Original Research Publication

Overview

You are going to do something that most people never do at any age: produce a genuine, original contribution to human knowledge and put it in front of the world to be judged. Not a book report, not a recreation of a known experiment, but a study you designed, around a question nobody had answered the way you will answer it, with data you collected and analyzed honestly — and then submitted to a venue with real reviewers who can reject it. This is the full research cycle, run at professional standards, with your name on the result.

This matters far beyond academia. Research is the disciplined version of the most important skill an adult can have: figuring out what is actually true when the answer is not already known and the easy answers are wrong. A founder running an experiment on customer behavior, an engineer testing whether a design holds, a policy analyst measuring whether an intervention worked — all of them are doing research, whether or not they call it that. The habits you build here, of separating what you want to be true from what the data shows, of designing a test that could prove you wrong, of reporting the inconvenient result, are the habits that separate people who discover things from people who fool themselves and everyone around them.

The Deliverable

Two things, both judged by real-world standards:

1. A complete research paper or report, structured the way the field structures them — typically an abstract, an introduction stating the question and why it matters, a methods section detailed enough that someone else could replicate your study, results presented with honest data and clear figures, a discussion of what the findings mean and where they fall short, and a properly formatted reference list. Length follows the venue, but expect 3,000-6,000 words plus figures.

2. A real submission to a real external venue. This is the bar that makes it count. The venue must have reviewers who are not your family — a student research journal, a regional or national science fair with judging, an undergraduate research conference, a discipline-specific competition, or at minimum a public preprint server with an invited critique from a domain expert. "Done" means you submitted and received external response, whether acceptance, revision requests, or rejection. A rejection you learned from is a completed project. A paper sitting in a drawer is not.

The topic is yours, but the question must clear three bars: it must be answerable with methods and data you can actually access; it must be original in some real sense (a new question, a new population, a new method, or a known question in a context nobody checked); and it must be honest — designed so that the data could genuinely come out against your hypothesis.

Materials & Tools

Project Phases

Phase 1: Plan — Question, Hypothesis, and Design (Weeks 1-4)

Research lives or dies in the design, and the most expensive mistakes are the ones you make before collecting a single data point.

Find the question. Start from genuine curiosity or a genuine irritation — something you noticed that the available answers do not explain. Read widely in the area to find out what is already known; you cannot make an original contribution until you know where the edge of the known is. This is the literature review, and skipping it is the classic novice error: people "discover" things that were settled decades ago because they never checked. As you read, you are hunting for a gap — a question that is asked but not answered, answered but not in your context, or never asked at all.

Sharpen it into a hypothesis. A research question becomes science when you can state, in advance, what result would support your idea and what result would refute it. "Does music affect studying?" is a topic. "Instrumental music improves recall on a memorization task relative to silence, but lyrical music does not" is a hypothesis — specific, directional, and falsifiable. Write down explicitly what outcome would prove you wrong. If no possible result could prove you wrong, you do not have a hypothesis; you have a belief, and beliefs are not research.

Design the study that could answer it. This is where you earn the result. Specify exactly what you will measure, how, and on whom or what. Confront the questions that destroy weak studies: What is your control or comparison? How will you sample, and what biases does your sample introduce? What confounds could explain your result other than your hypothesis, and how will you rule them out? How much data do you need before a result means anything rather than nothing? Write the full methods section now, before any data exists, in enough detail that a stranger could run your study from it. Then bring it to your research mentor and ask them to attack it. A mentor who finds a fatal flaw in your design in week three has saved you from discovering it in week twelve, after the data is already useless.

End Phase 1 with a written research proposal: the question, the hypothesis, the falsification condition, and the full methods. This is your contract with reality.

Phase 2: Build — Collect and Analyze the Data (Weeks 5-14)

Milestone 1 — A clean pilot (Weeks 5-6). Before running the full study, run a tiny version of it — a handful of subjects, a few trials, one round of measurements. Pilots exist to find the problems your proposal could not anticipate: the survey question everyone misreads, the instrument that drifts, the procedure that takes four times as long as you planned. Fix the design based on the pilot. The pilot data does not go in your final results, because you changed the design after seeing it — using it would be a thumb on the scale.

Milestone 2 — Full data collection (Weeks 7-12). Run the study as designed and do not change the design midstream to chase a result you like. This is the discipline that defines honest research. Log everything contemporaneously in your research notebook — every data point, every anomaly, every deviation from protocol, dated as it happens. Backfilled notes are unreliable notes, and the day you cannot remember whether subject 14 followed the procedure is the day your study's integrity quietly dies. If something goes wrong, you record it; you do not erase it. The anomalies are often where the real findings hide.

Milestone 3 — Honest analysis (Weeks 13-14). Analyze the data using the method you specified in your proposal, not the method that happens to produce the prettiest result. This is the single most important integrity test in the entire project, and it is worth being blunt about why. There is a powerful, almost gravitational pull to keep slicing the data until something "significant" appears — drop the inconvenient subjects, try a different test, split the sample a new way. This is how honest people accidentally produce false findings, and it has corrupted entire fields of real science. The defense is simple and absolute: you decided how you would analyze the data before you saw it, in your proposal, and you run that analysis and report what it shows. If the result refutes your hypothesis, you report that. A clean null result, honestly obtained, is worth infinitely more than a "positive" result you tortured out of the data — and any competent reviewer can smell the difference.

Phase 3: Test & Refine — Write, and Submit to Honest Critique

Now write the paper, in the structure the field uses. The methods and introduction you can largely assemble from your proposal. The results section presents what you found, in clear figures and plain language, separating what the data shows from what you think it means. The discussion is where you interpret — and where you do the thing that marks a mature researcher: you state the limitations of your own study before a critic does. What could be wrong? What does your study not show? What would the next study do better? A paper that pretends to have no limitations announces that its author cannot see them, which is worse than having them.

Then hand the full draft to your mentor and ideally one other informed reader, and ask for the harshest honest critique. Treat their objections the way reviewers will: separate the criticism that exposes a real flaw in your reasoning or analysis from the criticism that is a difference in emphasis. Fix the real flaws. Check every citation against its source — misrepresenting what a paper found is the credibility-destroying error, and reviewers check. Revise the argument first, the prose second.

Phase 4: Present — Submit to the World

Submit to your chosen venue, formatted exactly to its requirements — formatting sloppiness is the easiest reason for a desk rejection and the easiest to avoid. Then wait for the response, because the external judgment is the part of this project that no parent or mentor can substitute for.

When the response comes, it will be one of three things, and all three are a successful completion of the project. Acceptance means strangers found your work sound — rare and worth celebrating. Revise-and-resubmit is the most common and most valuable outcome: real reviewers found real problems and are telling you how to fix them, which is precisely the apprenticeship into how knowledge actually gets made. Do the revisions; resubmitting a paper in response to review is the core loop of a research career. Rejection stings, but a rejection with reasons is a free expert teardown of your reasoning — read it as the gift it is, extract every fixable lesson, and either revise for a different venue or carry the lessons into your next study. The researchers who go furthest are not the ones who never get rejected. They are the ones who metabolize rejection into a better next paper.

A Worked Example: From Irritation to Finding

To see how the phases connect, follow one realistic study from spark to submission.

A student volunteers at a food pantry and notices something that bothers them: a lot of fresh produce gets thrown out, and the staff's explanation — "people just don't take it" — doesn't sit right, because they have watched people pass over produce they clearly wanted. That irritation is the spark. It is not yet research.

Phase 1. The literature review reveals that food-pantry "client choice" models have been studied, but mostly in large urban pantries, and almost nobody has measured what specifically drives produce uptake in small rural ones. There is the gap. The question sharpens: does the way produce is displayed — bagged-and-hidden versus loose-and-visible-at-the-entrance — affect how much of it clients take? The hypothesis becomes falsifiable: produce displayed loose and prominently will be taken at a higher rate than the same produce bagged and shelved in back. The falsification condition is explicit: if uptake is equal or lower under prominent display, the hypothesis is wrong. The methods get specified down to detail — which produce, measured how (weight taken per client visit), over how many weeks, with the display alternated week by week to control for which produce happened to be in stock. The mentor, a teacher with research training, catches a confound: produce type varies week to week, so the student adds the control of alternating display condition within the same produce types.

Phase 2. A two-week pilot reveals the first real problem: weighing produce per client is too slow and creates a line, changing client behavior. The student redesigns to weigh total produce out at end of day against total in, by condition — a cleaner measure that doesn't disturb the thing being measured. The pilot data is discarded, as it must be. The full collection runs six weeks, display condition alternating, everything logged daily in the notebook including the week the cooler broke (recorded, not erased, because it affects interpretation). At analysis, the pre-specified comparison shows prominent display increased produce uptake by a real and consistent margin — the hypothesis holds. The student resists the temptation to also report a dozen other slices of the data they didn't pre-plan; those go in a clearly-labeled "exploratory" note, not the main result.

Phase 3. The write-up states the finding plainly and then states the limitations honestly: one pantry, one region, six weeks, a broken-cooler week, no measure of whether the produce was actually eaten versus taken-and-tossed. The mentor and a second reader push on the generalizability, and the student narrows the claims to match what the data can bear.

Phase 4. The paper goes to a regional student research symposium with a poster session. A judge asks the sharp question — "how do you know the effect isn't just novelty, with people taking more of anything new at the entrance?" — and the student, who anticipated exactly this, points to the within-produce-type controls. The paper earns a revise-and-resubmit for a student journal. That is a complete, successful project: a real question, an honest study, a real finding, and external judgment that improved the work. The pantry, notably, changed its display based on the result — which is the kind of thing that happens when research is honest enough to be trusted.

On Research Integrity: The Habits That Outlast the Project

It is worth naming directly that the most important thing this project teaches is not a method but a disposition — a relationship with the truth that holds even when the truth is inconvenient. Every shortcut available to a researcher is a small lie told first to oneself: keeping the data point that "must be an error" because it hurts the hypothesis, running the analysis five ways and reporting the one that worked, softening a limitation that a reader deserves to know, citing a paper for a claim it doesn't quite make. None of these feel like fraud in the moment. Each feels like a reasonable judgment call. That is exactly why they are dangerous, and why the defense has to be a habit built in advance rather than a decision made under temptation.

The habits are simple to state and hard to keep: decide your analysis before you see your data; log everything as it happens, including what went wrong; report the result that emerges, not the one you hoped for; state your limitations before a critic finds them; and never let a source say something in your paper that it does not say in the original. A student who builds these habits in one honest study carries them into every consequential thing they will ever do — because the temptation to make reality look more convenient than it is does not end with research. It shows up in business, in argument, in self-assessment, everywhere a person stands to gain from a flattering version of the facts. The researcher's discipline is, at bottom, the discipline of refusing to fool yourself, and it is the most transferable thing in this entire project.

Success Criteria

• The research question is original in a real sense and the hypothesis is genuinely falsifiable — the student can state what result would have proven them wrong
• The methods were written before data collection, reviewed by a mentor, and detailed enough to replicate
• Data was collected honestly, logged contemporaneously, and analyzed by the pre-specified method — including reporting results that went against the hypothesis
• The paper meets the structure and quality standards of the chosen field, states its own limitations, and has no misrepresented sources
• The work was submitted to a real external venue and received external response, and the student can describe what they learned from that response

Common Pitfalls

• Skipping the literature review and "discovering" the known. If you don't map the edge of existing knowledge, you cannot stand at it. Read first.
• A hypothesis that cannot fail. If no result could refute your claim, you are not doing research. Write the falsification condition in advance.
• Changing the design midstream to chase a result. The moment you adjust the study after seeing data you like, the result stops being trustworthy. Lock the design; report what comes.
• Torturing the data until something is "significant." Pre-specify your analysis. A clean null beats a tortured positive every time, and reviewers can tell.
• Hiding limitations. Every real study has them. State yours before a critic does — it is the mark of someone who can see their own work clearly.
• Treating a drawer as a venue. Research that is never submitted is never finished. The external judgment is the point.

Extensions

• Turn revise-and-resubmit into a published paper. If your venue offered revisions, see the cycle all the way through to acceptance. Few experiences teach more about real intellectual work than getting a paper across the line.
• Build a research program, not a one-off. The best findings raise new questions. Use your discussion section's "future work" as the proposal for a second study, and treat your research as a multi-year arc rather than a single project.
• Collaborate. Real science is rarely solo. Find a peer with a complementary skill and co-author a study, learning to divide labor, share credit, and resolve disagreement about what the data means.
• Present it live. Submit to a conference or fair with an oral or poster component and defend your work in person against questioning. Real-time defense of your methods is a different and harder skill than writing them up.
• Make it the seed of a venture or a thesis. Original findings about how some corner of the world actually works are exactly what new companies, nonprofits, and bodies of writing are built on. If you found something true that others have not acted on, that gap may be an opportunity.