Preregistration

1 Definition and purpose

Preregistration is a research practice in which investigators publicly record key elements of a study before collecting or inspecting the data. These elements usually include the research question, hypotheses, design, planned measurements, and analysis strategy. The main purpose is to show which parts of the work were planned in advance and which parts emerged later during analysis.

In many disciplines, preregistration is used to strengthen the interpretive value of findings. It helps readers distinguish confirmatory tests, which evaluate predictions made beforehand, from exploratory analyses, which are used to search for patterns or generate new ideas. By creating a dated record, it also supports transparency and reduces the chance that results are shaped invisibly by post hoc choices.

1.1 Core concept

The core idea of preregistration is simple: document the intended study plan before the data can influence that plan. This can involve a short summary of the hypotheses or a fuller protocol with detailed analytic decisions. The preregistered record becomes a benchmark against which the final report can be compared.

Because the record is made in advance, it provides evidence that certain decisions were not adjusted after the results were known. This does not prevent researchers from changing course when necessary, but it makes such changes visible.

1.2 Scientific rationale

Preregistration addresses several common sources of bias in empirical research. When investigators have many reasonable choices about how to process data or test hypotheses, different choices may lead to different outcomes. If only the most favorable results are reported, the published literature can become distorted.

A preregistered plan encourages discipline in hypothesis testing and analysis. It can lower the risk of selective reporting, accidental overfitting, and overly confident claims based on chance findings. It also supports more reliable comparison across studies because readers can see how closely the final report followed the original plan.

1.3 Distinction from related practices

Preregistration is related to other forms of research documentation, but it is not identical to them. The defining feature is that it is completed before the relevant data are observed and is made publicly accessible in some form. Other documents may be private, may be written after data collection, or may serve different purposes.

1.3.1 Research protocols

A research protocol is a detailed plan for conducting a study. It may include background, methods, ethics, and logistics, and in some settings it is required for internal review or oversight. A preregistration can resemble a protocol, but it is usually intended for public disclosure and for marking the boundary between planned and post hoc work.

1.3.2 Registered reports

Registered reports are a publication format in which journals review the study design before data collection and may offer in-principle acceptance. Preregistration and registered reports share the goal of separating design from results, but they differ in procedure. A preregistration can be done independently of journal submission, whereas a registered report is tied to editorial review.

1.3.3 Data sharing and open science

Preregistration is part of a broader open science landscape that also includes data sharing, code sharing, open materials, and transparent reporting. These practices complement one another but serve different functions. Preregistration clarifies intent, while data and code sharing allow others to inspect and reproduce the analysis.

2 History and development

Preregistration emerged from longstanding concerns about scientific reliability and analytic flexibility. Its modern form developed gradually as researchers, editors, funders, and institutions sought better ways to document study plans before outcomes were known. Although the practice is now associated with digital repositories, its conceptual roots are older.

2.1 Early methodological roots

Earlier scientific traditions often relied on formal protocols, laboratory notebooks, and prospectively defined research plans. In clinical and experimental settings, advance documentation was valued for organizational and ethical reasons as well as for scientific rigor. These older practices did not always involve public posting, but they established the idea that a study should be planned before it is judged.

2.2 Growth in psychology and medicine

Preregistration gained broader visibility in psychology and medicine, where hypothesis testing and outcome reporting are central. In these fields, concerns about selective reporting and undisclosed analytic flexibility made advance documentation especially appealing. As online registries became easier to use, researchers could record plans in a way that was time-stamped and accessible to others.

2.3 Expansion across disciplines

The practice later spread to economics, education research, ecology, and other empirical fields. It proved adaptable because many disciplines face similar challenges: multiple analytic choices, incentives for positive findings, and the difficulty of distinguishing prediction from discovery after the fact. Different fields have tailored preregistration to their own methods, data types, and publication norms.

2.4 Influence of the replication crisis

Growing concern about replication failures and irreproducible results helped drive interest in preregistration. Researchers increasingly questioned whether some published findings had been overstated because of flexible analysis or selective publication. Preregistration became one of several reforms proposed to improve credibility and make research claims easier to evaluate.

3 Components of a preregistration

A preregistration can be brief or highly detailed, but most versions share a common structure. They identify the research question, describe the planned study, and specify how the data will be analyzed. The more precisely these elements are defined, the clearer the distinction between confirmatory and exploratory work.

3.1 Research questions and hypotheses

A preregistration typically begins with one or more research questions. In hypothesis-driven studies, investigators state what they expect to find and what outcome would count as support. Clear wording matters because vague predictions are hard to evaluate after the fact.

3.2 Study design

The design section explains how the study will be carried out. It usually covers the type of study, the participants or units being studied, the conditions or groups involved, and the overall structure of the investigation.

3.2.1 Sampling plan

The sampling plan describes how participants, cases, or observations will be selected. It may include inclusion and exclusion rules, recruitment methods, and target sample size. When feasible, specifying the sampling approach in advance helps prevent post hoc adjustments based on the collected data.

3.2.2 Experimental conditions

For experimental research, the preregistration identifies the conditions or treatments to be compared. It may explain how groups will differ and what manipulations will be applied. This is especially important in randomized experiments, where the interpretation of effects depends on the planned comparison.

3.2.3 Randomization and blinding

If randomization or blinding is used, the preregistration should explain the procedure. Randomization reduces systematic differences between groups, while blinding helps limit expectation effects. Describing these methods in advance makes it easier for readers to judge the rigor of the design.

3.3 Data collection procedures

This section covers how data will be obtained, recorded, and managed. It can include the timing of measurement, the instruments used, and any steps taken to ensure consistency. A clear account of collection procedures helps others understand how the raw observations were produced.

3.4 Outcome measures

Outcome measures are the variables that will be used to evaluate the hypotheses. Preregistration often identifies primary outcomes, secondary outcomes, and any composite measures. Defining them beforehand reduces ambiguity about which results matter most.

3.5 Statistical analysis plan

The statistical analysis plan is one of the most important parts of a preregistration. It states how the data will be processed and tested, including the main models, transformations, and decision rules. A good plan limits ambiguity without becoming unrealistically rigid.

3.5.1 Primary analyses

Primary analyses are the main tests intended to address the central hypotheses. They should be described with enough detail that another researcher could reproduce the intended procedure. This may include statistical models, significance thresholds, or estimation methods.

3.5.2 Secondary analyses

Secondary analyses examine additional questions that are not the central focus of the study. These may include subgroup comparisons, robustness checks, or alternative model specifications. In a preregistration, they are often clearly labeled so they are not mistaken for the primary confirmatory test.

3.5.3 Exclusion criteria

Exclusion criteria explain when data will be removed from analysis. Examples include failed attention checks, missing key measurements, or technical errors. Predefining these rules helps prevent selective removal of inconvenient observations.

3.6 Deviations and amendments

No preregistration can anticipate every complication. When changes become necessary, researchers may revise their plans, but they should document the modification and explain why it was made. Transparent amendment records preserve the value of preregistration while allowing reasonable adaptation to unforeseen circumstances.

4 Types of preregistration

Preregistrations vary in length, specificity, and purpose. Some are little more than a time-stamped note of a study’s main claims, while others resemble complete protocols. The appropriate format depends on the field, the question, and the intended use.

4.1 Simple preregistration

A simple preregistration records the essential predictions and broad method in a concise form. It is often used when the main goal is to establish a dated public record. This approach is quick to prepare, but it may leave some analytic choices open.

4.2 Detailed preregistration

A detailed preregistration includes extensive information about sampling, measurement, exclusions, and statistical modeling. It can be especially useful for studies with many possible analytic paths. The added specificity makes later deviations easier to identify.

4.3 Flexible and exploratory preregistration

Some researchers use preregistration in a more flexible way, combining planned hypotheses with room for discovery. In these cases, the document may specify which elements are confirmatory and which parts are exploratory. This allows a study to remain open to unexpected findings without losing transparency.

4.4 Clinical trial registration

Clinical trial registration is a specialized form of preregistration used in medical research. It generally records the study purpose, design, intervention, and outcomes before participant enrollment begins. Because health studies can affect patient care and public trust, this form of advance registration is especially significant.

4.5 Archival and retrospective registration

Archival registration refers to recording a study plan in a repository that preserves the dated version. Retrospective registration, by contrast, occurs after some data may already have been collected. Retrospective entries can still be useful for documentation, but they do not offer the same evidentiary value as prospective preregistration.

5 Platforms and registries

Preregistrations are usually deposited in online platforms or registries that provide a timestamped record. Some are general-purpose, while others are tailored to specific disciplines or study types. The choice of platform often depends on the level of detail needed and the norms of the field.

5.1 Open Science Framework

The Open Science Framework is a widely used platform for preregistration and related research materials. It allows users to create time-stamped records, store versions of documents, and link preregistrations to datasets or code. Its flexibility makes it suitable for many kinds of projects.

5.2 ClinicalTrials.gov

ClinicalTrials.gov is a major registry for clinical studies. It is often used to record medical trial details such as interventions, outcome measures, and enrollment information. Because it is built for health research, it plays a central role in trial transparency.

5.3 AsPredicted

AsPredicted is a streamlined preregistration platform designed for concise entries. It asks researchers to answer a small set of focused questions about their hypotheses and analyses. Its simplicity makes it attractive for studies that need a quick, standardized record.

5.4 Field-specific registries

Some disciplines use specialized registries that reflect their methods and reporting standards. These may be designed for economics, ecology, social science, or other areas. Field-specific systems can make preregistration more practical by matching the language and workflow of the discipline.

5.5 Versioning and time-stamping

Versioning and time-stamping are central to the credibility of preregistration. A timestamp shows when a plan was recorded, and version history shows how it changed over time. These features help distinguish the original plan from later revisions and preserve an audit trail.

6 Benefits

Preregistration offers several methodological advantages. Its value is not limited to any single field, because it supports clearer reasoning about what a study did and did not test. The strongest benefits are usually tied to transparency and interpretability.

6.1 Transparency

By making the plan public in advance, preregistration increases openness about how a study was conceived. Readers can see which outcomes and analyses were intended from the start. This visibility makes it easier to evaluate the credibility of the final report.

6.2 Reduction of p-hacking and HARKing

Preregistration can reduce p-hacking, meaning the practice of trying many analyses until a statistically significant result appears. It can also discourage HARKing, or hypothesizing after the results are known. Both behaviors can blur the line between testing and searching, whereas preregistration makes that line more explicit.

6.3 Improved reproducibility

When the planned methods are documented in advance, other researchers can more easily reproduce the study logic. Even if exact replication is not possible, the preregistered record clarifies what was intended and how the results should be interpreted. This can improve confidence in cumulative findings.

6.4 Clearer separation of confirmatory and exploratory work

Preregistration helps researchers and readers separate confirmatory claims from exploratory ones. This distinction is important because the two kinds of analysis answer different questions and carry different evidentiary weight. A clear boundary prevents exploratory discoveries from being mistaken for strong prior predictions.

6.5 Public accountability

A preregistered study creates a public commitment to a plan. That commitment can encourage careful design and more disciplined reporting. It may also increase trust among readers, reviewers, and collaborators because the study path is visible from the outset.

7 Limitations and criticisms

Preregistration is widely praised, but it is not a complete solution to problems in research quality. Critics note that it can be demanding to use and may not suit every project equally well. Its value depends heavily on how thoughtfully it is implemented.

7.1 Added administrative burden

Preparing a preregistration takes time, especially when the study design is complex. Researchers may need to anticipate many analytic details before data collection begins. For small teams or fast-moving projects, this added work can be a practical obstacle.

7.2 Reduced flexibility in unexpected situations

Research often encounters unforeseen problems such as missing data, technical failures, or sampling difficulties. A preregistered plan cannot always anticipate these issues. If researchers treat the document as an inflexible rulebook, they may struggle to adapt sensibly when circumstances change.

7.3 Incomplete preregistrations

Some preregistrations are too vague to prevent ambiguity. If key decisions are left unspecified, researchers may still have substantial flexibility during analysis. In such cases, the preregistration exists in form but offers limited protection against post hoc interpretation.

7.4 Misinterpretation of exploratory research

A risk of preregistration is that exploratory work may be undervalued if it is not clearly presented. Discovery-oriented analysis is a normal and important part of science. The challenge is not to suppress exploration, but to label it honestly so that its role in the research process is understood.

7.5 Suitability across different research traditions

Not all research traditions fit neatly into a preregistration model. Some studies are iterative, theoretical, or highly observational, making fixed advance plans less straightforward. Even so, many of these projects can still benefit from documenting goals and analytic intentions in advance when feasible.

8 Best practices

Effective preregistration depends on clarity, specificity, and honesty. The goal is not perfection, but a record that meaningfully constrains later reinterpretation. Good practice makes the study plan understandable to both specialists and informed readers.

8.1 Writing clear hypotheses

Hypotheses should be stated in precise, testable terms. Ambiguous predictions are difficult to evaluate and may weaken the value of the preregistration. A clear hypothesis identifies the expected direction, outcome, and comparison when possible.

8.2 Specifying analysis decisions in advance

The more important the analysis choice, the more clearly it should be documented. This includes transformations, thresholds, covariates, and model-selection rules. Advance specification helps prevent silent changes that could influence the outcome.

8.3 Distinguishing confirmatory and exploratory analyses

Best practice is to label confirmatory analyses separately from exploratory ones. This distinction should appear in the preregistration and in the final article. Doing so allows discovery-oriented work to be appreciated without being mistaken for preplanned testing.

8.4 Updating protocols transparently

If the plan must change, the revision should be recorded with an explanation. Transparent updates preserve the historical record and show that the change was not hidden. This is especially useful when unexpected methodological issues arise during data collection.

8.5 Sharing preregistrations with datasets and code

Linking the preregistration to the final dataset and analysis code improves traceability. Readers can compare the intended and actual procedures more easily. When combined with open materials, this practice creates a more complete account of the research process.

9 Applications by field

Preregistration is used in many domains, but it takes different forms depending on the research culture and methods. Some fields rely on highly structured statistical testing, while others use preregistration mainly to document decisions and strengthen transparency. The same principle can therefore support different kinds of inquiry.

9.1 Psychology

Psychology has been one of the most visible fields for preregistration. Researchers often use it to define hypotheses, outcome measures, and exclusions before collecting participant data. It is especially helpful in experiments with multiple plausible analytic approaches.

9.2 Medicine and clinical research

In medicine, preregistration is closely tied to trials, outcomes, and patient safety. It helps ensure that endpoints are set in advance and that treatment effects are assessed consistently. The practice is widely viewed as a standard component of responsible clinical research.

9.3 Economics

Economists use preregistration to record empirical strategies, treatment comparisons, and statistical specifications. It can be especially useful in experimental and quasi-experimental work where several modeling decisions are possible. The resulting record helps clarify which conclusions were planned before the analysis began.

9.4 Education research

Education researchers use preregistration to improve the credibility of intervention studies and policy evaluations. It helps define the target population, outcome metrics, and comparison groups ahead of time. This can be valuable in settings where results may influence institutional practice.

9.5 Behavioral sciences

Behavioral science often combines experimental, observational, and survey-based methods. Preregistration can be adapted to all of these formats by documenting the central predictions and the main analytic pathway. It is particularly useful when multiple behavioral outcomes might be measured.

9.6 Ecology and environmental science

In ecology and environmental science, preregistration can support studies of field experiments, monitoring programs, and observational datasets. It may be used to specify sampling windows, model structures, or decision criteria before analysis begins. This is helpful when data are complex and the number of plausible tests is large.

10 Evaluation and impact

Researchers have studied preregistration not only as a method, but also as an intervention with measurable effects. The evidence suggests that it can improve some aspects of reporting, although its impact varies by context and implementation. Ongoing evaluation remains important because the practice continues to evolve.

10.1 Evidence for improved reporting

Studies of preregistered projects often find better documentation of hypotheses, analysis choices, and deviations from plan. This does not guarantee better scientific conclusions, but it can make the research record easier to interpret. Improvements are usually strongest when preregistrations are detailed and followed closely.

10.2 Effects on publication outcomes

Some work has examined whether preregistration changes the likelihood of publication or the type of results reported. The findings are mixed, in part because publication depends on many factors beyond registration. Even so, preregistration may reduce the visibility of selective outcome reporting by making the original plan public.

10.3 Meta-research on adherence

Meta-research has explored how often published studies comply with their preregistered plans. Such work shows that adherence is uneven and that researchers sometimes omit important details or revise plans without clear explanation. These studies have helped identify where preregistration guidance and tools need improvement.

10.4 Future directions

Future development is likely to focus on better usability, stronger standards, and closer integration with journals and data repositories. Improved templates and discipline-specific guidance may make preregistration easier to adopt without excessive burden. As the practice matures, its role will likely remain tied to transparency, accountability, and clearer separation of planned and discovered results.