high score is appropriate only when greater pragmatism serves the intended research question.

Does a Pragmatic Trial Have to Be Randomized?

The term pragmatic trial commonly refers to a randomized trial designed to inform practical decisions, although pragmatic characteristics can also be discussed in other evaluation designs.

Randomization is fully compatible with ordinary practice conditions.

For example, universities may be randomized to offer a new learning program while retaining their existing instructional arrangements.

Randomization can support causal inference without requiring that every aspect of intervention delivery be tightly controlled.

Researchers should not confuse pragmatic implementation with observational allocation.

Can Pragmatic Trials Be Blinded?

Yes. Blinding and pragmatism address different methodological concerns.

Blinding can reduce certain biases when knowledge of treatment assignment influences behavior, outcome assessment, or analysis.

However, participant or provider blinding may be impractical for interventions that visibly change ordinary practice.

For example, students and instructors may know whether a digital learning platform has been introduced.

Outcome assessors or analysts may nevertheless remain blinded where feasible.

Researchers should use appropriate safeguards against bias rather than assume that pragmatic trials must be unblinded.

How Do Eligibility Criteria Differ?

Explanatory trials may use narrower eligibility criteria to reduce variation or focus on participants expected to respond under particular conditions.

Pragmatic trials may use broader criteria when the target decision concerns a diverse population encountered in routine practice.

For example, a university-wide adoption decision may require evidence from students with different prior preparation, academic disciplines, and support needs.

However, broad eligibility is not automatically preferable.

Eligibility criteria must still reflect intervention safety, feasibility, the target population, and the research question.

Researchers should consider how broad eligibility criteria should be in pragmatic trials.

How Does Intervention Delivery Differ?

Explanatory trials may require providers to follow detailed procedures to establish a particular treatment contrast.

Pragmatic trials may permit more variation when provider discretion is part of routine implementation.

For example, an explanatory educational trial might require every instructor to deliver identical feedback sessions.

A pragmatic trial might require the same essential feedback functions while allowing instructors to adapt examples and scheduling.

However, flexible delivery should not make the intervention impossible to identify.

The appropriate level of intervention standardization in pragmatic trials depends on which components define the treatment strategy.

What Happens to Usual Care?

Usual care often serves as a comparator in pragmatic trials because the practical decision concerns whether a new intervention should replace or supplement existing services.

However, usual care may vary across providers, institutions, and participants.

For example, universities may differ in tutoring availability, instructional technology, and academic advising.

These differences may be part of the environment in which the intervention is intended to operate.

Researchers should describe usual care sufficiently and consider whether variation across sites should be permitted.

Usual care should not be treated as a self-explanatory comparison condition.

How Do Outcome Choices Differ?

Explanatory trials may prioritize outcomes that are sensitive to a proposed mechanism or useful for understanding intervention effects under selected conditions.

Pragmatic trials often emphasize outcomes that matter directly to practical decisions.

For example, an educational intervention might improve performance on a narrowly defined feedback-recognition task.

That result may be valuable for understanding a learning process, but universities deciding whether to adopt the program may also need evidence about independent writing quality, course completion, or other consequential outcomes.

Researchers should consider the importance of meaningful outcomes in pragmatic trials.

The outcome should correspond to the research question, not merely be convenient to collect.

Do Explanatory and Pragmatic Trials Use Different Statistical Methods?

Not necessarily. Both require analyses appropriate to their randomization design, outcomes, and estimands.

However, their intended effects may differ.

A pragmatic trial may emphasize the effect of assignment to a treatment strategy under ordinary adherence and implementation conditions.

An explanatory trial may seek an effect under more tightly specified treatment receipt or adherence conditions.

These questions cannot always be answered by simply choosing between intention-to-treat and per-protocol analyses.

The ICH E9(R1) framework emphasizes defining the estimand before selecting analytical methods.

Post-randomization adjustment for adherence or treatment receipt may require additional causal assumptions.

Does Pragmatism Mean Lower Internal Validity?

No. A pragmatic trial can have strong internal validity when allocation, outcome measurement, missing data, and analysis are handled appropriately.

Allowing ordinary variation in treatment delivery does not automatically introduce bias for the intended treatment-policy effect.

However, pragmatic conditions may create particular challenges, including differential recruitment, missing outcomes, contamination, and inconsistent measurement.

These challenges require methodological safeguards.

Researchers should distinguish flexibility that belongs to the treatment strategy from deviations that compromise the intended comparison.

Does an Explanatory Trial Have Poor External Validity?

Not automatically. Explanatory trials may provide useful evidence for populations and settings resembling those studied.

However, restrictive eligibility, specialized providers, or unusually intensive implementation support may limit applicability to routine conditions.

For example, an intervention that succeeds with expert facilitators may not produce the same effects when delivered by ordinary instructors without equivalent support.

Researchers should distinguish credible estimation under selected conditions from justified generalization to other conditions.

External validity depends on the relationship between the study and target settings, not simply the explanatory label.

Can Pragmatic Trials Make Mechanisms Harder to Explain?

Potentially. When interventions vary across providers and settings, the observed effect may reflect several interacting processes.

For example, a digital learning intervention may improve achievement through feedback, increased practice, or changes in instructor behavior.

A pragmatic trial may estimate the overall effect of introducing the program without identifying the contribution of each mechanism.

Process evaluation and mechanism-focused studies can help investigate these questions.

Researchers should consider whether greater pragmatism complicates explanations of intervention effects.

Can Explanatory and Pragmatic Evidence Complement Each Other?

Yes. Different trials may answer complementary questions about an intervention.

An explanatory study may establish whether a particular treatment strategy produces an effect under selected conditions.

A pragmatic study may investigate whether introducing that strategy produces useful outcomes under ordinary implementation.

However, there is no universal requirement that an explanatory trial must precede a pragmatic trial.

The appropriate sequence depends on existing evidence, intervention development, feasibility, ethical considerations, and the decisions requiring support.

Researchers should select designs according to unresolved questions rather than assume a fixed progression from efficacy to effectiveness.

How Should Researchers Choose Their Trial Orientation?

The starting point should be the decision the research is intended to inform.

If the central question concerns whether an intervention can produce a hypothesized effect under specified delivery conditions, more explanatory design choices may be appropriate.

If the central question concerns whether institutions should introduce the intervention under ordinary conditions, more pragmatic choices may be justified.

Researchers should then examine eligibility, recruitment, setting, delivery, adherence, follow-up, outcomes, and analysis individually.

PRECIS-2 can help identify whether these choices are coherent with the intended purpose.

The goal is not to maximize pragmatism or control. It is to design a trial capable of answering the question that matters.

Two Trials of the Same AI-Assisted Research Writing Program

Suppose the explanatory trial finds a meaningful improvement, while the pragmatic trial finds a smaller effect.

This difference might reflect variation in providers, participant populations, implementation support, treatment receipt, or other contextual conditions.

However, comparing results from two separately conducted trials does not automatically identify which factor caused the difference.

The example demonstrates that explanatory and pragmatic trials can produce different but complementary evidence without one being inherently more scientifically rigorous than the other.

Common Misconceptions About Explanatory and Pragmatic Trials

Choose the Trial Orientation That Fits the Decision You Want to Inform

Begin by specifying whether your primary objective concerns efficacy under selected conditions or effectiveness under circumstances relevant to routine practice.

Then align individual design decisions with that objective rather than selecting a trial label first and forcing the methods to fit it.

Researchers should also consider how to position a trial on the pragmatic–explanatory continuum rather than relying on a single descriptive label.

A defensible trial is one whose design, estimand, and analysis correspond to the scientific or practical question it is intended to answer.

Before Classifying a Trial as Explanatory or Pragmatic

Frequently Asked Questions About Explanatory and Pragmatic Trials

Explanatory and Pragmatic Trials Answer Different Questions, Not Different Standards of Scientific Quality

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.