How to Conduct Experiments
Drop a Mentos into soda and a foamy geyser erupts - but some sodas erupt far higher than others. Anyone can watch it happen. A scientist tests it fairly and lets the evidence decide. This lesson teaches you the scientist's toolkit.
What You'll Be Able to Do
By the end of this lesson, you will be able to:
- Prepare students for the upcoming melting-ice-on-four-surfaces investigation by naming the tools they will use there.
- Set honest scope: students assemble a hypothesis, sort variables, classify observations, practice inference, and analyze supplied data. They do not design or run an experiment, collect measurements, or generate the soda data here.
- Goal setting orients attention toward what to look for
- Advance organizer for the hypothesis, variables, and observation work ahead
- Understand to Apply
- DOK 1 to 3 (most goals name and classify; the supplied-data goal asks for a claim backed by evidence)
- Five short goal cards, one idea each
- Practice standards tagged for transparency
- Plain "be able to do" phrasing
Words You'll Meet
Choose a card to see what each word means.
- Pre-teach the ten terms students meet again in context below.
- Give a definition to return to when a word reappears.
- Pre-teaching vocabulary lowers load during the explore tasks
- Collapsed cards reduce extraneous load until a word is needed
- Remember to Understand
- DOK 1
- One card open at a time
- Click to reveal, no hover
- Short, plain definitions
Three Everyday Mysteries
Science doesn't start in a lab. It starts the moment you notice something and ask, "Wait... why does that happen?" Click each card and see what questions pop into your head.
- Open with real phenomena so a question, not a definition, starts the inquiry.
- Show that every investigation begins with noticing something puzzling.
- Curiosity gap primes attention before any Explore instruction
- Phenomenon-based learning anchors the abstract method in concrete events
- Understand
- DOK 2 (students generate questions from an observation)
- Click to reveal each mystery, no hover
- Icon paired with short scenario text
- No prior knowledge required to engage
What Is a Hypothesis?
Before scientists test anything, they make a prediction: what they think will happen. When they add why they think so, the prediction becomes a hypothesis.
A hypothesis is an educated guess that explains an observation based on your prior knowledge. A good hypothesis is testable and explains the cause-and-effect relationship between variables.
Choose one
Your turn - no drop-downs. Think back to the Thirsty Plant mystery. Write your own if - then - because hypothesis about how the amount of water affects how a plant grows.
- Define a hypothesis as a testable, reasoned prediction, not a random guess.
- Scaffold construction, not independent authorship: students assemble a hypothesis from drop-down choices, then may try one in their own words.
- Worked examples and a three-part frame scaffold the structure
- Cause-and-effect modeling links the "if" to the "then"
- Misconception checking: testable versus untestable cards counter "a hypothesis is just a guess"
- Understand to Apply
- DOK 2 (students assemble and check a hypothesis, not justify a full design)
- Drop-down builder first; the own-words step is optional
- Predict gate before the builder, with immediate feedback
- Key term defined in place before use
What Are the Variables?
Every fair experiment has three kinds of factors. Follow the color coding through the rest of the lesson: teal is what you change, orange is what responds, and green is what stays the same.
Choose one
If one plant got more sunlight and more water, you could never tell which one caused it to grow taller. Keeping everything else constant means the independent variable is the only possible cause of the change you measure. That's what makes a test fair.
- Separate the three variable roles and tie each to a color used lesson-wide.
- Have students choose why holding controls constant is what makes a test fair.
- Comparison and contrast distinguishes independent, dependent, and control roles
- Error analysis: the predict gate exposes why an uncontrolled factor ruins the test
- Misconception checking counters "changing more than one variable is okay"
- Understand to Analyze
- DOK 2 (sorting roles and choosing why a control keeps the test fair)
- Consistent color coding for the three roles
- Sorter gives item-level feedback on each choice
- Short, parallel role cards
Two Types of Observations
Scientists collect evidence in two flavors: with instruments and numbers, or with their senses and descriptions. Both are useful - but they're not the same.
Quantitative sounds like quantity - a number you can count or measure. Qualitative sounds like quality - a description from your senses.
- Distinguish quantitative from qualitative evidence before students collect their own data in a later hands-on investigation.
- Practice classifying sample lab-notebook entries, not just reading the definitions.
- Comparison and contrast pairs numbers against sensory descriptions
- The quantity / quality mnemonic supports retrieval
- Generation effect: students classify entries themselves
- Understand to Analyze
- DOK 2 (classifying each observation by type)
- Two-way sort keeps choices simple
- Icon and short quote for every item
- Immediate feedback per entry
What Is an Inference?
Detectives don't see the crime - they see the clues and work out what happened. Scientists do exactly the same thing with observations.
An inference is an interpretation you make based on your observations. You didn't see it happen - your brain filled in the most likely explanation.
Choose one
You can infer that they think it is going to rain. You never observed rain - you observed a raincoat and umbrella, and your brain connected the clues to the most likely explanation. That's an inference: an explanation built from observations.
Choose one
You can infer that it rained overnight or dew formed on the grass. Notice that the dry sidewalk is also an observation - it's a clue that points toward dew rather than rain. Better observations lead to better inferences.
A prediction is what you think will happen before testing. Observations and measurements are what actually happened. Evidence is the observations and measurements you use to support a claim. An inference explains what observations mean. A conclusion is the answer the evidence supports once the results are in - and it can only be as strong as that evidence.
- Draw the line between what you observe and the conclusion you build from it.
- Show that stronger observations lead to stronger inferences.
- Evidence-based reasoning connects clues to a likely explanation
- Detective analogy maps inference onto a familiar schema
- Callback to the opening wet-grass mystery supports elaboration
- Understand to Analyze
- DOK 2 (drawing and distinguishing inferences from given observations)
- Predict gate before each reveal
- Observation and inference shown side by side
- Concrete everyday examples first
Brain Check
Two quick questions before we put it all together. These are not graded. Pulling answers from memory now will help them stick.
- Force memory retrieval before the synthesis and quiz, while stakes are zero.
- Surface a wrong variable or observation/inference choice early enough to fix.
- Retrieval practice strengthens memory more than rereading
- Items apply ideas to a fresh Mentos scenario, not the exact examples taught
- Try-again loop turns an error into productive struggle
- Remember to Apply
- DOK 1 to 2 (kept at or below the formal quiz)
- Ungraded and low stakes, stated up front
- Retry button with feedback
- One short question at a time
Back to the Eruption
You started with a foamy mystery. Now you have the toolkit a scientist uses to investigate it. At the end of the quiz, you will use that toolkit to analyze results from a Mentos-and-soda investigation.
Everything in One Place
The words to know and the goals you are working toward, all in one spot.
| Term | Student-Friendly Definition |
|---|---|
| Hypothesis | An educated guess that explains an observation based on prior knowledge. A good one is testable and explains the cause-and-effect relationship between variables. |
| Independent variable | The factor in the experiment that you manipulate or change. |
| Dependent variable | The factor in the experiment that responds to the change - the thing you measure. |
| Control variables | The other factors kept constant to ensure any change measured was due to the independent variable. |
| Quantitative observation | An observation using precise measurements and numerical quantities: length, height, temperature, time. |
| Qualitative observation | An observation using sensory information to describe qualities: colors, textures, smells, tastes. |
| Prediction | What you think will happen, before testing. |
| Inference | An interpretation you make based on your observations. |
| Evidence | The observations and measurements you use to support a claim. |
| Conclusion | The answer the evidence supports, decided after the results are in. |
| Learning Goals | How You Showed It |
|---|---|
| Build a testable hypothesis using the if - then - because frame. | You used the Hypothesis Builder to assemble an if - then - because hypothesis for the Mentos question, checked whether its parts fit together, and could try one in your own words. |
| Identify independent, dependent, and control variables. | You sorted every factor in the plant experiment into its correct role and chose why control variables make a test fair. |
| Classify observations and draw inferences. | You sorted sample lab-notebook entries into quantitative and qualitative, and drew reasonable inferences from the raincoat and wet-grass clues. |
| Use supplied results to support a claim. | At the end of the quiz, you will analyze results from a Mentos-and-soda investigation and write a claim, evidence, and reasoning. |
- Tie question, hypothesis, fair test, observation, and inference into one chain.
- Map every learning goal to the evidence the student produced.
- Schema building connects the separate pieces into one method
- Coherent narrative returns to the opening eruption to close the loop
- Summary tables consolidate terms and goals in one place
- Understand to Analyze
- DOK 2 to 3 (relating the full inquiry sequence as a connected process)
- Three labeled beats break the method into steps
- Scrollable tables for term and goal review
- Essential question restated for self-check
Check Your Understanding
Ten questions covering everything in this lesson, including a brand-new experiment for you to analyze. Answer every question, then submit.
Scientists don't just know the answer. They explain their thinking.
Write your own explanation first. Then submit your work to compare your thinking with a model answer.
You did not collect these results. A science club ran this test and shared their measurements for you to analyze.
- Question: Which type of soda makes the tallest Mentos eruption?
- Changed (independent variable): the type of soda - diet cola, regular cola, or flat cola (regular cola left open overnight until it stopped fizzing, so most of its dissolved gas had escaped).
- Measured (dependent variable): eruption height - how high the foam rose above the bottle opening, measured with a meter stick against a wall.
- Kept the same (controls): same brand, same 2-liter bottle size, same room temperature, 5 Mentos dropped all at once with the same tube, and the same person measuring.
- Trials: three separate bottles of each soda.
| Soda | Eruption height (trials 1, 2, 3) | What the club noticed |
|---|---|---|
| Diet cola | 250, 230, 260 cm | A tall, fast column of foam |
| Regular cola | 90, 110, 80 cm | A shorter column of foam |
| Flat cola | 0, 0, 0 cm | A few bubbles; no foam left the bottle |
Heights are rounded to the nearest 10 cm.
Use only the supplied results above. Claim: Which soda do these results show made the tallest eruption? Evidence: Which specific measurements support your claim? Reasoning: Why do those measurements support it? Explain what the club kept the same and what the three trials show. Then check: does this evidence support the prediction you made at the start of the lesson?
Claim: In this test, diet cola made the tallest eruption.
Evidence: All three diet cola trials (250, 230, and 260 cm) were higher than every regular cola trial (90, 110, and 80 cm). The flat cola trials were 0 cm every time.
Reasoning: The type of soda was the only thing the club changed. They kept the brand, bottle size, temperature, number of Mentos, and drop method the same, so those factors cannot explain the difference in height. The heights varied a little from trial to trial, but diet cola was highest in all three, so the result was not a one-time fluke. These results show which soda erupted highest, not why. One reasonable inference: flat cola had lost most of its gas and did not erupt at all, so the dissolved gas is probably what forms the foam.
If you predicted a different soda, this evidence does not support that prediction - and changing your thinking when the evidence disagrees is exactly what scientists do. If your answer makes a claim, cites specific measurements, and explains why they support it, you have it.
- Have students reason from supplied Mentos results they did not collect: a claim, specific measurements, and reasoning.
- Rehearse the claim-evidence-reasoning routine students will use with their own data in the upcoming hands-on investigation.
- Generation and self-explanation surface gaps that multiple choice hides
- Mini CER matches an evidence-and-inference lesson: claim, evidence, reasoning
- Repeated trials let students judge consistency, not just a single reading
- Model answer revealed only after submitting supports honest self-check
- Analyze to Evaluate
- DOK 3 (construct and defend a conclusion from evidence)
- Supplied data table with a labeled setup, next to the prompt
- Required before submit, with a model answer to compare against
- The same closing routine students meet in every lesson
- Measure all four goals, including a brand-new experiment students must analyze.
- Offer practice mode for self-check and classroom mode for teacher reporting.
- Retrieval practice across the whole lesson in one sitting
- Feedback loops: every item has an answer explanation
- Transfer items apply the method to unfamiliar scenarios
- Understand to Apply
- DOK 1 to 2 (mixed recall and applied items)
- Plausible, evenly placed answer options
- Answer explanations provided
- Required name, teacher, and block before classroom submission
More Learning
Try it yourself: Pick one of your own “Why does that happen?” questions and plan a fair test on paper. Name the independent variable you would change, the dependent variable you would measure, and the controls you would keep the same. Then review your plan: Is there anything else you would need to control to make it a fair test?
- Move the method into a self-designed fair-test plan students write on paper.
- Use a peer swap to hunt for an uncontrolled variable in a partner's design.
- Transfer applies the full method to a student's own question
- Error analysis: critiquing a peer design targets weak fair tests
- Interest-driven extension sustains motivation past the lesson
- Apply to Analyze
- DOK 2 to 3 (designing an investigation and critiquing a peer's controls)
- Optional and self-paced, no penalty for skipping
- Linked next steps sorted by type and color
- Challenge restated in plain, concrete steps
Connections
Science and engineering share the same toolkit. Here is where these skills carry over.