Pre-Registration's Paradox: How a Tool Built for Transparency Is Quietly Suppressing Null Results

A Reform That May Be Reforming in the Wrong Direction

When the Open Science Framework began promoting pre-registration as a cornerstone of reproducible research in the early 2010s, the scientific community greeted it with considerable optimism. By requiring researchers to publicly document their hypotheses, methods, and analytic plans before data collection began, the protocol promised to close the loopholes that had enabled p-hacking, HARKing (Hypothesizing After Results are Known), and the selective suppression of inconvenient outcomes. The logic was sound. The implementation, however, may be producing consequences that nobody anticipated.

A growing body of methodological criticism now suggests that pre-registration, far from eliminating the burial of null results, may be redirecting that burial through a different mechanism entirely. Rather than manipulating analyses after the fact, researchers appear to be abandoning pre-registered studies altogether once those studies yield unexpected or null findings—leaving registered protocols as inert digital artifacts while the data they generated quietly disappears from the scientific record.

The Mechanics of Abandonment

To understand how this occurs, it is necessary to appreciate the institutional context in which pre-registration operates. A researcher who registers a study and subsequently obtains results that contradict the stated hypothesis faces a decision architecture that differs sharply from the one faced by researchers in the pre-registration era. Previously, a null result could be reframed, its analysis adjusted, or its manuscript simply withheld without any public trace of the failed inquiry. Pre-registration was designed to eliminate precisely that option by creating a timestamped public record.

What the reformers did not fully anticipate was that the existence of a public record does not obligate anyone to act on it. A pre-registered study that is never completed, never written up, and never submitted to a journal leaves behind only a registration entry—a placeholder that signals the existence of a question without ever communicating the answer. In the absence of mandatory outcome reporting tied to registration, the protocol creates accountability for the plan while leaving the results entirely unaccountable.

Data from the ClinicalTrials.gov registry, which operates under stronger reporting mandates than most academic pre-registration platforms, has long demonstrated that even legally required disclosure fails to achieve full compliance. A 2015 analysis published in The BMJ found that fewer than half of registered clinical trials reported results within two years of completion. If compliance is incomplete even when regulatory penalties apply, the situation in voluntary academic registries is likely considerably worse.

Incentive Structures Compound the Problem

The abandonment dynamic is not simply a matter of individual researcher misconduct. It is, in large measure, a predictable response to incentive structures that have not been reformed alongside the adoption of pre-registration. Academic careers in the United States continue to be evaluated primarily on the basis of publication counts, citation impact, and the successful acquisition of competitive grant funding. None of these metrics reward the completion and dissemination of null results.

A junior faculty member at a research university who pre-registers a study, collects data, and obtains a clean null result faces a calculus that is strikingly unfavorable. Publishing that null result requires identifying a journal willing to accept it—a task made easier by registered report formats but still considerably harder than placing a positive finding. The publication, once achieved, will likely generate modest citation counts. The time invested in writing it up could, in theory, have been redirected toward a new study with better prospects for a compelling positive outcome.

Given this landscape, abandonment is not irrational. It is, in fact, the behavior that institutional incentives reliably produce. Pre-registration changes the moment at which the decision to suppress occurs; it does not eliminate the underlying motivation.

The Registered Report as Partial Remedy

Some journals have responded to this structural problem by adopting the registered report format, in which peer review occurs in two stages. Reviewers evaluate the introduction and methods before data collection begins, and the journal commits to publishing the results regardless of outcome, provided the approved protocol was followed. This model directly addresses the abandonment problem by decoupling publication decisions from result valence.

The uptake of registered reports has grown substantially since the format was pioneered by Perspectives on Psychological Science in 2013. As of 2024, more than three hundred journals across disciplines ranging from neuroscience to economics offer the format. Yet registered reports still represent a small fraction of total scientific output, and their adoption is uneven across fields. In many areas of biomedical research, engineering, and the social sciences, the standard pre-registration model—without any commitment to publication—remains dominant.

Toward a More Complete Accounting

Addressing the abandonment problem requires interventions at multiple levels simultaneously. At the platform level, pre-registration systems should implement structured follow-up mechanisms that flag stale registrations and prompt researchers to report outcomes or formally document the reasons for non-completion. The Open Science Framework and AsPredicted already possess the technical infrastructure to do this; the question is whether the scientific community will collectively decide that such prompting is appropriate.

At the funding level, agencies including the National Institutes of Health and the National Science Foundation could condition grant renewals on the documented completion of previously funded pre-registered studies, regardless of outcome. Such a requirement would create a direct institutional incentive for follow-through that does not currently exist.

At the journal level, expanded adoption of registered reports and the continued development of null-result venues—such as the Journal of Articles in Support of the Null Hypothesis—provide partial solutions, though they depend on researchers choosing to submit rather than simply walk away.

None of these interventions addresses the root issue, which is that scientific careers are still built on positive discoveries rather than on the comprehensive mapping of what does and does not hold true. Until that calculus changes, any transparency mechanism will be vulnerable to the same pressures that created the original replication crisis. Pre-registration was a genuine step forward. Recognizing its limitations is not a repudiation of that step—it is the rigorous inquiry that the step itself demands.