Is Sunk-Cost Thinking Already Influencing Your Decision to Continue Developing the Idea?

Would You Still Choose This Research Idea If You Were Starting Today?

You have spent weeks reading the literature. The conceptual framework finally makes sense. You have drafted the proposal, developed an instrument, discussed the project with colleagues, and perhaps learned a new method specifically to conduct it.

Then something changes.

You discover stronger existing evidence than expected. A central assumption looks weaker. Recruitment seems less realistic. Another design could answer the important question more efficiently. The study still can be done, but its original rationale no longer looks quite as convincing.

At this point, past investment can become psychologically entangled with the decision about future investment. Instead of asking whether the idea still deserves another six months, you may find yourself thinking about the six months you have already spent. That is precisely when sunk-cost thinking deserves scrutiny.

Past Investment Should Not Be Mistaken for Evidence That Continuing Is Wise

The Relevant Decision Is About the Resources You Have Not Spent Yet

A sunk cost is a past investment you cannot recover

A sunk cost is a resource already committed that cannot be recovered by your present decision. In research, sunk costs may include time spent reviewing literature, developing a proposal, learning software, creating instruments, running preliminary analyses, collecting unusable data, preparing ethics documents, or repeatedly revising a project.

Continuing does not recover those resources. Stopping does not make them disappear. They belong to the history of the project either way.

The decision in front of you concerns what happens next.

The sunk-cost effect is a documented decision bias, but context matters

Experimental research has repeatedly examined the tendency for prior investment to influence subsequent choices even when those costs should not determine the prospective value of the options. Classic work by Arkes and Blumer described this as the sunk-cost effect, while later research has investigated how factors such as responsibility, waste avoidance, anticipated regret, and continued commitment may contribute to such decisions.

Researchers should nevertheless avoid turning “sunk cost” into a label for every decision to continue a difficult project. Persistence can be rational. A study may become harder than expected while remaining scientifically valuable and feasible. Additional investment may also unlock value from assets that already exist.

The diagnostic issue is narrower: are you continuing because the future work is still worth doing, or because stopping would make the past effort feel wasted?

Research creates particularly persuasive sunk costs

Research investments are rarely just financial. They become intellectual and sometimes social commitments.

You may have spent months mastering the literature. Your supervisor may have approved the topic. Colleagues may know you as “the person working on” that question. You may have presented the proposal publicly. Perhaps the study forms part of a thesis or grant application.

These investments can make abandonment feel like admitting that previous work was a mistake.

But that interpretation is unnecessary. A research idea can have been reasonable when you chose it and become less attractive after new information appears. Updating the decision is not evidence that the original decision was irrational.

Ask whether the rationale has changed since you began

Sunk-cost problems become especially relevant when the state of evidence changes.

Perhaps another research team publishes a stronger study. A systematic review resolves much of the uncertainty. Pilot work shows that your key measure performs poorly. Access to the intended population becomes unrealistic. A core theoretical assumption no longer appears credible.

None of these developments automatically means you should stop. They do mean the current project should be evaluated using the information available now rather than the information available when you first committed to it.

If the reason for asking the research question has substantially weakened, previous investment cannot restore that rationale.

Past work can have residual value without justifying the original project

Here is an important qualification: sunk costs are not necessarily worthless costs.

A literature review may support a different paper. An instrument may be useful in another project. Code can be reused. Pilot findings may expose a better question. Methodological expertise remains with you. A dataset may support a narrower analysis even when the original design becomes untenable.

Those are prospective benefits and should be included in your decision.

What you should not do is count the mere fact that you spent the resources as evidence that the original study deserves completion.

Beware of changing the rationale merely to keep the project alive

A project may begin with one contribution and gradually acquire another whenever the original rationale weakens.

The intervention is already well studied, but perhaps your population is different. That population has also been studied, but perhaps your institution is unique. The outcome is already established, but perhaps adding another mediator creates novelty.

Any individual revision might be defensible. The warning sign is a pattern in which the question repeatedly changes just enough to prevent the project from being abandoned.

This is why it helps to specify what evidence would make you change or abandon the research idea. Without such conditions, the rationale can become indefinitely adjustable.

Opportunity cost is the missing half of the decision

Continuing one project means not using those same resources elsewhere.

Another six months spent rescuing a weak idea could have been used to develop a stronger one. A grant devoted to a low-information study cannot simultaneously fund another project. Participants recruited into one study cannot give the same time to every alternative study.

Research-priority literature has long emphasized the importance of reducing avoidable waste and allocating finite research resources toward questions capable of producing useful evidence.

The correct comparison is therefore not “continue versus lose everything.” The comparison is “continue with this project versus use the remaining resources differently.”

Use the fresh-start test

Imagine that the current project belongs to somebody else.

You know everything you know now about the literature, feasibility, likely contribution, remaining workload, available methods, and competing opportunities. None of the past investment belongs to you.

Would you recommend starting the project today?

If yes, continuing may be entirely rational. If no, ask what has changed merely because you happen to be the person who already invested in it.

This is not a perfect test. Changing projects can itself impose real future costs, especially in degree programs, funded work, collaborative projects, or studies already involving participants. Those consequences belong in the prospective calculation. The test simply removes unrecoverable effort from the argument.

Do not confuse perseverance with escalation

Good research often requires persistence. Recruitment fails. Analyses need revision. Reviewers disagree. Instruments misbehave. A project should not be abandoned at the first inconvenience.

Perseverance remains rational when the expected contribution justifies solving the problem.

Escalation becomes concerning when each setback generates additional investment despite mounting evidence that the project's expected value is deteriorating. The difference lies not in how hard you work but in whether new evidence continues to justify the work.

The hardest question is sometimes what you would do with the remaining resources

Abandonment is easier to evaluate when there is a concrete alternative.

Suppose you have four months of research time remaining. You could spend them repairing a project whose contribution has weakened, or redesign the study around a sharper question. Perhaps you could analyze an existing dataset instead, or redirect the work toward an uncertainty that matters more.

Once alternatives become visible, continuing no longer feels like the default. It becomes one option competing for future resources.

Six Months of Work Do Not Automatically Justify Another Six

The six months were not magically recovered. Nor were they necessarily wasted. They produced knowledge that helped identify a stronger question. What changed was the decision rule: past effort stopped being treated as a reason to commit future effort to the original design.

Sunk-Cost Thinking Is Not an Argument for Abandoning Difficult Research

Evaluate the Research Idea From Today Forward

Ignore, temporarily, how much you have already invested. List what remains to be done, what the completed study could contribute given current evidence, and what else those resources could support.

This is closely related to asking what would make you abandon the idea before investing even more. The later you ask that question, the harder it may become to separate scientific reasons for continuing from the psychological cost of letting go.

Check Whether Past Investment Is Distorting the Decision

Questions About Sunk Costs in Research Decisions

Judge the Next Month of Research, Not the Months You Cannot Get Back

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.