Can Baseline Measurement Itself Change Participant Behavior?

Could Measuring Participants Before an Intervention Change What They Do?

You are preparing an intervention study and plan to assess participants before the intervention begins. Perhaps you will ask students about their study habits, test their knowledge, or measure their motivation. The assessment seems necessary because you want to establish their starting conditions.

But what if completing the assessment changes those conditions?

A student who answers questions about procrastination may become more conscious of missed deadlines. Someone who completes a knowledge test may recognize gaps in understanding and begin studying independently. Participants might change their behavior before receiving the intervention, simply because the research has drawn their attention to it.

This raises a methodological question: can baseline measurement influence the very outcome researchers intend to study?

Yes, Baseline Assessment Can Influence Subsequent Behavior

How Baseline Assessment Can Influence Research Outcomes

Measurement Is Not Always a Passive Observation

Researchers often treat baseline measurement as a procedure that records an existing condition without changing it. That assumption is reasonable for some measurements but less secure for assessments involving knowledge, beliefs, intentions, and behavior.

Answering questions can require participants to reflect on experiences they might otherwise ignore. Completing a performance test can provide practice. Being observed may influence how individuals behave, particularly when they recognize which behavior is being evaluated.

These possibilities matter because baseline measurement is part of the research procedure. If it changes participants' subsequent responses, the observed outcome may reflect both the intervention and the measurement process.

This does not mean that every pretest is problematic. Rather, researchers should consider whether the assessment could plausibly influence the outcome and whether the study design can accommodate that influence.

What Is Measurement Reactivity?

Measurement reactivity refers to changes in participants' responses or behavior associated with the act of measurement or observation.

For example, asking participants to record their daily screen time may make them more aware of device use. That awareness could lead some participants to reduce their screen time, even without receiving a separate behavior-change intervention.

Similarly, asking students to report how frequently they revise course materials may prompt reflection on their study routines.

Measurement reactivity can arise through several mechanisms, but researchers should not assume that any observed behavioral change was caused by the assessment. Natural change, concurrent experiences, and other study procedures may offer alternative explanations.

How Is the Question-Behavior Effect Different?

The question-behavior effect describes situations in which asking people questions about a behavior influences their subsequent performance of that behavior.

Questions about intentions, expectations, or past behavior may make particular actions more cognitively accessible or encourage participants to evaluate whether their behavior is consistent with their stated intentions.

For example, asking students whether they intend to complete weekly practice exercises could potentially influence their subsequent engagement with those exercises.

Research syntheses have reported evidence of question-behavior effects, although effect estimates and their generalizability vary with study characteristics, measurement procedures, and analytical decisions.

Researchers should therefore treat the effect as a plausible design consideration rather than assume that asking about a behavior will reliably increase or decrease it.

What Is Pretest Sensitization?

Pretest sensitization occurs when exposure to a pretest changes how participants respond to a subsequent intervention or assessment.

Suppose students complete a pretest containing questions about specific statistical concepts. The questions may reveal which topics deserve attention, encouraging students to study those concepts before or during the intervention.

The pretest might also make students more receptive to an instructional program that addresses the tested material.

In this situation, baseline assessment does more than influence behavior independently. It may modify the effect of the intervention itself.

This distinction is important because an intervention evaluated among pretested participants may not produce the same effect among individuals who receive the intervention without that preliminary assessment.

Distinguish Measurement Reactivity, Testing Effects, and Pretest Sensitization

These concepts overlap, and the terminology is not always used identically across disciplines. Their common implication is that assessment procedures may influence what researchers subsequently observe.

Can Baseline Measurement Improve Outcomes Without the Intervention?

Potentially. An assessment may encourage self-monitoring, reflection, information seeking, or behavioral adjustment.

In educational research, completing a diagnostic test might motivate students to review material. In health research, questions about physical activity may encourage participants to reconsider their activity patterns.

However, the direction of measurement effects is not necessarily beneficial. Assessments may increase anxiety, produce fatigue, discourage participants, or alter behavior in ways that depend on the context.

Researchers should avoid describing all assessment-induced changes as improvements. The relevant issue is whether the measurement changes the outcome or the response to intervention.

Does Giving Both Groups the Same Baseline Assessment Solve the Problem?

Not completely.

When intervention and control groups receive equivalent baseline assessments, measurement-related changes that affect both groups similarly may not bias their randomized comparison. This is one reason consistent assessment procedures are important.

Nevertheless, equal exposure to a pretest does not guarantee that its effects are identical across groups.

The intervention may reinforce information introduced by the pretest, while the comparison condition does not. Alternatively, the pretest might encourage independent behavior change among control participants, reducing the observed difference between groups.

Even if the randomized comparison remains internally valid, the estimated effect may describe the intervention under conditions that include pretesting. It may not represent the effect of delivering the intervention without that assessment.

This distinction concerns both causal interpretation and applicability to settings outside the study.

How Can Researchers Determine Whether Pretesting Changes Intervention Effects?

One established approach is the Solomon four-group design, which combines intervention assignment with the presence or absence of pretesting.

In its conventional form, participants are randomly assigned to four conditions:

This design permits researchers to examine intervention effects with and without pretesting and to investigate whether pretesting modifies the intervention effect.

For example, the intervention-control comparison among pretested participants can be contrasted with the corresponding comparison among participants who were not pretested.

The design has limitations. It requires additional participants or allocation groups, may increase logistical complexity, and can provide imprecise estimates of interactions unless adequately powered.

Furthermore, all groups still undergo a posttest, so the design does not eliminate every possible form of assessment reactivity.

A Solomon design is therefore useful when pretest sensitization is a substantive concern, not an obligatory feature of intervention research.

Should Researchers Eliminate Baseline Measurement to Avoid Reactivity?

Not automatically. Removing baseline assessment may reduce measurement-induced effects, but it can also eliminate valuable information.

Baseline measurements may establish eligibility, characterize participants, support adjustment for prognostic characteristics, and improve statistical precision.

In observational studies, appropriately timed baseline information may be necessary to address confounding.

Researchers should therefore evaluate the benefits of measuring the primary outcome at baseline against the potential for reactivity.

For some randomized studies, a posttest-only design may be suitable. For others, the information gained from pretesting may justify retaining it despite a plausible measurement effect.

How Can Baseline Measurement Reactivity Be Reduced?

Several design choices may help, depending on the mechanism of concern.

Researchers can avoid unnecessary questions that explicitly encourage behavioral intentions or self-evaluation when those questions are not required by the research objectives.

They may use less intrusive measurement procedures, existing records, or unobtrusive indicators when these provide valid information and satisfy ethical and privacy requirements.

For repeated performance assessments, equivalent alternative test forms may help reduce item familiarity, although their comparability must be established.

Measurement procedures should be standardized across relevant groups, and researchers should avoid giving feedback that could function as an unintended intervention unless feedback is part of the protocol.

Researchers should also consider the timing of baseline assessments. A short interval between pretesting and intervention may be useful for representing current conditions, but it could also intensify immediate priming or fatigue.

None of these strategies guarantees the absence of reactivity. Their appropriateness depends on the assessment and the causal question.

Can Measurement Reactivity Affect Follow-Up Assessments Too?

Yes. Repeated follow-up assessments may encourage self-monitoring, provide practice, or repeatedly remind participants of the study's target behavior.

For example, participants asked to report weekly exercise may become more attentive to physical activity than participants assessed only at the end of the study.

Thus, the number of follow-up measurements can influence not only the amount of information collected but also the experience of participation.

Researchers should distinguish the scientific value of frequent assessment from the possibility that assessment frequency changes the phenomenon being observed.

When a Study-Habits Questionnaire Becomes Part of the Intervention

Suppose the control group increases its study planning after completing the baseline questionnaire. That change alone would not prove measurement reactivity, because other explanations remain possible.

However, if a suitably designed randomized comparison showed different outcomes between pretested and non-pretested control participants, it would provide stronger evidence that the pretest influenced subsequent behavior.

The important lesson is that baseline assessment should be treated as a potentially active research procedure when its content overlaps with the behavior an intervention seeks to change.

Misconceptions About Baseline Measurement Effects

Assess Whether Your Baseline Procedure Could Function as an Intervention

Review the content and administration of each proposed baseline measure. Consider what participants may learn, notice, practice, or infer from completing it.

The concern is especially relevant when the assessment addresses the same knowledge, intentions, or behaviors targeted by the intervention.

Where measurement reactivity is plausible, describe the assessment procedures sufficiently for readers to understand the conditions under which the intervention was evaluated.

Acknowledge uncertainty rather than claiming that a measurement effect occurred without evidence capable of distinguishing it from alternative explanations.

Before Administering a Potentially Reactive Baseline Assessment

Frequently Asked Questions About Measurement Reactivity

Baseline Measurement Can Become Part of the Causal Process

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.