TL;DR
This post covers seven of the most common ways in which your high school research paper could go off course, and provides you guidelines and resources to come up with a plan to solve each kind of problem. Results problems are covered in items 1 to 3, access and support problems are tackled in items 4 and 5, while scope and submission problems are answered in items 6 and 7. Weighing a fresh start? The keep-or-switch table helps you decide. You can usually sort these out within a week or two and keep the project you have. No data, no mentor, or no time? You can work through it with a PhD mentor.
Why do high school research papers go off plan?
Plans change in nearly every study, and published scientists rewrite theirs too. Most of the common reasons research projects fail come down to planning choices you can adjust next time. See if any of the following sound familiar to you:
- You picked a question that needed more months than you had. “How does social media affect teens?” could fill a career. There is a very helpful article that can help you in sizing a topic, and we highly recommend going through it in detail: How To Pick the Best Research Topic: Broad vs Niche.
- You started full data collection without a pilot test, without doing a trial run that could have exposed flaws early.
- Your entire hypothesis hinged on one person, one lab, or one dataset.
- Your sample was underpowered. It held so few data points that a modest effect could stay hidden even if it was real.
- Nothing in your calendar could absorb a delay, so a single late reply pushed every later step back.
- Your hypothesis started as a hunch, before you’d read what earlier studies found.
Use the triage table next to judge how big your problem is.
What went wrong, and how serious is it?
It may feel like everything is on fire, but we highly recommend taking a couple of minutes to assess things calmly. Find your problem in the left column of the table below to see whether it usually calls for a restart. You’ll be relieved to know that in most cases, it doesn’t.
Read the second column as a general pattern. “Yes” means students can usually recover inside their current plan. “Often” means it depends on your deadline and on what you’ve already collected. The third column is the first thing to do today, before any bigger decision. The last column jumps to the full section for that problem.
If two rows fit, start with the one higher on the list. Results and data problems often explain the trouble further down. If you’re a parent reading this, this table can help you assist your child without taking over their project.
| Problem | Can you usually fix it without starting over? | Your first move | Jump to |
| Results came back null or surprising | Yes | Recheck the numbers, then treat the outcome as a finding | #1 |
| Method or equipment failed | Often | Change one thing per attempt and record each try | #2 |
| Too little data | Yes | Pick a path: collect longer, add public data, or call it a pilot | #3 |
| Lost your lab, dataset, or participants | Often | Take stock of what you hold, then search for open datasets | #4 |
| Mentor or teacher went quiet | Yes | Email a few yes-or-no questions with a reply-by date | #5 |
| Question is too wide or too tight | Yes | Restate your question in a single sentence | #6 |
| Rejected by a journal or competition | Yes | Split comments into required and optional changes | #7 |
Source: Horizon editorial guidance
7 things to do when your high school research paper goes off track
Skip straight to the problem that matches yours. There’s no need to read all seven. Each section closes with guidance on describing the problem in your paper. Reporting a setback accurately is a skill researchers use for their whole careers. Judges, editors, and teachers see derailed projects all the time, and they pay attention to how carefully you handled yours.
1. Your results are null or not what you expected
What it looks like: Maybe your two result groups came out about the same, or the effect ran in the direction you didn’t predict. What you have just experienced is called a null result, in which the data doesn’t show the effect you set out to test. Getting one doesn’t mean you made a mistake.
What to do this week: Start by checking your own work. Open the raw data, compare each value against your original records, and redo the statistics from scratch. If the numbers survive that check, sit down with your mentor and talk over what could explain them. Maybe the effect doesn’t exist. Maybe it’s real but too subtle for a sample your size to pick up. Or maybe your measurement tool couldn’t capture it.
How to write it up honestly: Describe the outcome in plain terms and leave your original hypothesis as written. In the discussion section, walk through those possible explanations. Never revise the hypothesis once you’ve seen the results, and never comb through the data looking for some other comparison that “worked.” That restraint is what makes your result useful. This article provides an in-depth explanation of the hypothesis writing and testing process.
2. Your experiment or method didn’t work
What it looks like: Sometimes it isn’t the logic but the equipment or the tool that fails. Perhaps your protocol doesn’t produce usable measurements, or your survey comes back with answers suggesting people misread the questions, or your painstakingly written code crashes on your actual data. A failed science experiment like this is frustrating, but the cause is usually in the method, and methods can be repaired.
What to do this week: Make additional attempts while changing one (and only one) thing per attempt. Record each try in your notebook: the date, what you adjusted, and the result. After a few days, you should have enough data to seek your mentor’s advice on whether to repair the method or move to a backup. This is doubly important if you are submitting your project to an ISEF-affiliated fair: when your project requires SRC or IRB approval (sign-off from a committee that reviews safety and ethics before research begins), the ISEF Rules for All Projects mandate any proposed change to your research plan to be approved again before you collect more data. You should go over this article to better understand when that approval applies.
How to write it up honestly: The methods section should match what you really did in the lab or the field. Add a short explanation naming the change and the reason for it. Readers trust a paper more when they can see how a problem was handled.
3. You don’t have enough data
What it looks like: Your survey drew 14 responses when you were counting on 100. Or you ran so few trials that you can’t tell a pattern from chance. Your sample size, meaning the number of people, samples, or trials in your study, is too small to support strong conclusions. Statistical power is how likely a study is to detect an effect that really exists, and small samples have low power. If terms like “statistical power” and “sample size” sound a little daunting and unfamiliar, we’ve got you covered with our Research terms glossary.
What to do this week: Pick the option that fits the time you have left:
- Keep collecting more data, if the calendar allows.
- Bring in an open, public dataset that addresses your question.
- Move to descriptive statistics (averages, ranges, and counts), which summarize data without testing for an effect.
- Recast the work as a pilot study, a small first run that checks whether a method works before anyone tries it at full scale.
How to write it up honestly: Give your sample size up front, in the abstract and again in the methods. Keep claims proportional to the data. List the small sample among your limitations and sketch the question a bigger study could answer.
4. You lost access to your lab, data, or participants
What it looks like: Sometimes your project plan gets hit by nothing but bad luck. It could be that your lab placement wrapped up earlier than planned, or your school didn’t approve the survey. Maybe a dataset you were reliant on moved behind a paywall. Or a partner who promised access went silent.
What to do this week: The first step is to take inventory: every bit of data, every note, every draft and source you’ve saved. Next, search for open datasets, or rework the question so you can answer it with secondary data, meaning information someone else already gathered. One powerful tool to help you with this is Google Scholar, and this article goes in-depth in explaining how to use it. Another avenue is to consider virtual research options. Some projects can also turn into a careful, organized review of published studies, and Original Research or a Literature Review compares the two routes. Make sure to read the rules for where you’ll be submitting your research before you change course though. ISEF’s rules, for example, list literature reviews as not appropriate for the fair.
How to write it up honestly: Describe the switch in your methods section. Spell out which data you gathered yourself and which came from another source, so no reader mistakes borrowed data for your own. A sentence or two is enough.
5. Your mentor or teacher stopped responding
What it looks like: Your emails have gone unanswered for a week or longer. Meetings get skipped, or the feedback that does arrive is too vague to act on.
What to do this week: Write one brief email that’s easy to answer. Ask two or three questions that need only a yes or no, and name the day you need a reply. Keep the tone friendly, since mentors often juggle many students. For example:
Hi Dr. Lee, my survey closed with 14 responses instead of 100. Could you answer three quick questions by Friday?
Can I add a public dataset? 2) Should I reframe this as a pilot study? 3) Can we meet for 15 minutes next week? Thanks, Maya
In the meantime, find someone else to read your draft: a science teacher, your school librarian, or a classmate with a sharp eye. You can read this post to better understand what your expectations from a research mentor should be, and tailor your emails and requests accordingly.
How to write it up honestly: Thank people in your acknowledgments for the work they really did. If your mentor approved the design but never saw the final draft, word the thank-you to match.
6. Your topic is too broad or too narrow, and now you’re behind
What it looks like: You’re midway through with 40 sources and still no clear argument. Or you’ve zoomed in so far that no data exists to answer your question. These feel like opposite problems, but both come from a question that doesn’t match the time and resources you have.
What to do this week: Restate your research question in a single sentence. If that’s hard, it’s still too broad. Drop any sub-question you won’t finish before the deadline. If the question is too tight, widen one part of it, such as the age group, location, or time period. Then pick a fresh finish date and plan backward from it. As this article details, you don’t really need as long as you might think to write a research paper, as long as your approach is optimized. In fact, you could do it well in as little as 90 days!
How to write it up honestly: In the paper, the narrower question simply is your question. That’s a normal part of research. Save the larger version for your future research section, where it shows readers you understand what comes next.
7. A journal or competition said no
What it looks like: A rejection email arrives, or a request to revise and resubmit, or your competition run ends after round one. Rejection happens to professional scientists all the time, and a request to revise is often good news.
What to do this week: Read every comment once, then wait a day before acting on it. Split the feedback into changes you must make and changes that would be nice. Handle the required ones first. You don’t need to stick to your original pick either, there are many more options available, both for journals and for competitions. If your research paper was written with a journal submission in mind, this article can help you adapt it into a competition submission. Thinking of sending one paper to several venues at the same time, or posting it as a preprint (a public version shared before peer review)? Read each journal’s policy first. There is a lot of learning to be had from the process of publishing independent research, whatever the outcome.
How to write it up honestly: With a resubmission, include a brief note that lists each change you made in response to reviewers. Keep it factual. How To Publish Research in High School explains how peer review works from start to finish.
Should you change your research topic or fix the one you have?
Usually, the better move is to repair the project you already have. Changing topics late in the process eats more time than most students plan for. A new topic means new background reading, a new method, possibly a new approval, and a fresh round of data collection, all squeezed into whatever weeks remain.
If you’re wondering, “Should I change my research question after bad results?”, narrowing it is usually the smarter move. A narrower question keeps the data and reading you already have. A brand new question throws most of that away. Use this table to check your situation. Parents can help here by asking which rows fit, then letting you decide.
| If this is true | Keep and fix | Consider switching |
| Part of your data is usable | Yes | No |
| You have nothing yet, and no way to collect data before the deadline | No | Yes |
| Your mentor sees a narrower version of the question | Yes | No |
| Your results are null but your method was sound | Yes | No |
| Your project needs approval or access you can’t get in time | No | Yes |
| The project works but your interest has faded | Yes | No |
Source: Horizon editorial guidance
If most of your rows land in the first column, stay with your project and use the seven sections above. If you do switch, take some time to understand how to choose a research topic that you can finish on time. Talk to your mentor before you commit to anything new, and keep your notes and data from the first attempt. Work you set aside now can become the starting point for a later project.
How do you plan your next project so this doesn’t happen again?
A research project contingency plan is a short list of backup steps you write before anything goes wrong. The students who recover fastest are usually the ones who planned for trouble in week one. These habits make the biggest difference:
- Test your method on a small scale during the first week.
- Name a backup method in your proposal, so a failed approach has a ready replacement.
- Get promises of lab space or datasets in writing, with names and dates.
- Schedule mentor meetings before the project starts.
- Create a structured schedule for your research, with a clear midpoint date to ask whether the question still fits your timeline.
- Understand how to draft a strong hypothesis. You should be clear on which results would support your hypothesis and which would count against it before the data even comes in.
- Leave two spare weeks before every deadline.
If your trouble is in the writing itself, such as a weak introduction or a messy discussion section, see Common Mistakes in High School Research Papers and How to Fix Them. Planning to extend this project next year? This blog will help you understand how continuation projects are reviewed.
Frequently asked questions
Your project isn’t ruined. Can a hypothesis be wrong? Yes, and that’s part of how science moves forward. Share what you found, leave the hypothesis as you first wrote it, and use the discussion section to explore why. However, spending some time and effort learning how to write and test your hypothesis goes a long way towards avoiding this issue.
Yes, if your method was sound and your write-up is accurate. Readers and judges care about how well you tested your question. A clear account of what went wrong, and why, shows careful thinking.
Tell them early, and put it in writing. Cover what happened, what you’ve tried, and two paths you’re weighing. Keep it short enough to read in under a minute.
Usually not, if you’re tightening the question instead of starting a new one. A narrower question lets you keep the work you’ve already done. The keep-or-switch table above can help you decide.
Often, yes. Many venues judge the quality of your method and reasoning, whichever direction your results went. Policies differ from one venue to the next, though. Read the author guidelines for every venue before you submit.
As a rough guide, spend one focused week on fixes, then talk with your mentor about a backup plan. If your deadline is under a month away, shorten that window to a few days. Treat this as guidance and adjust it to your timeline.
Yes. A mentor can help you weigh whether to repair the project, narrow it, or change direction, and can spot problems you’re too close to see. That outside view is often the fastest way to salvage a failed research project. Through Horizon, you work one-on-one with a PhD scholar or professor from schools such as Oxford, Stanford, and Columbia over several months of mentored research, and the program ends with a paper of about 20 pages.
Your project can still finish strong
Worried your project is too far off track to finish? Most aren’t. A setback you document clearly is part of research, and it often makes for a stronger paper. Pick your problem, take the first step this week, and keep a record of what you change. If you’d like a PhD mentor in your corner, apply to Horizon or join an info session.
Resources
External sources
- Rules for All Projects, Society for Science (ISEF)
- A Guide to Writing a Scientific Paper: A Focus on High School Through Graduate Level Student Research, PubMed Central
- JEI Author Guidelines, Journal of Emerging Investigators
More from Horizon
- What Is a Hypothesis? How to Write and Test One
- The Ultimate Guide to Writing a Research Paper in 90 Days
- How Long Does It Take to Write a Research Paper?
- 7 Steps to Turn Your Paper Into a Competition Submission
- What Does a Research Mentor Actually Do?
- How To Publish Research in High School
Image source: Horizon Academic Research Program