How to choose a science project to do at home that your kid finishes

Choose the project from a question your kid already has, offer three to five options, pitch it one step past what they can do alone, and pick one that ends as a record of their thinking, not a fizz. Home science rarely fails at the doing. It fails at the choosing.

Why doesn't a list of fifty experiments help you choose?

Because a longer list makes the choosing harder. Search for science experiments at home, or for science projects for kids, and you get fifty ideas, seventy, sometimes a thousand, and leave with a longer list. In a survey of 1,191 high school students, Grinnell, Dalley, and Reisch (2025) found the second most common complaint about a science project was that it was "not a good project" (17.9% of negative comments).

Kubrio's shelf is short on purpose. Every one of the projects on Kubrio is named for the thing the kid ends up holding, so picking one means picking an outcome.

Should your kid choose the project?

Yes, and more than once. Patall, Cooper, and Robinson (2008) pooled 41 studies on the effects of choice: choice raised intrinsic motivation (d = 0.30) and perceived competence (d = 0.59), with a far larger effect for children (d = 0.55) than adults (d = 0.25). Three to five options beat two and beat more than five, and two to four choices across a task (d = 0.61) beat one choice at the start (d = 0.21). Choice did not improve later learning (d = 0.10, not significant), so claim only engagement and finishing.

The science-fair data agrees. Grinnell, Dalley, and Reisch (2025) found 89.3% positive comments when taking part was optional, 52.6% when required, and 65.1% when required but the student picked the project. Students coached by a scientist gave 88.3% positive comments, against 58.7% without.

Kubrio works that way. A kid picks from the shelf, then keeps choosing inside the project: in Sail Darwin's Voyage to the Galápagos, which five observations deserve a field note in each chapter, and what to draw.

How do you find the question your kid already has?

Listen for the question they ask twice. Chouinard (2007) described children's questions as an information-requesting mechanism: kids ask, get an answer, and persist when the answer is inadequate. A question re-asked after a poor answer is your project. A question they drop was conversation.

Expect fewer with age. Engel (2011) counted curiosity episodes in classrooms over three months: kindergarten rooms produced two to five per two-hour stretch, fifth-grade rooms zero to two per visit. In one lesson the teacher asked ten questions and no child asked any.

If nothing surfaces, your kid does not dislike science. Hidi and Renninger (2006) describe interest as four phases beginning with a triggered situational interest, sparked by what is around the child. Phase one calls for a short, vivid project.

Noticing is trainable. Kubrio's guide on how to teach your kid about nature is about exactly that habit: notice something, name it, connect it to what's around it.

How hard should the project be?

One step past what your kid can do alone, with you supplying the step. Jirout and Klahr (2012) define curiosity as a threshold of desired uncertainty, the amount of not-knowing a child moves toward. Below it they are bored; above it they leave.

Alfieri and colleagues (2011) pooled 164 studies of discovery learning: unassisted discovery lost to explicit instruction (d = −0.38), while assisted discovery, with feedback, worked examples, scaffolding, and elicited explanations, beat other instruction (d = 0.30). Kirschner, Sweller, and Clark (2006) add the rule for how much: the advantage of guidance recedes only once the learner has prior knowledge to guide themselves. Scale help to what your kid knows about the subject, not their age.

Bjork and Bjork (2011) call these "desirable difficulties", with a boundary: a difficulty turns undesirable when the learner lacks the background knowledge to meet it.

Kubrio pitches projects the same way: the kid makes the thing by hand first, the AI renders or lays it out, and help arrives as a next question rather than an answer. The hand stays the hero. The evidence behind guided making is in Kubrio's article on the science of hands-on learning.

What should the project end as?

Something that still exists next month. There are two kinds of projects for kids. The first kind ends on the fridge door. The second kind ends as a real thing other kids can read, watch, or play. Kubrio makes the second kind. Kids make them by directing the same AI the pros use.

A demo evaporates. The volcano erupts, the counter gets wiped, and by Tuesday there is nothing to point at. What survives is a record of the thinking: a journal, a chart, a film, an explainer. Papert and Harel argued in Situating Constructionism (1991) that learning goes best when the maker is "consciously engaged in constructing a public entity." A fizzing cup is not one. A journal with a name on the cover is.

Decide the ending before the materials. Sail Darwin's Voyage to the Galápagos is built backwards from it: your kid boards the Beagle as the ship's naturalist, writes five field notes per chapter, and publishes as they go, and the chapters become an expedition journal in their name.

What do you say while they work?

Ask more than you explain. Callanan and colleagues (2017) recorded 82 families with children aged 3 to 11 and found children's conceptual engagement was predicted by parents' critical-thinking questions, evidence talk, and requests for explanation, while their explanation statements were negatively correlated with it. "What makes you think that?" produces the thinking that narrating the answer suppresses.

Crowley and colleagues (2001) watched families at museum exhibits and found parents were three times more likely to explain the science to sons than to daughters, though they talked to both equally about using the exhibit and about the evidence. None of them meant to. Notice which you are doing, and count.

Kubrio holds the same line. Krea, Tek, and Brio, the AI crew in every project, ask a better question and never hand over the answer. Claire coaches you, not the project.

Ten science projects to do at home, by age

Offer three to five of these, never ten.

Ages 6 to 8

• The cold-spot map. "Where is the coldest place in our home?" A thermometer and sticky notes. Ends as a hand-drawn temperature map.
• One square of ground. "What lives in one square of our yard?" A magnifying glass and a notebook. Ends as a dated field journal in their name.
• The hidden-sugar shelf. "Which snack hides the most sugar?" Six packets, a teaspoon, small jars. Ends as a photo chart of the jars in order.
• Which way is up? "Do seeds know which way is up?" A jar, paper towel, four beans, water. Ends as a ten-day photo strip.

Ages 9 to 11

• Flight test report. "Which paper airplane flies furthest?" Three designs, five throws each, a tape measure. Ends as a one-page report with a bar chart.
• The dirtiest thing I touch. "Which surface here is the dirtiest?" Four slices of bread in sealed bags. Ends as a labeled photo series.
• Does music move your heart? "Does music change how fast your heart beats?" A stopwatch and two very different tracks. Ends as a chart and a short explainer film.
• Which soap cuts grease? "Do expensive dish soaps work better?" Oil, water, four jars, four soaps. Ends as a consumer-report poster with times and prices.

Ages 12 to 13

• The one-week trash audit. "How much of our trash could have been recycled?" Gloves and a kitchen scale. Ends as a stacked bar chart and a rule sheet above the bin.
• Is your phone lying about your steps? "My phone says I walked 8,000 steps. Is that true?" A measured stretch and a clicker. Ends as an error chart and a verdict.

Decide the audience first. Every Kubrio season ends with Demo Week, December 1–7, 2026, when kids premiere what they made.

What to do at home this week

1. Write down every question your kid asks out loud for three days. Mark the ones they ask twice.
2. Turn the two strongest into projects, add three from above, and offer those five.
3. Ask which one they want, then ask what they think will happen.
4. Agree on the ending in one sentence, and the date it gets shown.
5. Book two more choices: materials, and how they show the result.
6. While they work, ask three times for every time you explain.

Ingenious projects that inspire kids to make things — until they find the thing only they could make.

Start with Sail Darwin's Voyage to the Galápagos: your kid boards the Beagle as the ship's naturalist and finishes with an expedition journal in their own name. Or explore all the projects. When you're ready, start here.

FAQ

What about a science fair project? The same rules apply, and the choice matters most. Grinnell, Dalley, and Reisch (2025) found 89.3% positive comments when taking part was optional, 52.6% when required, and 65.1% when required but the student picked the project. If the fair is compulsory, hand over the choosing. Any of the ten above works as an entry.

What age can my kid start? Six is a reasonable floor; these suit kids 6 to 13. At the younger end the science is observing and recording rather than testing a hypothesis, which is why the first four end in maps, journals, and charts. Difficulty sits in the record-keeping, not the chemistry.

How long should a science project at home take? Longer than an afternoon, shorter than a term. Most run on short daily sittings across one or two weeks, because the change over time is the point.

What if my kid quits halfway? Assume the project was mispitched before you assume anything about your kid. Bjork and Bjork's boundary is the usual cause: a stretch works only when the groundwork is there. Drop one variable, or go back to the five options and let them choose again.