No-Treatment vs. Usual-Care vs. Active-Control vs. Placebo Groups: Which Should You Use?

What Are You Actually Comparing Your Intervention Against?

Suppose you want to evaluate a new educational intervention, psychological treatment, or clinical procedure. You have identified the intervention, selected your outcomes, and decided that a comparison group is necessary. But what should that group receive?

Should participants receive nothing, continue their usual care, receive an established intervention, or undergo a placebo procedure?

These choices are not interchangeable. A new intervention may outperform no treatment yet offer little improvement over an established alternative. It might also appear effective against a placebo while proving impractical in routine service delivery.

The comparator determines the contrast your study estimates. Choosing it therefore requires more than finding a convenient group against which to measure change.

Choose the Comparator That Represents Your Research Question

How the Four Comparator Types Change the Meaning of Your Findings

Why Comparator Selection Is a Scientific Decision

In an intervention study, the treatment effect is defined relative to a particular alternative. Researchers sometimes describe this as the estimand's treatment contrast: the difference between outcomes under specified intervention conditions, for a defined population and outcome.

A statement such as "the intervention improved performance" is incomplete unless the reader understands what would otherwise have happened. Was improvement greater than among participants receiving no additional support? Greater than under routine instruction? Greater than with an established educational program?

Random assignment can improve the comparability of groups, but it does not make different comparators scientifically equivalent. The intervention contrast remains determined by the study conditions.

What Is a No-Treatment Control?

A no-treatment control group does not receive the intervention being evaluated or another intervention specifically designated as its substitute during the study period.

This does not necessarily mean participants receive absolutely no services. Researchers must specify whether ordinary background services, independent treatment-seeking, or other support remain available.

No-treatment controls may help establish whether introducing an intervention produces better outcomes than not introducing it. They are especially relevant when no established effective intervention exists, treatment is genuinely optional, and withholding the study intervention does not create unacceptable risk.

However, differences between groups may reflect more than the intervention's specific ingredients. Additional attention, participant expectations, contact with researchers, and changes in behavior associated with receiving an intervention can contribute to observed differences.

Where an effective intervention is already available, withholding it solely to create a no-treatment condition may be scientifically unnecessary or ethically unacceptable.

What Is a Usual-Care Control?

A usual-care control group continues to receive the services, treatment, or educational provision that participants would ordinarily receive outside the experimental intervention.

In clinical research, this might involve routine management according to local practice. In education, it could mean the existing curriculum, instructional approach, or student-support arrangements.

Usual care is particularly useful when the research question concerns the added value of introducing a new intervention into an existing system.

For example, an institution might evaluate whether an AI-supported formative feedback system improves learning outcomes compared with its existing feedback practices.

The difficulty is that usual care may vary across institutions, practitioners, and participants. One school might provide weekly individualized feedback, while another relies on occasional whole-class comments. Calling both conditions "usual care" without documenting their content can obscure the interpretation.

Researchers should therefore describe the services available, their intensity, the personnel delivering them, and any meaningful variation between sites.

What Is an Active-Control Group?

An active-control group receives another intervention intended to produce a relevant effect, rather than no intervention or an inert substitute.

The comparator may be an established treatment, a competing program, or another credible intervention selected for a particular scientific purpose.

An active-control design is useful when researchers want to know whether a new intervention is better than an existing alternative, similarly effective, or associated with different benefits and burdens.

Importantly, the presence of an active control does not automatically make a study a noninferiority or equivalence trial. Those objectives require explicitly defined hypotheses, justified margins, appropriate sample-size calculations, and suitable analysis.

Active controls also vary in purpose. An established effective treatment serves a different scientific role from an attention-matched activity designed primarily to account for contact time or participant engagement.

What Is a Placebo Control?

A placebo control is designed to resemble the experimental intervention while omitting the component or components believed to produce its specific therapeutic effect.

In a drug trial, the placebo may resemble the investigational medication without containing its active ingredient. In procedural or behavioral research, constructing a credible placebo can be substantially more difficult.

Placebo controls can help reduce differences in expectations and support blinding. They may improve the ability to distinguish an intervention-specific effect from changes associated with the treatment context.

Nevertheless, a placebo comparison does not automatically isolate every mechanism responsible for improvement. Placebo responses, spontaneous recovery, regression to the mean, reporting behavior, and other influences require careful interpretation.

Placebo use is also subject to ethical constraints, especially when participants would otherwise receive an established effective intervention.

How Do These Four Comparators Differ?

Can Usual Care and Active Control Be the Same Thing?

Yes. These terms describe different characteristics of a comparator.

"Usual care" describes what participants ordinarily receive. "Active control" indicates that the comparator includes an intervention expected to have an effect.

When routine practice consists of an established effective treatment, the same comparison condition can reasonably be described as both usual care and an active control.

Conversely, usual care may involve monitoring or supportive services without a specific active treatment. The protocol should explain the actual condition rather than relying on labels alone.

Why the Comparator Affects the Size and Interpretation of an Effect

Imagine that a new learning intervention produces an average posttest score of 82. If the no-treatment group averages 70, the observed difference is 12 points. If a well-established alternative produces an average of 80, the difference is only 2 points.

These results are not contradictory. They represent different comparisons.

Effect estimates also depend on participant characteristics, implementation fidelity, outcome measurement, adherence, and study context. It would be incorrect to assume that changing the comparator necessarily produces a predictable change in effect size.

In noninferiority trials, comparator selection is particularly consequential. Researchers need evidence that the active comparator has an established effect under sufficiently comparable conditions. This relates to assay sensitivity, or the study's ability to distinguish an effective intervention from a less effective or ineffective one.

What Ethical Limits Apply to Comparator Selection?

Scientific usefulness does not by itself justify withholding effective treatment.

The World Medical Association's Declaration of Helsinki provides important guidance on comparing interventions with the best proven intervention and on circumstances in which placebo or no intervention may be ethically permissible.

In particular, researchers must consider whether participants assigned to a less effective or inactive condition would face additional risks of serious or irreversible harm because they are not receiving the best proven intervention.

The appropriate ethical assessment depends on the population, condition, available evidence, research setting, and applicable oversight requirements.

A trial may also provide the experimental intervention in addition to existing care. Such an add-on design asks whether the new component offers benefits beyond the services participants already receive, rather than whether it should replace those services.

Researchers should address the ethical justification during protocol development, not after selecting a convenient comparator.

Choosing a Comparator for an AI-Assisted Feedback Intervention

The final condition should not automatically be called a placebo. Generic feedback may itself influence learning, making it an active or component-control condition rather than a genuinely inert substitute.

Suppose the university's actual decision is whether to adopt the AI system instead of its current feedback arrangements. A usual-care comparison would likely provide more directly applicable evidence than a no-treatment comparison.

Researchers would also need to consider whether the AI system supplements or replaces faculty feedback. These are different intervention strategies and should not be treated as the same causal contrast.

Common Mistakes When Selecting Control Conditions

Match the Comparator to the Decision Your Research Must Inform

Begin by specifying what decision the evidence is intended to support. If your institution is considering replacing an established service, comparing the new intervention with no service may be insufficient. If you are investigating whether an experimental component has a specific effect, a credible placebo or component-control condition may be more informative, provided it is feasible and ethical.

Where multiple questions are important, additional comparator arms may be useful, but they increase recruitment requirements, analysis complexity, and resource demands. The decision to add a second comparison group should follow a clearly justified scientific need.

Before finalizing the protocol, establish what makes the selected comparison appropriate for the target population and intended inference. Also determine whether differences between conditions could arise from unequal contact time, intervention intensity, or access to other services.

If the ideal comparator would require withholding beneficial care, evaluate ethically acceptable alternative comparisons rather than assuming that scientific control takes precedence over participant welfare.

Before Finalizing Your Comparator, Check These Decisions

Questions About Choosing Between Control Groups

The Right Comparator Is the One That Makes Your Conclusion Meaningful

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.