Asking Questions & Planning Investigations
Learn to spot testable questions, write If/then/because hypotheses that could be proven wrong, and plan investigations with measurements, trials, and clear evidence.
What you'll do in this lesson
A voice-first session with the Crimsora tutor on Asking Questions & Planning Investigations, then targeted practice and FRQs — with the tutor adapting to where you get stuck.
What this lesson covers
Every real investigation starts with a question — but not every question can be answered by doing science. "Which flower is prettiest?" and "Do marigolds grow taller in full sun than in shade?" sound similar, yet only one of them can be settled by collecting measurements. Learning to tell those apart is the first job of a scientist, and it is the first job in this lesson.
Once you have a question worth testing, you need two more things before you touch a single seed, thermometer, or stopwatch: a hypothesis stated clearly enough that the data could prove it wrong, and a plan that says exactly what you will measure, how many times you will measure it, and what pattern in the numbers would count as evidence. By the end of this lesson you will be able to take a vague wondering and turn it into an investigation another student could run without asking you a single question.
Once you have a question worth testing, you need two more things before you touch a single seed, thermometer, or stopwatch: a hypothesis stated clearly enough that the data could prove it wrong, and a plan that says exactly what you will measure, how many times you will measure it, and what pattern in the numbers would count as evidence. By the end of this lesson you will be able to take a vague wondering and turn it into an investigation another student could run without asking you a single question.
Testable Questions vs. Questions Science Cannot Answer
A testable question is one you could answer by making observations or measurements. It points at something you can change or compare, and something you can record with numbers or careful descriptions.
Questions fail the testable check for three common reasons. Some are about opinion or value — beauty, fairness, what someone "should" do. Science can tell you which soda dissolves an eggshell fastest, but not which soda tastes best. Some are too vague — "Does music affect plants?" doesn't say what music, what plants, or what about the plants. Some are impossible to gather evidence for with the tools and time you actually have, like "What did dinosaurs dream about?"
A good repair trick: take a fuzzy question and add three details — what you will change, what you will measure, and over what time or amount. "Does light matter for plants?" becomes "Do radish seedlings grown under 12 hours of light per day reach a greater height in centimeters after 10 days than seedlings grown under 4 hours per day?" Same curiosity, but now it can actually be answered with evidence.
Questions fail the testable check for three common reasons. Some are about opinion or value — beauty, fairness, what someone "should" do. Science can tell you which soda dissolves an eggshell fastest, but not which soda tastes best. Some are too vague — "Does music affect plants?" doesn't say what music, what plants, or what about the plants. Some are impossible to gather evidence for with the tools and time you actually have, like "What did dinosaurs dream about?"
| Question | Testable? | Why |
|---|---|---|
| Is it wrong to keep fish in a tank? | No | Asks about right and wrong, not measurable facts |
| Do goldfish swim faster in warmer water? | Yes | Compare swim speed at two water temperatures |
| Are plants amazing? | No | Opinion, nothing to measure |
| Does soil type affect how tall bean plants grow in 14 days? | Yes | Names the change, the measurement, and the time |
Writing an If/Then/Because Hypothesis That Could Be Wrong
A hypothesis is a proposed answer to your testable question, stated so clearly that data could contradict it. The classroom form has three parts:
If names the thing you change. Then names the measurable result you predict. Because gives your scientific reasoning — the idea from what you already know that makes the prediction sensible.
Example: "If radish seedlings receive 12 hours of light per day instead of 4, then their average height after 10 days will be greater, because more light means more photosynthesis, which supplies more sugar for building new cells."
The most important property of a hypothesis is that it is falsifiable — it must be possible for the results to disagree with it. "If I water plants, then something will happen" cannot be wrong, so it is useless. "If I water plants with salt water, then their average height after 10 days will be less than plants given tap water" can absolutely turn out wrong, which is exactly what makes it scientific.
Being wrong is not failure. A hypothesis that the data contradicts still teaches you something real, and scientists publish those results all the time. Students often try to protect their hypothesis by fudging measurements or ignoring one weird plant. Resist that. Your job is to test the idea honestly, not to defend it.
Also notice that the "because" clause is where science reasoning shows up. "Because I think so" or "because that's what happened last time" is not reasoning. Point to a mechanism: photosynthesis, dissolving, friction, insulation, surface area. A strong because tells your reader why the world would behave that way.
If names the thing you change. Then names the measurable result you predict. Because gives your scientific reasoning — the idea from what you already know that makes the prediction sensible.
Example: "If radish seedlings receive 12 hours of light per day instead of 4, then their average height after 10 days will be greater, because more light means more photosynthesis, which supplies more sugar for building new cells."
The most important property of a hypothesis is that it is falsifiable — it must be possible for the results to disagree with it. "If I water plants, then something will happen" cannot be wrong, so it is useless. "If I water plants with salt water, then their average height after 10 days will be less than plants given tap water" can absolutely turn out wrong, which is exactly what makes it scientific.
Being wrong is not failure. A hypothesis that the data contradicts still teaches you something real, and scientists publish those results all the time. Students often try to protect their hypothesis by fudging measurements or ignoring one weird plant. Resist that. Your job is to test the idea honestly, not to defend it.
Also notice that the "because" clause is where science reasoning shows up. "Because I think so" or "because that's what happened last time" is not reasoning. Point to a mechanism: photosynthesis, dissolving, friction, insulation, surface area. A strong because tells your reader why the world would behave that way.
Planning the Investigation: Measurements, Trials, and Evidence
A plan is complete when another student could follow it without asking you anything. Three parts do most of that work.
What will be measured. Name the quantity and the unit, and say how you will measure it. "Growth" is not a measurement; "height in centimeters from soil surface to the tip of the tallest leaf, measured with a ruler each morning at 8:00" is. Turning a fuzzy idea into a specific measuring procedure is called giving an operational definition.
How many trials. One plant, one drop, one run tells you almost nothing, because living things and messy materials vary. Running several trials and averaging them lets a real pattern show through the noise. In a classroom investigation, five per group is a reasonable minimum, and more is better. Repeating also lets you spot an oddball result that came from a mistake.
What counts as evidence. Decide before you start what result would support your hypothesis and what result would contradict it. "If the 12-hour group's average height is at least 2 cm greater than the 4-hour group's average, I will count that as support. If the averages are within 0.5 cm, I will say the light difference had no clear effect."
Write the plan down as numbered steps in your notebook before you begin. Deciding what counts as evidence ahead of time keeps you honest when the numbers come in.
What will be measured. Name the quantity and the unit, and say how you will measure it. "Growth" is not a measurement; "height in centimeters from soil surface to the tip of the tallest leaf, measured with a ruler each morning at 8:00" is. Turning a fuzzy idea into a specific measuring procedure is called giving an operational definition.
How many trials. One plant, one drop, one run tells you almost nothing, because living things and messy materials vary. Running several trials and averaging them lets a real pattern show through the noise. In a classroom investigation, five per group is a reasonable minimum, and more is better. Repeating also lets you spot an oddball result that came from a mistake.
What counts as evidence. Decide before you start what result would support your hypothesis and what result would contradict it. "If the 12-hour group's average height is at least 2 cm greater than the 4-hour group's average, I will count that as support. If the averages are within 0.5 cm, I will say the light difference had no clear effect."
| Plan part | Weak version | Strong version |
|---|---|---|
| Measurement | See how they grow | Height in cm, measured daily at 8:00 |
| Trials | Try it | 5 seedlings per group, 10 days |
| Evidence | See which is better | Compare group averages; 2 cm difference counts as an effect |
Where Students Usually Go Wrong
The first common slip is writing a prediction and calling it a hypothesis. "The plant in the sun will be taller" is a prediction — it says what will happen but not why. Adding the because clause turns it into a hypothesis you can actually reason about.
The second slip is a question that is technically answerable but so broad the investigation collapses. "How does the environment affect animals?" cannot be planned. Narrow it until exactly one thing is changing and exactly one thing is being measured.
The third slip is running a single trial. If one seedling in the bright group grows 9 cm and one in the dim group grows 8 cm, you have learned nothing — individual seeds differ that much on their own. Multiple trials and an average are what make a difference believable.
The fourth slip is deciding what counts as evidence after seeing the data. It is very tempting to look at a 0.3 cm gap and announce that light helped. Setting your threshold in advance prevents that.
A fifth slip is measuring something that does not match the question. If your question asks about growth rate but you only record the final height, you cannot answer what you asked. Read your question and your measurement side by side and check that they match.
Finally, students often forget to record units, times, and tools. "Height: 7" is not data — seven centimeters, seven inches, or seven millimeters? A data table with labeled units, written before the first measurement, prevents a whole afternoon of confusion later when you graph the results.
The second slip is a question that is technically answerable but so broad the investigation collapses. "How does the environment affect animals?" cannot be planned. Narrow it until exactly one thing is changing and exactly one thing is being measured.
The third slip is running a single trial. If one seedling in the bright group grows 9 cm and one in the dim group grows 8 cm, you have learned nothing — individual seeds differ that much on their own. Multiple trials and an average are what make a difference believable.
The fourth slip is deciding what counts as evidence after seeing the data. It is very tempting to look at a 0.3 cm gap and announce that light helped. Setting your threshold in advance prevents that.
A fifth slip is measuring something that does not match the question. If your question asks about growth rate but you only record the final height, you cannot answer what you asked. Read your question and your measurement side by side and check that they match.
Finally, students often forget to record units, times, and tools. "Height: 7" is not data — seven centimeters, seven inches, or seven millimeters? A data table with labeled units, written before the first measurement, prevents a whole afternoon of confusion later when you graph the results.
Key terms
- Testable question.
- A question that can be answered by collecting observations or measurements, because it names something to change and something to measure.
- Hypothesis.
- A proposed answer to a testable question, usually written in If/then/because form, that data could either support or contradict.
- Falsifiable.
- Able to be shown wrong by evidence. A statement that no possible result could contradict is not a scientific hypothesis.
- Prediction.
- A statement of what result you expect. It becomes a hypothesis when you add reasoning explaining why that result should happen.
- Operational definition.
- A precise statement of how a quality will be measured, such as defining plant growth as height in centimeters from soil to the tallest leaf tip.
- Trial.
- One complete run of an investigation. Repeating trials and averaging reduces the effect of random variation.
- Evidence.
- Recorded data used to support or challenge a claim. Useful evidence is measured with units and compared against a standard you set in advance.
- Investigation plan.
- A written procedure naming what will be measured, how many trials will be run, and what results would count as support for the hypothesis.
Worked example
A student writes in her notebook: "Is music good for plants?" Turn this into a testable question, write an If/then/because hypothesis, and plan an investigation that names the measurement, the number of trials, and what will count as evidence.
Step 1 — Check whether the question is testable. "Good" is a value word, and "music" is vague. As written, the question cannot be answered with measurements, so it must be rewritten.
Step 2 — Add the three missing details. What will change? Hours of music per day. What will be measured? Bean seedling height in centimeters. Over what time? Twelve days. Revised question: "Do bean seedlings exposed to 3 hours of classical music per day grow taller in 12 days than seedlings kept in silence?"
Step 3 — Write the hypothesis in If/then/because form. "If bean seedlings are exposed to 3 hours of music per day, then their average height after 12 days will be no different from seedlings kept in silence, because plant growth depends on light, water, and nutrients, and sound waves do not supply any of those." Notice this hypothesis could be wrong — if the music group ends up clearly taller, the data contradicts it, which is exactly what a good hypothesis allows.
Step 4 — Name the measurement. Height in centimeters, from the soil surface to the tip of the tallest leaf, measured with a metric ruler every morning at 8:30.
Step 5 — Name the trials. Six seedlings in the music group and six in the silent group, all in identical cups with the same soil, same water amount, and same light. Twelve days total.
Step 6 — State what counts as evidence. Compute each group's average final height. If the two averages differ by 2 cm or more, count that as evidence that music had an effect. If they differ by less than 2 cm, report no clear effect.
The original wondering is now a plan another student could carry out exactly as written.
Step 2 — Add the three missing details. What will change? Hours of music per day. What will be measured? Bean seedling height in centimeters. Over what time? Twelve days. Revised question: "Do bean seedlings exposed to 3 hours of classical music per day grow taller in 12 days than seedlings kept in silence?"
Step 3 — Write the hypothesis in If/then/because form. "If bean seedlings are exposed to 3 hours of music per day, then their average height after 12 days will be no different from seedlings kept in silence, because plant growth depends on light, water, and nutrients, and sound waves do not supply any of those." Notice this hypothesis could be wrong — if the music group ends up clearly taller, the data contradicts it, which is exactly what a good hypothesis allows.
Step 4 — Name the measurement. Height in centimeters, from the soil surface to the tip of the tallest leaf, measured with a metric ruler every morning at 8:30.
Step 5 — Name the trials. Six seedlings in the music group and six in the silent group, all in identical cups with the same soil, same water amount, and same light. Twelve days total.
Step 6 — State what counts as evidence. Compute each group's average final height. If the two averages differ by 2 cm or more, count that as evidence that music had an effect. If they differ by less than 2 cm, report no clear effect.
The original wondering is now a plan another student could carry out exactly as written.
Practice questions
Which of the following is a testable question?
- Which is the most beautiful type of butterfly?
- Should students be allowed to keep classroom pets?
- Do mealworms move toward a damp paper towel more often than a dry one in 5 minutes?
- Are insects important?
Answer: Do mealworms move toward a damp paper towel more often than a dry one in 5 minutes?
A testable question names something to change and something to measure. The mealworm question changes the paper towel condition and counts how many mealworms move toward each within a set time, so it produces data. "Most beautiful" and "should students be allowed" ask about opinions and values, which evidence cannot settle. "Are insects important?" gives nothing to measure at all.
A student writes: "If I put a plant in the closet, then something will change about it, because plants react to their surroundings." Explain why this is not a usable hypothesis, and rewrite it so it could be shown wrong.
Answer: It is not falsifiable, because "something will change" is true no matter what happens. A better version: "If a bean plant is kept in a dark closet for 7 days, then its leaves will turn yellow and it will gain less than 1 cm in height, because without light it cannot photosynthesize to make the sugars needed for growth and chlorophyll."
A hypothesis must specify a measurable result that the data could contradict. "Something will change" cannot fail — any result confirms it, so the investigation cannot teach you anything. The rewrite names exactly what is measured (leaf color and height change in centimeters) over a set time, and its because clause points to a real mechanism, photosynthesis. Now a plant that stays green and grows 4 cm would clearly contradict the hypothesis.
Two students test whether cold water or room-temperature water dissolves sugar faster. Each student stirs one cup once and records the time. They report that cold water is slower. Name two things missing from their plan and explain why each matters.
Answer: They ran only one trial per condition, and they never defined what counts as "dissolved" or what size difference would count as evidence.
With a single trial in each cup, a difference could easily come from stirring speed, sugar clumping, or reaction time on the stopwatch rather than from temperature. Running at least five trials per condition and averaging lets the real pattern show. Second, "dissolved" needs an operational definition — for example, no visible grains when the cup is held up to a window — so both students judge the endpoint the same way, and they should decide in advance how large a time gap counts as a real difference.
FAQ
- What is the difference between a hypothesis and a prediction?
- A prediction says what you expect to happen. A hypothesis says what you expect and explains why, using science reasoning. "The seeds in warm soil will sprout first" is a prediction. Adding "because warmer temperatures speed up the chemical reactions inside the seed" turns it into a hypothesis. The If/then part is the prediction; the because part is what makes it a hypothesis.
- Is it bad if my hypothesis turns out to be wrong?
- No. A hypothesis is a well-reasoned guess, not a promise. Results that contradict your hypothesis are still real evidence and still teach you something about how the world works. What matters is that you report the data honestly, explain what actually happened, and suggest what you would investigate next. Changing your data to match your hypothesis, on the other hand, is the one thing that ruins an investigation.
- How many trials should I run?
- Enough that one unusual result cannot control your conclusion. For a middle-school investigation, five trials per condition is a sensible floor, and living things like seeds or mealworms usually need more because individuals vary so much. Always report the average of your trials plus the individual values, so a reader can see how spread out your results were.
- How do I know what should count as evidence?
- Decide before you collect data. Pick a difference large enough that it could not plausibly come from measuring error — for example, at least 2 cm of height, or at least 5 seconds of dissolving time. Then compare group averages against that threshold. Choosing the threshold in advance keeps you from talking yourself into seeing a pattern in numbers that are basically the same.
Learn this with a teacher, not a page
The Crimsora tutor teaches Asking Questions & Planning Investigations live — explaining on a whiteboard, asking you questions, and adapting to where you get stuck.