Precision, Sample Size and Replication: a skill companion
A small set of reusable sheets that grow one inquiry skill: choosing equipment fine enough for the change being measured, taking enough repeats for the result to settle, and recording the work fully enough that another group could replicate it, with digital tools where they help. Print the scaffolds once and slot them into the science lessons you already teach.
What a skill companion is
Inquiry skills are not a topic of their own. They grow inside the science units a class teaches all year, from Body Systems to Chemical Equations. So this pack is not a full term of lessons. It is three reusable scaffolds, a map of where they fit, and a short stand-alone lesson for teaching the skill on its own first.
Start here: five minutes
- Read the pairing map on the next page: it shows the instrument, the sample size and the record for each topic.
- Print the instrument chooser when a class plans a measurement, and the repeats sheet when it has a set of runs to work with.
- Cut out the cards once. They are reused all year.
- Open the free interactive unit on your board when you want a worked example of the skill.
- Run the one-page mini-lesson first if you want to teach the skill before folding it into a topic.
The step up from Year 8
In Year 8 students chose equipment and repeated trials. In Year 9 they decide how precise is precise enough (a step no bigger than a tenth of the change), how many repeats are enough (until the running mean settles), and what record lets someone else replicate the result. Data loggers and video make many precise readings cheap, but the judgement stays with the investigator.
Slot the skill into your science lessons
Every measurement in the year asks the same three questions: is the instrument precise enough, is the sample big enough, and could another group replicate it? This map answers them for one measurement in each Year 9 topic.
| When you teach | The measurement | Precise enough | A useful sample | Record for replication |
|---|---|---|---|---|
| Body Systems and Negative Feedback (AC9S9U01) | Heart rate after exercise | A finger pulse sensor, or beats counted for a full 60 seconds | At least 20 students, two trials each | How and where the pulse was taken, and when |
| Reproduction in Animals and Plants (AC9S9U02) | Roots on cuttings | A hand lens, counting roots longer than 2 mm | 10 cuttings in each group | Photos with a ruler in shot; the counting rule |
| The Carbon Cycle (AC9S9U03) | Indicator colour around a leaf | A printed colour chart, or a colour sensor on a data logger | Three tubes at each lamp distance | The lamp, its distances and the colour chart |
| Wave and Particle Models of Energy (AC9S9U04) | Time for a pulse along a slinky | Slow-motion video, counted frame by frame | Ten pulses at each stretch | The stretch in cm and the frame rate |
| Conservation of Energy and Efficiency (AC9S9U05) | Bounce height of a ball | Video against a tape measure marked in cm | Five drops from each height | The ball, the floor and how the height was read |
| The Atom and Radioactive Decay (AC9S9U06) | Dice left after each throw | Counts: exact, so precision is not the problem | Five runs of 60 dice, pooled | The removal rule and every run’s counts |
| Chemical Equations and Conservation of Mass (AC9S9U07) | Mass lost from an open cup | An electronic balance reading to 0.01 g | Three trials for each mass of bicarbonate of soda | Volumes, masses and when each reading was taken |
| Any topic with a data logger | Any reading that changes over time | A sensor whose step is no bigger than a tenth of the change | Enough runs for the running mean to settle | The sensor, its interval and the raw file |
The three moves, and the picture that backs each one
When you want a worked example on the board, open the interactive unit and use the picture that matches the move.
- Repeat until the mean settles: “Drain time across more and more repeats”.
- Log every run, and check a stray one while the gear is still set up: “Twelve logged runs of the same test” and the stray run.
- Keep the record another group needs: “Judge the case for replicable data”.
Precise enough for the change?
An instrument is precise enough when its smallest step is no bigger than about a tenth of the change you want to see. A step as big as the change hides it; a step a tenth of the change shows it clearly.
A. Choose the instrument
| The change, about | Instrument 1 | Instrument 2 | Choose, and why |
|---|---|---|---|
| The mass an open cup of vinegar loses as bicarbonate of soda fizzes in it: 0.8 g | a kitchen scale that reads to 1 g | an electronic balance that reads to 0.01 g | |
| The drop in temperature of a cup of warm water in 5 minutes with a fan: 3 °C | a digital probe that reads to 0.1 °C | a glass thermometer marked every 1 °C | |
| The time a pulse takes to cross a stretched slinky: 0.6 s | a slow-motion video at 240 frames a second, counted frame by frame (about 0.004 s a frame) | a stopwatch started and stopped by hand (reaction time about 0.2 s) | |
| The height a tennis ball bounces back up from a 1 m drop: 70 cm | a guess against marks on the wall every 10 cm | a tape measure marked in centimetres, read from a video |
B. My own measurement
Teacher note: precision is not accuracy. A balance that reads to 0.01 g but was not zeroed gives precise, wrong numbers; ask students to zero, or tare, before every reading.
How many repeats, and does it replicate?
One run can mislead. The running mean, the mean of all the runs so far, jumps about at first and then settles; when it stops moving, the sample is big enough. Another group that follows the same method and gets a matching mean has replicated the result.
A. Class 9A’s timings
The time (s) for a pulse to cross the same slinky, stretched to 3 m, read from slow-motion video. Fill the running mean to three decimal places.
| Run | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 |
| Time (s) | 0.66 | 0.55 | 0.62 | 0.57 | 0.64 | 0.58 | 0.60 | 0.59 | 0.61 | 0.60 | 0.59 | 0.60 |
| Running mean |
B. When is the sample big enough?
After how many runs does the running mean stay within 0.005 s of the mean of all twelve? What would one run alone have told you?
C. Does it replicate?
Class 9D followed the same written method with their own slinky and phone: 0.60, 0.63, 0.61, 0.59, 0.62, 0.60, 0.61, 0.62, 0.60, 0.61, 0.62, 0.61.
Precise, a useful sample, or replicable?
Cut out the cards. Lay down the three grey heading cards, then sort the practices under them. On the blank cards, write a practice your own investigation needs for each heading.
Teacher note: the piles are not the same size. The answer sheet lists which card is which, and why.
One swing, ten swings, many runs
Use this stand-alone lesson to teach the skill on its own, before you fold it into a science topic. A pendulum of string and a washer swings in about 1.4 seconds when it is 50 cm long. Timed by hand, one swing is swamped by reaction time, about 0.2 seconds at each end. Timing ten swings and dividing by ten shrinks that error tenfold; repeating, and comparing with other groups, shows when the answer has settled and whether it replicates.
We are learning to
- choose a way of measuring that is precise enough for the quantity,
- take enough repeats for the mean to settle,
- record the work so another group can replicate the result.
Success criteria
- I can explain why timing ten swings is more precise than timing one.
- I can say when my running mean has settled, and whether my result matches another group’s.
You need
- for each pair: about 60 cm of string, a heavy washer or nut, a ruler, a stopwatch or phone timer, and tape to fix the string to a desk edge,
- the instrument chooser and the repeats sheet (scaffolds 1 and 2), and the cards from scaffold 3, one set per table. Keep the swing clear of faces.
Lesson flow (about 30 minutes)
| 5 min | One swing Pairs set the pendulum to 50 cm and time one swing, five times, writing every time down. Ask: “Your five times disagree. Is the pendulum changing, or is the timing?” |
| 10 min | Ten swings Now time ten swings and divide by ten, five times. Show the board (see the box). Ask: “The hand still adds about 0.2 seconds at each end. Why does that matter less now?” |
| 10 min | Settle and replicate Pairs work the running mean of their ten-swing results, then write their mean, string length and method on the board. While the class compares, tables sort the cards. Ask: “Do the pairs agree? What would another class need from us to check our answer?” |
| 5 min | Make it better Pairs name one change that would make the result more precise (for example, a slow-motion video) and one record that would help another class. |
Running it shorter? Stop after the ten-swing timings, and pick up the running mean and the cards in your next science lesson.
Watch for these ideas
- Thinking a digital stopwatch is precise to 0.01 seconds when a hand starts and stops it.
- Stopping the repeats as soon as one result looks right.
- Deleting a stray result instead of checking the apparatus and recording why.
Make it easier, make it bigger
- Easier: give pairs a table with columns for ten swings, the time for one, and the running mean.
- Bigger: time twenty swings, or film the swing in slow motion, and compare the spread of the results with the ten-swing ones.
Answers and look-fors
The next sheet has the instrument choices, the worked running mean and replication check, the card key, and a quick three-level guide.
Answers and look-fors
Instrument chooser
| The change | Choose | Why |
|---|---|---|
| The mass an open cup of vinegar loses as bicarbonate of soda fizzes in it: about 0.8 g | Instrument 2: an electronic balance that reads to 0.01 g | A tenth of the change is 0.08 g: a step of 0.01 g is fine enough; 1 g is not. |
| The drop in temperature of a cup of warm water in 5 minutes with a fan: about 3 °C | Instrument 1: a digital probe that reads to 0.1 °C | A tenth of the change is 0.3 °C: a step of 0.1 °C is fine enough; 1 °C is not. |
| The time a pulse takes to cross a stretched slinky: about 0.6 s | Instrument 1: a slow-motion video at 240 frames a second, counted frame by frame (about 0.004 s a frame) | A tenth of the change is 0.06 s: a step of 0.004 s is fine enough; 0.2 s is not. |
| The height a tennis ball bounces back up from a 1 m drop: about 70 cm | Instrument 2: a tape measure marked in centimetres, read from a video | A tenth of the change is 7 cm: a step of 1 cm is fine enough; 10 cm is not. |
Repeats and replication
Class 9A’s running mean: 0.660, 0.605, 0.610, 0.600, 0.608, 0.603, 0.603, 0.601, 0.602, 0.602, 0.601, 0.601. The mean of all twelve is 0.601 s. From run 6 on, the running mean stays within 0.005 s of it, so about 6 runs make a useful sample; one run alone said 0.66 s, 0.059 s too long. Class 9D’s mean is 0.610 s, 0.009 s from 9A’s, while single runs spread over 0.11 s: the classes agree, so the result replicates. The check needs both classes to have recorded the stretch (3 m), the timing method and frame rate, and every run.
Data cards
| Card | Practice | Pile | Why |
|---|---|---|---|
| A | Drop the ball ten times from each height and use the mean bounce. | Useful sample size | Ten repeats, so one odd bounce barely moves the mean. |
| B | Weigh the cup on a balance that reads to 0.01 g, because the mass lost is under a gram. | Precise | A step a hundredth the size of a gram can show a change under a gram. |
| C | Record the logger’s interval, the battery type and every reading, high or low. | Replicable | The full record lets another group repeat the work and compare. |
| D | Measure the recovery time of 20 students, not 3. | Useful sample size | More students, so one unusual person cannot set the answer. |
| E | Take cuttings from 10 plants for each number of leaves, not from 1. | Useful sample size | Ten plants in each group, not one that might be unusual. |
| F | Time the slinky pulse from a slow-motion video instead of a hand-held stopwatch. | Precise | Frames a few thousandths of a second apart, not a hand that reacts in about 0.2 s. |
| G | Write the slinky’s stretch and the camera’s frame rate beside the results. | Replicable | Another class can set up the same slinky and the same timing. |
| H | Read the ruler with your eye level with the top of the ball, to the nearest millimetre. | Precise | A reading rule that removes the error of looking from above or below. |
| I | Pool five runs of 60 dice each, instead of one run, before drawing the graph. | Useful sample size | Three hundred dice smooth out the chance ups and downs of sixty. |
| J | Use a digital probe reading to 0.1 °C for a drop of about 2 °C. | Precise | A step a twentieth of the change. |
| K | Share the full data table and the method, so another class can check the result. | Replicable | An independent check needs the method and every value. |
| L | Keep adding runs until the running mean stops changing by more than 0.005 s. | Useful sample size | A rule for when the sample is big enough. |
A quick three-level guide
| Move | Working towards | At standard | Beyond |
|---|---|---|---|
| Precision | uses whatever instrument is to hand | chooses a step no bigger than a tenth of the change, and a reading rule | tells precision from accuracy, and zeroes or checks the instrument |
| Sample size | takes one reading, or stops at a result that looks right | repeats until the running mean settles | decides the number of repeats from the scatter before starting |
| Replication | keeps only the results that fit | records the equipment, settings and every reading | compares with an independent group and judges whether the results agree |
A student at standard picks an instrument fine enough for the change, repeats until the mean settles, and keeps a record another group could follow. The pendulum in the mini-lesson: about 1.4 seconds a swing at 50 cm, whatever the washer.