Why this is hard to get right
A Science Teacher Races the Clock Before Lab Day
Maria teaches 7th-grade Earth Science at a public middle school with 28 students per class. She's running a chemistry unit and needs a baking soda and vinegar experiment guide ready before Friday's lab session — three days away. She also has two IEP students who need simplified instructions and a class that has never used lab equipment before.
She starts the way most teachers do: she searches Google for a pre-made worksheet. She finds three options. One is too advanced — it assumes students understand molarity. Another skips safety notes entirely, which her department head requires in writing. The third is formatted for a demonstration, not a hands-on lab.
She decides to write the guide herself. She spends 45 minutes drafting it. The instructions are clear enough, but when she re-reads them she realizes she forgot the reflection questions her curriculum coordinator expects. She adds those, then notices the materials list references equipment her school doesn't stock.
She rewrites the materials section. Now the word count is too long for a single page. She cuts it down, but loses the learning objectives in the trim. At this point she's spent over an hour and still doesn't have a finished guide.
The core problem isn't skill — it's structure. Writing an experiment guide requires you to hold at least six things in your head simultaneously: learning objectives, safety constraints, materials, step-by-step instructions, appropriate reading level, and assessment questions. Missing any one of these forces a rewrite.
When Maria finally tries using an AI assistant, she types: "Write a science experiment for my class." The output is a volcano model — a demonstration activity, not a hands-on student experiment. It includes no safety notes, no reflection questions, and assumes access to craft supplies she doesn't have.
The issue isn't the AI. It's the prompt. An AI without context guesses. It picks the most common, most generic experiment it can find in its training data and produces something that looks complete but fits no specific classroom.
When Maria builds a structured prompt — specifying grade level, the exact scientific concept, student experience, required sections, word count, and tone — the output changes completely. The AI produces a guide with numbered steps written at a 7th-grade reading level, a clear materials list using household items, two safety callouts, and three reflection questions that connect to the unit's learning objectives.
She reviews it in under five minutes and it's ready to print. The difference isn't the AI's capability. It's the information she gave it. A structured prompt is the difference between a generic document and a classroom-ready resource that saves her an hour every time she uses it.
Common mistakes to avoid
Omitting the Student Experience Level
Without specifying prior lab experience, the AI defaults to assumptions — often writing for students who already know basic lab safety and terminology. This produces guides that confuse novice students or bore advanced ones. Always state whether students have zero, some, or extensive lab experience so vocabulary, pacing, and step complexity match your actual class.
Skipping Safety and Equipment Constraints
Science guides without safety callouts are unusable in most schools. If you don't specify your environment — no open flames, no glass beakers, household materials only — the AI may include steps or materials you can't use. Name your constraints explicitly. The AI won't assume them, and rewriting safety sections after the fact wastes time.
Requesting a Guide Without Listing Required Sections
An experiment guide is a structured document with defined components: objectives, materials, safety notes, steps, and reflection questions. If you don't list these sections, the AI often omits one or more. Most commonly it skips objectives and reflection questions — the two sections most likely to satisfy curriculum standards and department requirements.
Using a Topic That's Too Broad
Prompts like 'a chemistry experiment' give the AI too much latitude. It may return a demonstration, a thought experiment, or a multi-day project when you need a 30-minute in-class activity. Narrow the topic to a specific concept and a specific reaction or process — for example, 'the effect of temperature on dissolving rate' rather than just 'chemistry.'
Forgetting to Specify Format and Length
Without a word count or format requirement, AI-generated experiment guides often run too long, use paragraph-style instructions instead of numbered steps, or include headers that don't match your curriculum template. State your preferred format — numbered steps, bullet lists, table — and a maximum word count so the output fits your worksheet or handout layout.
Not Defining the Learning Objective
A guide without a tied learning objective is an activity, not a lesson. If you don't tell the AI what students should be able to explain or demonstrate after the experiment, the output won't reinforce your unit's goals. Provide one concrete learning objective — for example, 'students will explain why a gas is produced during an acid-base reaction' — and the guide will align to it automatically.
The transformation
Write a science experiment students can do in class.
**Role:** Act as a middle school science educator. **Task:** Create a step-by-step experiment guide on chemical reactions using baking soda and vinegar. **Audience:** 7th-grade students with limited lab experience. **Requirements:** 1. Include learning objectives, materials list, and safety notes. 2. Write simple, numbered instructions. 3. Add 3 reflection questions. 4. Keep the guide under 250 words and use a friendly, instructional tone.
Why this works
Role Anchors Voice and Accuracy
The After Prompt opens with "Act as a middle school science educator." This single instruction shifts the AI's register — it writes with instructional intent, uses classroom-appropriate language, and frames safety as a teacher would. Without a role, the AI writes as a general assistant, which produces inconsistent tone and missed pedagogical cues.
Specificity Eliminates Guessing
The After Prompt names the exact concept — "chemical reactions using baking soda and vinegar" — and the exact audience — "7th-grade students with limited lab experience." This removes ambiguity entirely. The AI doesn't choose a topic, guess a grade level, or assume prior knowledge. Every output decision flows from defined parameters rather than statistical defaults.
Section Requirements Ensure Completeness
The numbered requirements in the After Prompt — learning objectives, materials list, safety notes, numbered instructions, and reflection questions — function as a structural checklist. The AI treats each item as a required output element. This prevents the most common failure mode: a guide that reads well but is missing a section your curriculum or department requires.
Constraints Shape Usability
"Keep the guide under 250 words" and "use a friendly, instructional tone" are constraints that directly affect classroom usability. The word count keeps the guide to one printable page. The tone instruction ensures instructions don't read as dry technical procedures — they read as guidance a student can follow confidently. Constraints turn generic output into a ready-to-use document.
Reflection Questions Elevate Learning Depth
The After Prompt explicitly requests "3 reflection questions." This transforms the guide from a procedural activity into a learning experience. Reflection questions prompt students to connect observations to concepts, which is what separates a science experiment from a science demonstration. Specifying the number prevents the AI from adding too many or skipping them altogether.
The framework behind the prompt
The Science Behind Science Experiment Guides
Writing a science experiment guide is a document design problem as much as a content problem. Every section — objectives, materials, procedure, safety, reflection — serves a distinct cognitive function for the student. Remove one section and the guide loses structural integrity. Understanding why each section exists helps you prompt for it correctly.
Learning objectives anchor the experiment to a transferable concept. Without one, students experience the experiment as a novelty rather than an investigation. In instructional design terms, objectives operationalize Bloom's Taxonomy: they define the expected cognitive level (recall, application, analysis) so that both the experiment and its assessment questions align to a single target.
Procedural steps work through Cognitive Load Theory, developed by John Sweller in the 1980s. Novice learners have limited working memory. Numbered steps, one action per step, offload cognitive sequencing to the document — freeing the student's attention for observation and thinking. This is why format specificity in your prompt matters: prose instructions force students to parse sequence themselves, adding unnecessary load.
Safety notes are not just liability protection. Research in science education shows that explicit safety instruction before an activity reduces off-task behavior during it. When students know what to expect, they're less distracted by uncertainty.
Reflection questions activate metacognitive processing — the practice of thinking about one's own thinking. The NGSS (Next Generation Science Standards) framework specifically requires sense-making activities after investigations, not just data collection. Reflection questions that ask students to explain why, rather than what, push toward the deeper end of Bloom's Taxonomy and improve long-term retention.
Understanding this structure helps you write prompts that don't just produce readable guides — they produce pedagogically sound ones.
Prompt variations
Role: Act as an elementary school science teacher with experience in hands-on learning for young students.
Task: Create a step-by-step experiment guide exploring how plants absorb water using celery stalks and food coloring.
Audience: 3rd-grade students with no prior lab experience.
Requirements:
- Write a single learning objective in one sentence.
- List materials using only household or grocery-store items.
- Include two safety notes written in simple language.
- Write 5 numbered steps using short sentences at a 2nd-grade reading level.
- Add 2 observation prompts students can answer by drawing or writing.
- Keep the total guide under 200 words.
Role: Act as a high school chemistry teacher preparing a formal lab investigation.
Task: Write a structured lab guide for an experiment measuring the rate of reaction between hydrogen peroxide and potassium iodide at different concentrations.
Audience: 10th-grade students who have completed one semester of chemistry and are familiar with basic lab safety protocols.
Requirements:
- State two measurable learning objectives aligned to reaction rate concepts.
- Include a materials list with quantities and concentrations.
- Write a formal safety section covering chemical handling and disposal.
- Provide 8 numbered procedure steps using precise scientific language.
- Add a data table template students will complete during the experiment.
- Include 4 analysis questions that require students to interpret results and connect them to collision theory.
- Keep the guide under 400 words.
Role: Act as a STEM program facilitator running an informal after-school workshop.
Task: Design a 20-minute experiment guide exploring static electricity using a balloon, wool fabric, and small paper pieces.
Audience: Mixed-age group of students ages 8 to 12 with varying science backgrounds. Some students may have learning differences.
Requirements:
- Write one simple learning goal in plain language.
- List only materials available at a dollar store or grocery store.
- Include one safety reminder.
- Write 6 numbered steps using plain, encouraging language accessible to a 3rd-grade reading level.
- Add a "What did you notice?" section with 2 open-ended questions.
- Include one fun extension challenge for students who finish early.
- Keep the total guide under 220 words and use a warm, energetic tone.
Role: Act as a homeschool science curriculum designer.
Task: Create a kitchen-science experiment guide exploring density using water, oil, and honey in a clear glass.
Audience: A 9-year-old child working independently at home with light parent supervision. No specialized equipment available.
Requirements:
- Begin with one learning objective written in child-friendly language.
- List only materials found in a typical home kitchen.
- Include a brief parent note with one supervision tip.
- Write 7 numbered steps the child can follow independently.
- Add a "Think about it" section with 2 questions connecting the observation to the concept of density.
- Suggest one way to extend the experiment using a different liquid.
- Use a conversational, encouraging tone and keep the guide under 250 words.
When to use this prompt
Science Teachers
Create consistent experiment guides that match grade-level expectations and support hands-on learning across units.
Curriculum Designers
Develop supplemental experiment activities that align with objectives and support inquiry-based instruction.
STEM Program Leaders
Build repeatable experiment templates for after-school programs or workshops with mixed-age groups.
Homeschool Educators
Produce structured, easy-to-follow science activities tailored to different ages and experience levels.
Pro tips
- 1
Specify the exact scientific concept you want students to explore so instructions stay focused.
- 2
Define your students’ experience level to ensure the AI selects appropriate vocabulary and pacing.
- 3
Clarify safety constraints to avoid steps or materials that don’t fit your environment.
- 4
State your preferred format so the guide fits your teaching style or curriculum needs.
A single experiment guide can serve multiple ability levels if you build differentiation directly into your prompt. Use this technique: add a requirement that says "Include a scaffolded extension section with three tiers: a basic observation task, an intermediate analysis question, and an advanced design challenge."
This mirrors Universal Design for Learning (UDL) principles, which recommend providing multiple means of engagement and expression. Instead of writing three separate guides, you build one guide with three entry points. Students who struggle complete the observation task. Students who are on pace complete the analysis question. Students who are advanced take on the design challenge — for example, proposing a modified experiment to test a new variable.
A second advanced technique is "anticipatory set" language. Add this to your prompt: "Open the guide with one surprising question or interesting fact that connects the experiment to a real-world phenomenon." This gives students a reason to engage before they even pick up a material. Research in science education consistently shows that activating prior curiosity improves both attention and retention during procedural activities.
Finally, if your school uses lab notebooks, add: "Include a lab notebook prompt at the start — a space for students to record their hypothesis before beginning." This turns the guide into a complete learning document, not just a set of instructions.
Science experiment guides built for print don't always translate to digital or flipped classroom formats. Here's how to adapt the same prompt structure for those contexts.
For a flipped classroom, add this requirement: "Include a 'Before Class' section with one video recommendation topic and two pre-reading questions students answer at home. Label the in-class section separately." This structures the guide across both the asynchronous and synchronous phases of instruction without requiring you to write two separate documents.
For digital submission, request: "Format all student response sections as clearly labeled text boxes with instructional prompts inside — for example, 'Type your hypothesis here.' Include a word count target for each response." This makes the guide compatible with Google Docs or any PDF-with-form-fields tool your district uses.
For recorded demonstration experiments, where students observe rather than conduct: "Reframe all procedural steps as observation prompts. Replace 'Add one teaspoon of baking soda' with 'Notice what happens when the teacher adds baking soda. Record your observation below.'" This small language shift changes the guide's entire function — from a hands-on procedure to an active observation protocol — without changing the underlying structure.
If you teach multiple science units per year, you don't need to rebuild your prompt each time. Create a reusable prompt template by abstracting the specific concept, grade level, and materials into variable lines you swap out.
The stable structure stays the same every time:
- Role: act as a [grade level] science educator.
- Task: create a step-by-step experiment guide on [concept].
- Audience: [grade and experience level].
- Requirements: learning objective, materials list, safety notes, numbered steps, reflection questions.
- Constraints: under [word count] words, [format], [tone].
The only lines you update per unit are the concept, the grade level, the materials constraints, and any safety specifics tied to that unit. This approach turns prompt-building from a creative task into a 3-minute editing task.
For curriculum designers working across multiple teachers, this template approach also supports consistency across classrooms. When every teacher uses the same base prompt structure, the resulting guides share a recognizable format, which reduces cognitive load for students who move between teachers or review past experiments. Consistency in document structure is an underrated learning support — students learn the format once and use that schema to navigate every guide they receive for the rest of the year.
When not to use this prompt
When This Prompt Pattern Doesn't Fit
This structured experiment guide prompt works best for single-session, procedure-driven activities with a defined learning objective. There are contexts where a different approach serves you better.
Don't use this format for open-ended inquiry projects. If you want students to design their own investigation — choose their own variable, write their own procedure — a prescriptive step-by-step guide works against the goal. Instead, prompt for an inquiry framework that gives students a question, a set of available materials, and guiding constraints without writing the steps for them.
Don't use it for demonstrations you'll perform while students observe. The format assumes student action. A demonstration guide needs a different structure: presenter cues, talking points, anticipated student questions, and timing markers.
Don't use it when your district requires a specific template. If your school has a mandatory lab report format with defined fields, prompting for a free-form guide will produce something you have to reformat anyway. In that case, paste your district's template into the prompt and ask the AI to populate it — a different task entirely.
Don't use it as a substitute for expert review of high-risk experiments. AI-generated safety notes are a starting point, not a safety certification. Any experiment involving heat, electricity, strong acids or bases, or equipment with injury potential requires review by a qualified science educator before classroom use.
Troubleshooting
The AI writes a demonstration, not a student-led experiment
Add this line explicitly: "Write this as a hands-on student experiment — every step should be performed by the student, not observed from a teacher demonstration." Also specify: "Use second-person instructions — 'You will add...' rather than 'The teacher will show...'" This distinction in voice directly controls who performs each action in the output.
The safety section is either missing or too vague to use
Make safety a numbered requirement, not a general request. Write: "Include exactly 2 safety notes. Each must name a specific hazard and a specific action — for example, 'Vinegar may irritate eyes. If splashed, rinse immediately with water for 15 seconds.'" Generic prompts like 'include safety tips' produce generic output like 'be careful with materials.'
The reflection questions are too easy or don't connect to the learning objective
Tie your reflection questions directly to the objective. Add: "Write 3 reflection questions that require students to explain why the reaction occurred, not just what happened." You can also specify Bloom's taxonomy levels: "Include one recall question, one application question, and one question that asks students to predict what would change if one variable were different."
The guide is too long to fit on one printed page
Set a strict word cap and add a format instruction: "Keep the total guide under 250 words. Format steps as single sentences. List materials in two columns to save space." If the output still runs long, follow up with: "Reduce each procedural step to 15 words or fewer without removing any essential action."
The vocabulary is too advanced for the grade level specified
Reinforce the vocabulary constraint with a reading level benchmark: "Write all procedural steps at a 6th-grade reading level. Avoid words with more than 3 syllables unless you define them inline immediately after first use." You can also add: "After drafting the guide, flag any technical term and replace it with plain language if it hasn't been defined."
How to measure success
How to Evaluate Your AI-Generated Experiment Guide
A strong experiment guide passes a simple structural and usability check. Run through this evaluation before you print or distribute it.
Structural completeness:
- Learning objective is present and uses measurable language ("students will explain..." not "students will learn...")
- Materials list names every item needed with quantities
- Safety notes identify a specific hazard and a specific response action
- Steps are numbered and each contains one action only
- Reflection questions require explanation, not just observation recall
Usability check:
- Word count fits your format — a single worksheet typically requires under 300 words
- Reading level matches your students — try the Flesch-Kincaid grade level tool if unsure
- Materials are all available in your actual setting — cross-check your supply list
- Steps can be followed in sequence without referring back to earlier sections
Learning alignment:
- The reflection questions connect directly to the stated learning objective
- The experiment produces an observable result that students can link to the concept
- No step requires knowledge students don't yet have at this point in the unit
If any item fails, add the missing element as a specific follow-up instruction rather than regenerating the full guide from scratch.
Now try it on something of your own
Reading about the framework is one thing. Watching it sharpen your own prompt is another — takes 90 seconds, no signup.
Build a classroom-ready experiment guide tailored to your grade level, topic, and safety constraints — in minutes.
Try one of these
Frequently asked questions
Add a specific line to your prompt requirements such as: "Include a simplified instructions version written at a 2nd-grade reading level for students who need additional support." You can also request that key vocabulary words be bolded and briefly defined inline. Naming the accommodation directly is more effective than general phrases like 'make it accessible.'
Yes. Add a constraint like: "All materials must be available in a standard home kitchen — no lab equipment." You can also specify: "Design this for a student working independently without adult supervision." These two additions shift every step, material choice, and safety note to fit an at-home context without requiring a full rewrite of your prompt.
This almost always means the safety and materials constraints weren't explicit enough. Add a dedicated line listing prohibited items — for example: "Do not include open flames, glass containers, or chemicals not available at a grocery store." Being specific about what you cannot use is as important as listing what you can.
Name the standard directly in your prompt. For example: "Align this experiment to NGSS standard MS-PS1-2, which requires students to analyze and interpret data on the properties of substances." The AI will shape the learning objective, reflection questions, and procedure language to support that standard. Vague phrases like 'curriculum-aligned' produce no measurable alignment.
Yes. Add this to your requirements: "Generate two versions — a student-facing guide and a separate teacher guide that includes expected observations, sample answers to reflection questions, and one common misconception to address during debrief." This doubles the output's value without doubling your work.
Specify the timeline explicitly: "Structure this as a 3-day experiment. Day 1: setup and initial observations. Day 2: data collection. Day 3: analysis and reflection." Break your requirements list into sections matching each day. The AI will organize steps and questions across the timeline rather than compressing everything into a single procedure.
Set a hard word limit in your prompt — "Keep the total guide under 250 words" — and add: "Format for a single printed page with standard margins." If output still runs long, follow up with: **"Condense the materials list to a single line per item and reduce each step to one sentence maximum."
Refresh the prompt whenever your class context changes — new grade level, new unit topic, different experience level, or updated safety policies. A well-structured prompt is reusable across topics. Simply swap the concept, adjust the audience line, and update any materials or safety constraints. The structure itself doesn't need rebuilding.