How Do You Position a Trial on the Pragmatic-Explanatory Continuum?

How Can You Determine Whether Your Trial Is Explanatory, Pragmatic, or Somewhere Between?

You are designing a randomized trial of an AI-assisted research writing program across several universities. The study includes a broad range of postgraduate students, ordinary instructors, and existing academic support arrangements.

However, the intervention also requires a standardized set of core activities, trained outcome assessors, and carefully specified analytical procedures.

Some features resemble a pragmatic trial, while others appear more explanatory.

Should the study be labeled pragmatic because it takes place in ordinary universities? Does requiring a standardized intervention make it explanatory? Can the same trial legitimately have characteristics of both?

These questions cannot be answered reliably through a single label. They require examining how individual design choices align with the scientific or practical question the trial is intended to answer.

Evaluate the Trial Across Multiple Design Domains Rather Than Assigning a Binary Label

Using the Pragmatic-Explanatory Continuum to Design and Describe Trials

What Is the Pragmatic-Explanatory Continuum?

The pragmatic-explanatory continuum describes how trial design choices may align with different research purposes.

An explanatory orientation emphasizes understanding intervention effects under selected conditions.

A pragmatic orientation emphasizes informing decisions about treatment strategies under conditions relevant to ordinary practice.

However, most trials do not fit neatly into one category.

They may combine broad eligibility with tightly specified intervention delivery, or use ordinary practitioners while collecting specialized outcome measures.

The continuum recognizes this multidimensional variation.

It is therefore more informative to examine specific design domains than to describe a trial as simply pragmatic or explanatory.

Why Is the Continuum More Useful Than a Binary Classification?

Binary labels can conceal important design differences.

For example, two trials may both describe themselves as pragmatic even though one uses broad eligibility and ordinary delivery while the other restricts participation and requires intensive specialist supervision.

Both may include features relevant to practical decisions, but their design profiles differ.

The continuum encourages researchers to specify which choices support their intended purpose and which may require reconsideration.

It also prevents the mistaken assumption that a trial must be entirely explanatory or entirely pragmatic.

What Is PRECIS-2?

PRECIS-2 stands for Pragmatic Explanatory Continuum Indicator Summary 2.

It was developed to help trial designers assess whether their planned methods are consistent with the trial's intended purpose.

The framework evaluates nine design domains using a five-point rating scale.

The ratings are commonly displayed on a wheel that makes the trial's design profile visible.

PRECIS-2 is intended primarily as a trial design aid, although it has also been used to characterize completed trials.

It is not a risk-of-bias instrument, a measure of methodological quality, or a direct test of generalizability.

What Are the Nine PRECIS-2 Domains?

The domains should be considered together, but each should receive its own justification.

How Does the PRECIS-2 Scoring Scale Work?

Each domain is rated on a scale from 1 to 5.

A score of 1 represents a very explanatory design choice, while a score of 5 represents a very pragmatic choice.

Intermediate scores indicate positions between these orientations.

Ratings require judgment based on the actual trial design and intended practice conditions.

Researchers should not assign scores merely according to whether a particular procedure sounds controlled or flexible.

For example, a standardized intervention may receive a pragmatic rating when identical delivery is how the intervention would ordinarily be implemented.

Should Researchers Calculate an Overall PRECIS-2 Score?

The PRECIS-2 wheel is intended to show the trial's profile across domains rather than reduce its design to one number.

Although researchers may calculate descriptive summaries, a total or average score can conceal important differences.

For example, a trial may have highly pragmatic eligibility but a very explanatory primary outcome.

An average could obscure the fact that the selected outcome does not directly inform the intended practical decision.

Researchers should prioritize domain-level interpretation and the coherence of the overall design.

There is no universal PRECIS-2 cutoff that establishes whether a trial qualifies as pragmatic.

How Should the Eligibility Domain Be Rated?

The eligibility domain concerns whether participants resemble those who would ordinarily receive the intervention.

For example, a university-wide learning program may be intended for students with different academic backgrounds and levels of digital literacy.

A trial excluding students with low prior achievement or limited technological experience may be more explanatory in eligibility.

However, restrictions justified by safety or the intervention's intended population are not automatically inappropriate.

Researchers should consider how broad eligibility criteria should be in pragmatic trials in relation to the target population.

How Is Recruitment Different From Eligibility?

Eligibility determines who may participate. Recruitment concerns how eligible individuals are identified, approached, and enrolled.

A trial may allow a broad population but recruit only highly motivated volunteers through specialized research advertisements.

Such recruitment may differ from the pathways through which individuals would ordinarily receive the intervention.

For example, a university-wide program might normally be offered through existing courses rather than through a voluntary research-interest mailing list.

Researchers should therefore rate eligibility and recruitment separately.

What Does the Setting Domain Assess?

The setting domain concerns how closely trial locations resemble those in which the intervention would ordinarily be used.

For example, an educational intervention evaluated in ordinary university classrooms may be more pragmatic in setting than one evaluated exclusively in a specialized research laboratory.

However, the label university does not automatically establish a pragmatic setting.

Participating institutions may have unusually advanced resources or specialized personnel.

Researchers should compare the actual study settings with the environments relevant to the intended decision.

What Does the Organization Domain Assess?

The organization domain considers the resources, expertise, infrastructure, and organizational arrangements required to deliver the intervention.

For example, a program requiring specialist facilitators and intensive research-team supervision may be more explanatory when ordinary institutions would not have those resources.

However, specialized delivery may be entirely consistent with usual practice for some interventions.

Researchers should determine whether the required organizational arrangements are part of the intended treatment strategy.

The domain should not be rated solely according to the absolute cost or complexity of the intervention.

What Does Flexibility of Delivery Assess?

This domain concerns the extent to which providers can deliver the intervention according to ordinary professional or institutional practice.

A trial may permit instructors to adapt examples, schedules, or delivery formats while preserving essential intervention functions.

Alternatively, it may require identical procedures beyond what ordinary implementation would demand.

The rating depends on the intended practice environment.

Researchers should consider how much intervention standardization is compatible with a pragmatic trial rather than assume that unrestricted flexibility is desirable.

What Does Flexibility of Adherence Assess?

The flexibility-of-adherence domain concerns the extent to which the trial uses procedures to encourage participant adherence beyond those expected in usual practice.

For example, frequent research-team reminders, additional coaching, or incentives designed specifically to increase treatment receipt may make the trial more explanatory in this domain.

However, adherence support may be part of the intended intervention package.

Researchers should distinguish ordinary treatment support from additional procedures introduced solely for research purposes.

The rating should reflect the intended treatment strategy and practice conditions.

What Does the Follow-Up Domain Assess?

The follow-up domain considers how closely outcome follow-up resembles the procedures that would occur in usual practice.

A trial requiring frequent specialized assessments may be more explanatory when such procedures are not ordinarily used.

However, the scientific need for valid outcome measurement remains important.

Researchers should not sacrifice essential follow-up or create excessive missing data merely to obtain a more pragmatic rating.

The relevant question is whether follow-up procedures impose additional demands beyond those appropriate for the trial's purpose.

What Does the Primary Outcome Domain Assess?

The primary outcome domain considers whether the selected outcome is relevant to the people and organizations making or experiencing the treatment decision.

For example, an educational intervention may be evaluated using a short knowledge quiz or an independently completed research proposal.

If universities are deciding whether to adopt the intervention to improve research writing competence, proposal quality may be more directly relevant.

However, outcome validity and feasibility also matter.

Researchers should consider the selection of meaningful outcomes in pragmatic trials rather than choose measures solely for convenience.

What Does the Primary Analysis Domain Assess?

The primary analysis domain concerns whether the analytical approach estimates an effect relevant to the intended practical decision.

Pragmatic trials often emphasize effects of assignment to treatment strategies under ordinary adherence and implementation conditions.

However, the appropriate analysis depends on the estimand and randomization design.

An intention-to-treat approach may support a pragmatic question, but it does not automatically make the entire trial pragmatic.

Likewise, a per-protocol analysis does not automatically establish an explanatory effect.

Researchers should define the target effect and choose methods appropriate to its causal assumptions.

Can a Trial Be Highly Pragmatic in Some Domains and Explanatory in Others?

Yes. This is one of the central reasons to use a multidimensional framework.

For example, a trial may recruit students through ordinary courses, use existing university facilities, and permit flexible instructional delivery.

At the same time, it may use specialist outcome assessments and intensive adherence monitoring.

These features produce a mixed profile.

Researchers should examine whether the differences are justified by the intended research question.

A mixed profile is not automatically a design weakness.

How Should Researchers Assign Ratings?

Ratings should be based on a clear description of each design domain and the intended usual practice comparator.

Researchers may conduct a structured discussion involving trial designers, implementation specialists, practitioners, and other relevant stakeholders.

Disagreements can reveal uncertainty about what ordinary practice actually involves.

For example, investigators may believe that weekly specialist supervision is routine, while participating institutions report that such support is unavailable.

Resolving the disagreement may require additional information about the target setting.

Ratings should be justified rather than treated as purely mechanical calculations.

Should PRECIS-2 Be Applied Before or After Trial Design?

Its principal value lies in prospective trial design.

Researchers can use the framework to identify choices that do not align with the intended explanatory or pragmatic objective.

For example, a trial intended to inform university-wide adoption may discover that its eligibility criteria exclude many ordinary recipients.

The design can then be reconsidered before recruitment begins.

PRECIS-2 can also be used descriptively after a trial has been conducted, but retrospective ratings do not correct design limitations.

Can PRECIS-2 Scores Be Compared Across Trials?

Comparisons may be informative when ratings are applied consistently and interpreted within the relevant research contexts.

However, the meaning of a score depends on the intervention and what constitutes usual practice.

For example, highly standardized delivery may be routine for one treatment but unusual for another.

Researchers should therefore avoid interpreting numerical differences as direct measures of scientific quality or generalizability.

Comparisons should consider domain-level justifications rather than rely exclusively on totals.

Does a More Pragmatic Profile Guarantee External Validity?

No. A trial may resemble usual practice in participating settings while remaining poorly representative of other target institutions.

For example, a pragmatic educational trial conducted exclusively in well-resourced universities may not provide sufficient evidence for institutions with limited connectivity.

PRECIS-2 helps assess alignment with pragmatic design intentions, but it does not directly establish transportability.

Researchers should separately evaluate whether participants, settings, and treatment strategies correspond to the population and decision of interest.

Does a More Explanatory Profile Mean Lower Scientific Quality?

No. Explanatory design choices may be appropriate when researchers seek to establish efficacy under specified conditions or investigate a narrowly defined treatment strategy.

For example, specialist delivery and carefully controlled procedures may be necessary to evaluate a complex intervention during early development.

The quality of the trial depends on whether its design and analysis support the intended inference.

PRECIS-2 should not be used as a ranking system in which the highest pragmatic score automatically wins.

How Does PRECIS-2 Relate to Risk-of-Bias Assessment?

PRECIS-2 and risk-of-bias tools address different questions.

PRECIS-2 evaluates how trial design choices align with explanatory or pragmatic purposes.

Risk-of-bias tools such as RoB 2 evaluate whether particular features of a randomized trial may bias its effect estimate.

A trial may be highly pragmatic but have important bias concerns.

Another may be highly explanatory and have strong internal validity.

Researchers should not substitute PRECIS-2 ratings for methodological quality assessment.

How Should Researchers Use the PRECIS-2 Wheel?

The PRECIS-2 wheel visually represents the nine domain ratings.

Ratings closer to the center indicate more explanatory choices, while ratings toward the outer edge indicate more pragmatic choices.

The shape helps reveal differences between domains.

For example, a trial may have high ratings for eligibility and setting but lower ratings for adherence flexibility and follow-up.

Rather than treating this pattern as a defect, researchers should examine whether those choices correspond to the intended question.

The wheel is most useful as a structured discussion tool rather than a decorative label of trial quality.

How Should the Trial's Position Be Reported?

Researchers should describe the intended explanatory or pragmatic objective and the design choices supporting it.

When PRECIS-2 is used, domain ratings and their justifications should be reported sufficiently for readers to understand the assessment.

A wheel may provide a useful summary, but it should not replace detailed descriptions of eligibility, recruitment, intervention delivery, outcomes, and analysis.

Researchers should avoid claiming that a trial is fully pragmatic based on one or two design features.

The most informative reporting explains how the overall design corresponds to the decision the trial is intended to support.

Applying PRECIS-2 to an AI-Assisted Research Writing Trial

These ratings are illustrative judgments rather than empirically validated scores for an actual trial. Formal PRECIS-2 assessment would require consideration of the full domain guidance and agreement about the intended usual practice conditions.

The study is relatively pragmatic in eligibility, setting, primary outcome, and analysis, while retaining more explanatory features in other domains.

This mixed profile does not make the trial inconsistent or invalid.

Instead, it identifies specific design decisions that require justification and provides a structured basis for refining the trial before implementation.

Common Misconceptions About the Pragmatic-Explanatory Continuum

Use the Continuum to Improve Trial Design, Not Merely to Label It

Begin with the research question and the decision the trial is intended to inform.

Then examine each PRECIS-2 domain and determine whether its design choices align with that purpose.

Researchers should also consider whether the study primarily addresses efficacy or effectiveness and how that objective should shape the treatment strategy and estimand.

The purpose of positioning a trial on the continuum is to make its design choices more coherent and transparent, not to achieve a predetermined pragmatic score.

Before Positioning a Trial on the Pragmatic-Explanatory Continuum

Frequently Asked Questions About the Pragmatic-Explanatory Continuum

A Trial's Position on the Continuum Is a Design Profile, Not a Quality Ranking

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.