# Science Fair Research Plan Template

By the end of this article you will have a complete, judge-ready research plan for a science fair project. You will know what goes in each section, how a filled-in plan looks for a real experiment, how to set up a matching [lab notebook](/knowledge/molecular-biology/lab-notebook), and how to fix the mistakes that most often send plans back to students for revision.

What you need on hand before you start:

- A topic question you can actually test with materials available to you
- Access to a word processor (Google Docs, Microsoft Word, or LibreOffice Writer)
- A bound composition notebook or a printed lab notebook template PDF for your raw data
- Your school's or science fair's specific forms, if it requires them (many regional fairs use the forms published by the Society for Science for its affiliated fairs)
- About two hours of focused drafting time

A research plan is the document you write before you touch any equipment. It tells a reviewer what you will measure, why you expect a particular result, and how you will analyze the data. Most science fairs require one, and many require that it be approved before you begin hazardous work such as growing bacteria, working with vertebrate animals, or handling chemicals. Writing it well is the single highest-leverage step in the whole project, because a vague plan produces a project that cannot be judged fairly.

## Why a Written Plan Matters More Than Students Think

Structured pre-planning is a habit borrowed directly from professional science. In formal research, a [data management](/blog/guides/data-management-basics-principles-processes-and-best-practices) plan is a document describing how a researcher will collect, document, describe, share, and preserve the data generated by a project, and it is now a required part of most grant applications [1]. Scientific oversight bodies evaluate studies on their rigor and transparency, and a proactive written plan is the key element used to demonstrate both [2].

There is also a reporting reason to plan carefully. A systematic review of observational health studies found that heterogeneity in how researchers report their methods can hinder efforts to combine and [compare results across studies](/blog/research-skills/how-to-compare-results-across-multiple-studies-a-framework-for-evaluating-consistency), and the review classified reporting into defined areas such as methodological justification, measurement strategies, and the analytic plan [3]. When you fill in a research template with those same elements, your project becomes easier for judges to evaluate and easier for you to write up later.

Professional study planning also shows up in what researchers call pre-specification. A scoping review of target trial emulation studies in oncology, where investigators design observational analyses to mimic randomized trials, found that all 94 identified studies reported their treatment strategies, outcomes, and data analysis plans, but only about 75.5 percent clearly defined their time zero, the point at which follow-up begins [4]. Time zero is the science fair equivalent of deciding exactly when your experiment starts counting. If you skip that decision in your plan, your data become hard to interpret later.

## The Science Fair Research Plan Template Format

![Five-step flow showing the ten research plan sections grouped from question through data management and bibliography.](/article-images/illustrations/science-fair-research-plan-template-plan-sections-order.webp)

Most science fair research plan template formats share the same skeleton. The exact section names vary between fairs, but the content below covers every requirement you are likely to meet. Plan on 2 to 5 pages total for a middle school project and 5 to 8 pages for a senior high school project.

| Section | Purpose | Typical length |
|--|--|--|
| Title and contact block | Identifies the project and student | 3 lines |
| Question | States the testable question | 1 sentence |
| Background research | Shows the literature basis for the hypothesis | 2 to 4 paragraphs |
| Hypothesis | One falsifiable prediction with a direction | 1 sentence |
| Materials | Complete equipment and reagent list with quantities | Bulleted list |
| Variables | Independent, dependent, and controlled variables | Short table or list |
| Procedure | Numbered steps a stranger could repeat | Half to one page |
| Data analysis plan | How raw numbers become a conclusion | One paragraph plus formulas |
| Risk and safety assessment | Hazards, supervision, and [biosafety level](/knowledge/diagnostics/emerging-tech/biosafety-levels-1-4-a-comparative-framework-for-diagnostic-laboratory-design) if applicable | Short paragraph |
| Data management | Where data live, how they are backed up, what is shared | Short paragraph |
| Bibliography | Citable sources, at least 3 for most fairs | List |

### Question and Hypothesis

The question must be testable, which means it must be answerable with data you can collect. "Which fertilizer is best?" is not testable because "best" is undefined. "Does a 10-10-10 NPK fertilizer solution increase the mean height of radish seedlings more than water alone over 21 days?" is testable because it names the treatment, the organism, the measurement, and the time window.

The hypothesis is a one-sentence prediction with a direction. "Radish seedlings treated with a 10-10-10 fertilizer solution will show a greater mean increase in height over 21 days than seedlings treated with an equal volume of water." Notice the structure: treatment, comparison group, measurement, direction.

### Variables

Name all three types. The independent variable is what you deliberately change (fertilizer vs. water). The dependent variable is what you measure (height in centimeters). Controlled variables are everything you hold constant: seed variety, pot volume, soil type, light exposure, watering schedule, and temperature. Judges routinely probe students on controlled variables, so list at least five.

### Materials and Procedure

Write the procedure so that a classmate who has never seen your project could repeat it exactly. Use numbered steps in the imperative mood ("Fill each pot with 500 mL of potting mix" rather than "I filled pots"). Include replication in the procedure, not just in the analysis plan. Replication means repeating the measurement on multiple independent subjects. Ten plants per treatment group is a realistic and defensible number for a plant growth project, and it gives you enough data to compute both a mean and a standard deviation per group.

### Data Analysis Plan

![Decision path asking about counts, two groups and three or more groups to pick chi-square, t-test, ANOVA or descriptive statistics.](/article-images/illustrations/science-fair-research-plan-template-which-stat-test.webp)

This is the section most students leave nearly blank and most judges weight heavily. Your plan should state:

1. What you will calculate (means, standard deviations, growth rates)
2. What comparison or statistical test you will use, matched to your data type
3. What result would lead you to reject the hypothesis

For a two-group comparison of a continuous measurement, a common entry-level approach is an independent two-sample t-test, which compares the means of two groups while accounting for the spread within each group. State your significance threshold in the plan. An alpha level of 0.05 is the conventional default, meaning you accept a 5 percent risk of declaring a difference real when it is not.

For a three-group comparison, such as three light levels, a one-way analysis of variance (ANOVA) tests whether at least one group mean differs from the others. If your data are counts (number of seeds germinated), plan a chi-square test instead. If you are in middle school and statistics are beyond your coursework, a well-drawn graph of group means with error bars plus a plain-language comparison is acceptable, but say so explicitly in the plan.

### Risk and Safety Assessment

Every plan needs one, even for a low-risk project. Identify hazards (chemicals, heat, biological agents, electrical equipment, human or animal subjects), state your supervision arrangements, and name the safety equipment you will use. If your project involves microorganisms, most fairs will ask you to state the biosafety level, which is a four-tier classification of containment practices for biological work, with standard K-12 experiments using non-pathogenic strains falling at the lowest tier. Projects involving human participants or vertebrate animals almost always require prior approval and signed forms, so check your fair's rules before drafting this section.

## A Free Science Fair Research Plan Template You Can Copy

Here is a complete, free science fair research plan template in plain markdown. Paste it into any word processor and fill in the bracketed fields. This same text works as a generic research template for class labs, independent projects, and camp programs.

```markdown
# Research Plan

**Project title:** [Title, ideally a question or descriptive phrase]
**Student name:** [Full name, grade]
**School and teacher:** [School name, teacher name]
**Date submitted:** [YYYY-MM-DD]

## 1. Question
[One sentence ending in a question mark.]

## 2. Background Research
[2-4 paragraphs citing at least 3 sources. Explain the science
behind your question and cite each source in brackets, e.g. [5].]

## 3. Hypothesis
[One sentence: If [treatment], then [dependent variable] will
[direction], because [mechanism].]

## 4. Materials
- [Item, quantity, source]
- [Item, quantity, source]

## 5. Variables
- Independent variable: [What you change, with levels]
- Dependent variable: [What you measure, with units]
- Controlled variables: [At least 5, each with how it is held constant]

## 6. Procedure
1. [Step]
2. [Step]
...
n. [Include replication, timing, and measurement details]

## 7. Data Analysis Plan
- Raw data recorded: [variable, units, frequency]
- Calculations: [mean, standard deviation, growth rate formula]
- Statistical test: [t-test / ANOVA / chi-square / descriptive only]
- Significance level: [alpha = 0.05]
- Decision rule: [I will reject my hypothesis if ...]

## 8. Risk and Safety Assessment
- Hazards: [List]
- Supervision: [Who, where, when]
- Safety equipment: [Goggles, gloves, etc.]
- Special approvals needed: [Human subjects / vertebrate animals /
  microorganisms, with forms]

## 9. Data Management
- Data recording location: [Bound notebook pages / spreadsheet file]
- Backup: [Photo of notebook pages saved weekly to cloud drive]
- Sharing: [Data available to judges on request at fair day]

## 10. Bibliography
1. [Author. Title. Publisher or journal. Year. URL]
2. [...]
3. [...]
```

When this template is filled in correctly, the expected output is a document where every bracket has been replaced by concrete text, the procedure contains no first-person pronouns except where your fair requires them, and the analysis section contains at least one named calculation or test. If any section is still shorter than two sentences, it is not done.

## Numbered Walkthrough: Filling In the Template With a Real Example

This walkthrough builds a science fair research plan template example for a plant growth experiment. Follow the steps in order.

### Step 1: Draft the question and hypothesis

Question: "Does supplemental blue LED light increase the mean height of radish (Raphanus sativus) seedlings over 21 days compared with identical plants under white light?"

Hypothesis: "If radish seedlings receive 2 hours of supplemental blue LED light per day in addition to 6 hours of white light, then their mean height at day 21 will be greater than that of seedlings receiving white light alone, because blue wavelengths promote vegetative growth in seedlings."

### Step 2: List materials with quantities

```text
- Radish seeds (Raphanus sativus, Cherry Belle variety), 45 seeds
- Potting mix, 6 L (one bag)
- Plastic pots, 250 mL, 30 units
- Full-spectrum white LED grow light, 20 W, 1 unit
- Blue LED grow light, 20 W, 1 unit
- Ruler with millimeter markings, 1
- Spray bottle for watering, 2
- Timer outlets for lights, 2
```

### Step 3: Define variables and controls

```text
- Independent variable: light treatment (white only vs. white + blue)
- Dependent variable: seedling height in cm, measured daily
- Controlled variables:
  1. Seed variety and source (one packet, same lot)
  2. Potting mix (one bag, mixed before potting)
  3. Water volume (10 mL per pot per day)
  4. Temperature (20-24 °C, same room)
  5. Photoperiod (8 hours of total light per day via timer)
```

### Step 4: Write the procedure

```text
1. Fill each of 30 pots with 250 mL of potting mix.
2. Plant one radish seed 5 mm deep in each pot.
3. Randomly assign 15 pots to Group A (white light only) and
   15 pots to Group B (white light plus blue supplemental light).
   Use a random number generator to assign pot positions.
4. Place all pots on the same shelf. Water each pot with 10 mL
   daily at 7:00 AM.
5. Set timers so every pot receives 8 hours of light daily.
   Group B receives 6 hours white + 2 hours blue; Group A
   receives 8 hours white.
6. Starting at day 3, measure each seedling's height in
   millimeters from soil surface to the tallest cotyledon tip
   at the same time each day. Record values in the bound
   notebook, one table page per day.
7. Continue for 21 days. Mark the day 21 measurement as the
   endpoint for all plants (time zero for growth analysis is
   the first day every seedling in a pot has emerged).
8. Photograph each group weekly against a labeled ruler.
```

### Step 5: Compute a worked example of the analysis

Suppose day 21 data for Group A (white only) and Group B (white plus blue) look like this, in centimeters:

```text
Group A heights (n = 15): 8.2, 7.9, 9.1, 8.4, 8.8, 7.6, 9.0,
8.3, 8.1, 8.7, 8.0, 8.5, 8.9, 8.2, 8.6

Group B heights (n = 15): 9.4, 10.1, 9.8, 9.2, 10.4, 9.6,
9.9, 10.0, 9.3, 9.7, 10.2, 9.5, 9.9, 10.3, 9.8
```

Compute each group mean. For Group A, the sum of the 15 values is 128.3, so:

```text
mean A = 128.3 / 15 = 8.55 cm
mean B = 147.1 / 15 = 9.81 cm
difference = 9.81 - 8.55 = 1.26 cm
```

Also [compute the standard deviation](/blog/data-analysis/how-to-calculate-standard-deviation) for each group, which measures how widely individual plants vary around the group mean. A spreadsheet function such as `=STDEV.S(range)` computes it directly. Then run a two-sample t-test, for example with `=T.TEST(A1:A15, B1:B15, 2, 3)` in a spreadsheet, where the arguments are the two data ranges, a two-tailed test, and the assumption of unequal variances.

Expected output in your notebook:

```text
Group A: mean = 8.55 cm, SD = 0.46 cm, n = 15
Group B: mean = 9.81 cm, SD = 0.36 cm, n = 15
Difference in means = 1.26 cm
t-test two-tailed p-value = 2.0E-07 (p < 0.05)
Decision: reject the null hypothesis, support the hypothesis
```

In your plan itself, you would not have these numbers yet. You would instead write: "I will compute the mean and standard deviation of day 21 heights for each group and compare the groups with a two-tailed independent t-test at alpha = 0.05. I will reject my hypothesis if the supplemented group's mean height is not at least 1 cm greater than the control group's."

### Step 6: Write the risk assessment

Radish seeds and potting mix carry no special hazard in this design. The plan states: hazards are limited to low-voltage LED lighting, supervision is by the classroom teacher during watering and measurement, safety equipment is not required beyond standard laboratory conduct, and no special approvals are needed because the project involves no human subjects, vertebrate animals, or microorganisms.

### Step 7: Write the data management paragraph

This is where the plan borrows from professional practice. A data management plan describes how data will be collected, documented, shared, and preserved, and adopting those practices makes a project more transparent and reproducible [1]. For the radish project, the plan states that raw measurements go directly into a bound notebook on the day they are taken, notebook pages are photographed weekly and backed up to a cloud drive, and a summary spreadsheet with one row per plant per day is kept alongside the notebook. Professional research data stewardship covers exactly this span, from study design through collection, analysis, storage, and sharing, and dedicated data stewardship roles exist across research institutions to support reproducibility and transparency [5]. You will not have a data steward, but you can act like one for your own project.

## Setting Up the Matching Lab Notebook

Your research plan and your notebook are two halves of one system. The plan is the promise. The notebook is the proof you kept it.

A lab notebook template PDF is useful because you can print pages in advance and never scramble for structure on a measurement day. Set up the notebook like this:

1. Number every page in the top outer corner. Never remove pages.
2. Reserve page 1 for a title page and pages 2 and 3 for a table of contents you fill in as you go.
3. On page 4, glue or tape a printed copy of your approved research plan.
4. Create one data table per measurement day. Each table needs the date, the treatment group for each row, and one column per measured variable with units in the header.
5. Record errors and surprises in a dated "observations" section on each day's page. A seedling that falls over is data, not a mistake to hide.
6. Never erase. Draw a single line through an error, write the correction beside it, and initial it.

The connection between notebook habits and scientific quality is well documented in professional settings. Reviews of how studies are reported consistently find that clearly documented measurement and data handling strategies are a core component of transparent research reporting [3]. If a judge asks why one plant in your control group is short, your notebook's dated observation that the pot drained poorly is what turns an anomaly into evidence of careful work.

## Common Mistakes and Limitations

**Writing the question and procedure after collecting data.** If your fair requires plan approval, this can disqualify the project outright. Even where it does not, a retrospective plan cannot be trusted, and neither can the data.

**A hypothesis that cannot be wrong.** "Plants will grow differently with different lights" is unfalsifiable because any outcome fits it. Every hypothesis needs a stated alternative: what specific result would lead you to reject it.

**Missing replication.** One plant per treatment tells you about that plant, not about the treatment. Plan at least 10 independent subjects per group so you can compute a meaningful mean and standard deviation.

**Confusing repetition with replication.** Measuring the same plant ten times gives you ten repeated measures of one subject, which reduces your measurement error but does not tell you how plants vary. Replication means different subjects.

**No time zero.** Professional reviews found that a clearly defined starting point was the element most often missing even in mature research programs [4]. Define in your plan exactly when counting begins, for example "the first day a seedling emerges in each pot," rather than leaving it ambiguous.

**An analysis plan disconnected from the data type.** A t-test on counts of germinated seeds is the wrong tool. Counts call for a chi-square test or similar. If you are unsure which test fits, describe the data and your comparison goal in the plan and ask a teacher before collecting data.

**Ignoring human subjects rules.** Surveys, taste tests, and any project involving people require informed consent procedures and, at most fairs, prior review. Vertebrate animal work has additional oversight requirements. Check your fair's rules before you recruit a single participant.

**Data living in only one place.** A single notebook can be lost, and a single spreadsheet file can be deleted. Plan a weekly backup from day one. This mirrors the professional requirement that data plans address preservation and quality across the whole project lifecycle [1].

**A note on limits of this template.** The format here reflects common science fair requirements, but each fair publishes its own forms and rules, and affiliated fairs may require specific signed paperwork before any experimental work begins. Confirm your fair's current requirements directly. Also, a template cannot choose a good question for you. A plan built around a question you cannot test with available materials will be approved nowhere.

## How to Adapt the Template for Different Project Types

### Engineering and computer science projects

Replace "hypothesis" with a design goal and success criteria. State the requirement the design must meet, for example "a robot that completes the 2-meter course in under 15 seconds with at least 8 out of 10 successful runs." Your dependent variables become performance metrics and trial success rates. The analysis plan should state how many trials you will run and what improvement threshold counts as success.

### Observational and survey projects

You change nothing, so there is no independent variable. Instead, define your sampling plan: who or what you will observe, how many, how you will select them, and how you will avoid sampling bias, which is a systematic error in who or what gets observed. State your survey questions verbatim in the plan. Observational work benefits especially from clear pre-specified reporting, since heterogeneity in how observational studies define and measure their variables is a documented barrier to comparing results [3].

### Chemistry and microbiology projects

Add concentration, volume, and reagent source details to the materials list, and state the biosafety precautions in the risk section. For bacterial work with non-pathogenic strains, state that cultures are sealed, incubated at specified temperatures, and disposed of by the disposal method your fair or school requires. These projects most often require prior approval, so submit the plan early.

## How Judges Use Your Plan

Judges typically read your plan before or during your presentation, and they use it to check three things: whether your method can answer your question, whether you controlled the variables that could have confounded the result, and whether your conclusion follows from your data. Scientific evaluation in professional contexts centers on exactly this kind of assessment of rigor and transparency, performed by independent reviewers using a documented plan [2]. When you can point to a line in your plan and the matching pages in your notebook, you demonstrate the same chain of accountability.

One practical tip: bring both documents to the fair. A plan that matches a complete notebook is the strongest physical evidence a judge can hold that your work is yours and that it was done properly.

## Frequently Asked Questions

### How long should a science fair research plan be?

Two to five pages for middle school and five to eight pages for senior high school projects, including the bibliography. Length matters less than completeness, so a tight two-page plan with all ten sections beats a padded eight-page one.

### Where can I find a free science fair research plan template?

Copy the free template in this article directly into Google Docs or Word, or download the forms your affiliated fair publishes, since those are free and tailored to that fair's exact rules. Many schools also keep approved templates from prior years.

### What is the difference between a research plan and a research paper?

The research plan is written before the experiment and predicts what you will do, while the research paper is written after and reports what you found. The plan contains your hypothesis and procedure, and the paper contains your results and conclusions.

### Do I need approval for my research plan before starting?

Most science fairs require plan approval before experimental work begins, and approval is always required before projects involving human participants, vertebrate animals, or potentially hazardous biological or chemical agents. Check your fair's rules and submit early, because review can take weeks.

### What statistical test should a middle school student use?

Descriptive statistics are acceptable at that level. Compute the mean and standard deviation for each group, graph the means with error bars, and state in plain language whether the groups differ by more than the spread within groups. Say in your plan that this is the analysis you will do.

### How many sources do I need in my bibliography?

At least three is the common minimum for most fairs, though more is better for senior projects. Use primary sources where possible, meaning original research articles or authoritative reference material, and cite each source within your background section, not just at the end.

### What is a controlled variable and how many do I need?

A controlled variable is any factor you hold constant across all groups so it cannot explain your results, such as soil type, watering volume, and temperature in a plant experiment. There is no official minimum, but listing five or more with the method for holding each constant is standard practice and what judges expect.

### Can I change my procedure after my plan is approved?

Small procedural adjustments are normal and should be documented in your notebook with a dated note, but substantive changes to the question, treatments, or safety profile usually require a revised plan and re-approval. When in doubt, ask your fair's review committee before changing anything.

## Sources

1. [Data Management Plans in the genomics research revolution of Africa: Challenges and recommendations.](https://pubmed.ncbi.nlm.nih.gov/34506960/)
2. [Managing the science in the presence of financial conflict of interest.](https://pubmed.ncbi.nlm.nih.gov/31349765/)
3. [A systematic review of research reporting practices in observational studies examining associations between 24-h movement behaviors and indicators of health using compositional data analysis.](https://pubmed.ncbi.nlm.nih.gov/39371105/)
4. [Scoping review: Investigating target trial emulation approaches in oncology research.](https://pubmed.ncbi.nlm.nih.gov/41762536/)
5. [Data stewardship in Italy: Findings from the first national survey.](https://pubmed.ncbi.nlm.nih.gov/41816254/)