03 · What You Need to Know
Most Research Advances Knowledge in Smaller Steps
Transformative research is not the standard every study must meet
Some research changes how an entire field thinks. It introduces a powerful theory, reveals an unexpected phenomenon, establishes a new method, overturns a longstanding assumption, or creates a new line of investigation. Such contributions deserve attention.
They are not, however, an appropriate benchmark for every research project.
Even funding systems that explicitly support transformative research distinguish such programs from conventional research. The U.S. National Institutes of Health, for example, describes its Transformative Research Award as supporting unusually bold projects with the potential for major impact and explicitly distinguishes that mechanism from conventional investigator-initiated research.
That distinction is useful because it exposes a common category error. A study can be scientifically worthwhile without being transformative. If every acceptable project had to overturn a paradigm, most researchers would spend considerably more time overturning paradigms than actually accumulating evidence.
Small contributions are part of cumulative knowledge building
Research rarely advances through isolated breakthroughs alone. Claims become credible because they are examined repeatedly, with different samples, measures, designs, analytical choices, settings, and assumptions.
Incremental science can include replication, validation, methodological refinement, improved measurement, extension to relevant populations, more precise estimation, or testing of conditions under which an established finding does and does not hold.
Recent discussion of incremental science in educational psychology similarly characterizes scientific progress as cumulative and programmatic, with smaller studies potentially establishing evidence needed for later and larger investigations.
This means the question “Is my contribution large?” may be less informative than “Where does this contribution sit in the accumulation of evidence?”
An initial finding appears A study provides evidence for a potentially important relationship or effect.
Other studies test its reliability Researchers examine whether the finding can be reproduced and how sensitive it is to methodological choices.
Later studies refine its boundaries Evidence accumulates about where, when, for whom, and under what conditions the finding holds.
The field develops a more defensible conclusion Knowledge becomes more precise because many contributions collectively establish what one study could not.
No individual step in that sequence necessarily transforms the field. Yet removing the apparently “small” steps would leave the larger conclusion poorly supported.
Novelty and importance are not the same thing
A particularly important distinction is between how new something is and how much it matters.
Current NIH guidance for evaluating research importance explicitly separates these ideas. Reviewers are asked to consider what will be learned, how valuable that knowledge will be, and how novelty affects the importance of the project. Crucially, the guidance notes that some important projects may use existing methods to answer a critical question, in which case the absence of substantial innovation does not necessarily reduce their importance.
Novelty
How does the study differ from what has already been done or known?
Importance
How consequential is the knowledge gained, uncertainty reduced, capability improved, or decision informed by the study?
A study can therefore be highly novel but unimportant. Imagine applying a sophisticated new analytical technique to a question nobody has a compelling reason to answer. The method may be new in that context, yet the resulting contribution could remain trivial.
Conversely, a study might use entirely conventional methods to answer a consequential question for which the existing evidence is inadequate. Its methodological novelty could be minimal while its contribution is substantial.
This is why judging whether a research question is important enough to study requires more than demonstrating that nobody has conducted your exact study before.
A replication may make a small but important contribution
Suppose an influential finding has been demonstrated in only one or two studies. A carefully designed replication may produce little conceptual novelty. The variables, hypothesis, and basic analytical logic might already be familiar.
What changes if the replication is successful?
Confidence in the finding may increase. If the result does not replicate, researchers may need to reconsider its robustness, measurement, context, or underlying explanation. Either outcome can improve the evidential basis of the field when the original claim is important and genuine uncertainty remains.
The contribution is therefore not “we studied the same thing again.” It is the reduction of uncertainty about whether an important finding is reliable.
Testing a meaningful boundary can extend existing knowledge
Another common incremental contribution involves asking whether an established relationship holds under circumstances where there is a good reason to be uncertain.
Suppose a learning effect has been studied extensively among adults but rarely among younger learners. Extending the work to adolescents might be worthwhile if developmental differences provide a credible reason to expect the effect, mechanism, or appropriate interpretation to differ.
By contrast, repeating a study in another university simply because “no study has been conducted at University X” is not automatically a meaningful extension.
The difference lies in the reason the new context matters.
Watch Out
A change in population, institution, country, platform, subject area, or year does not automatically create a meaningful contribution. Explain why that difference could plausibly change the phenomenon, test the limits of existing knowledge, address an important evidence deficit, or matter to people who need the answer.
Greater precision can itself be useful
Sometimes researchers already know approximately what happens, but they do not know it precisely enough.
A larger or better-designed study may produce a more precise estimate. Improved measurement may distinguish effects that previous instruments blurred together. Better sampling may provide stronger evidence about a population. Additional observations may clarify how much an effect varies across contexts.
These contributions can appear modest because they do not necessarily change the direction of an established conclusion. Yet precision matters when the magnitude of an effect influences theory, policy, resource allocation, intervention design, or subsequent research.
“We already know there is an effect” does not necessarily mean “we know enough about the effect for the purposes that matter.”
Methodological refinements can accumulate into substantial improvements
A new method does not have to replace an entire research paradigm to be useful.
A study might improve the reliability of a coding procedure, reduce bias in a measurement process, simplify data collection, validate an instrument for an appropriate population, improve an algorithm's performance under a consequential condition, or identify an analytical choice that affects interpretation.
Each improvement may appear small in isolation. If the method is widely used, however, a modest improvement can influence many subsequent studies.
This illustrates why contribution size cannot be judged solely by how dramatic the paper sounds. A technically small change applied to an important bottleneck may be more consequential than a conceptually flashy addition to a peripheral problem.
A contribution can be important because it corrects rather than expands
Researchers naturally associate contribution with adding something: another finding, variable, model, theory, dataset, or application.
Sometimes the more valuable contribution is subtraction or correction.
A study may show that a widely cited relationship is weaker after controlling for a measurement problem. It might demonstrate that two supposedly distinct constructs are difficult to distinguish empirically. It may identify a coding error in a commonly used dataset, reveal that an effect depends heavily on one analytical assumption, or establish that a measure performs poorly in a particular context.
The contribution may occupy only a narrow part of the literature, but it can improve the reliability of everything built on that part.
Small contributions can matter more when they address consequential uncertainty
Imagine two studies.
The first introduces a completely new variable into an already crowded model, but there is little reason to believe knowing its association will change theory, practice, or subsequent investigation. The second uses an established method to determine whether an influential result remains credible under a methodological condition that previous studies overlooked.
The first study may appear more novel. The second may contribute more.
This is why the importance of the underlying uncertainty matters. NIH's current review guidance similarly asks reviewers to evaluate the importance of the proposed research itself rather than merely the importance of the broader field or topic.
Studying an important topic does not automatically make every possible question about that topic important.
Ask what becomes different after your contribution
A practical way to evaluate a modest contribution is to imagine the literature before and after your study.
Before your study, what can researchers reasonably claim? After your study, what could they claim more confidently, precisely, broadly, narrowly, or cautiously?
Perhaps an effect previously demonstrated in one context is now supported across several. Perhaps a broad claim now needs qualification. Perhaps an instrument can now be used with stronger evidence in a particular population. Perhaps a previously uncertain estimate is precise enough to inform another study. Perhaps researchers discover that an assumption they routinely make requires reconsideration.
If you can articulate that difference, you can begin to evaluate whether the contribution is worth making.
Small is not the same as redundant
Incremental research still needs a reason to exist.
If twenty strong studies have already established a relationship across the populations and conditions that matter, conducting a twenty-first nearly identical study may add very little. If an instrument has already been thoroughly validated for the population and purpose you intend, another validation study may have little marginal value.
The important concept here is not simply contribution but marginal contribution: what additional knowledge does this particular study provide beyond what is already reasonably known?
As evidence accumulates, another similar study may contribute progressively less unless it addresses a remaining uncertainty.
| Small but potentially meaningful |
Small and potentially trivial |
| Replicates an important but uncertain finding |
Repeats an already well-established result without addressing remaining uncertainty |
| Tests a theoretically meaningful boundary condition |
Changes the location or sample without explaining why the difference matters |
| Improves measurement of an important construct |
Creates another instrument despite adequate existing measures without a clear advantage |
| Provides a substantially more precise estimate |
Adds observations that do not meaningfully improve inference |
| Corrects a consequential methodological weakness |
Makes a technical modification with no meaningful effect on what can be concluded |
| Adds evidence needed for a cumulative research program |
Produces another isolated finding with no clear connection to an unresolved question |
Not every publication venue expects the same magnitude of contribution
Expectations also vary among disciplines, journals, funding schemes, degree programs, and article types.
Some venues explicitly seek highly innovative or transformative work. Others prioritize methodological rigor, reproducibility, useful datasets, replication, validation, or solid extensions of existing knowledge.
For example, editors of the Journal of Cheminformatics have stated that they do not assess submissions purely on scientific novelty, but also consider utility, availability, and the scientific contribution itself. That is one journal's editorial position rather than a universal rule, but it illustrates why “Is this novel enough?” and “Is this worth contributing?” are not always the same editorial question.
You should therefore understand the standards of the community and venue in which your work will be evaluated. But do not reverse the logic and select a question merely because you think its contribution will satisfy a publication threshold. Whether you should study a question because it appears publishable is a different decision from whether the question is genuinely worth pursuing.
A modest contribution should be described modestly
If your study makes an incremental contribution, there is no need to disguise it with inflated language.
A paper that tests the robustness of an important result does not need to claim that it “revolutionizes” understanding. A study extending evidence to a theoretically relevant context need not announce that it “fills a critical void” unless that characterization is genuinely defensible.
Describe exactly what changes.
“This study tests whether the relationship persists under conditions not examined in previous work” is often more informative than “This groundbreaking study addresses a major gap.” The first statement allows the reader to evaluate the contribution. The second mostly asks the adjective to do the reviewing.
The right question is whether the contribution justifies the study
A contribution should ultimately be evaluated relative to what the research requires.
A modest study using existing data to resolve a narrow but meaningful uncertainty may be easy to justify. The same expected knowledge gain may be harder to justify if the project requires substantial funding, years of work, scarce samples, burdensome participation, or significant ethical exposure.
There is no universal equation that converts contribution size into research worth. Proportionality matters.
This is also why a balance among scientific importance, practical relevance, and personal interest can be more useful than searching for a single threshold of contribution. A modest but well-justified study may be exactly the research worth doing next.
