How Do You Use a Pilot Study in an HSC Science Investigation?

You want to investigate how much a spring stretches when you add different masses. The lightest mass barely moves your measuring pointer. The heaviest stretches the spring so far that it doesn’t return to its original length. Should you just collect more measurements and hope the results become clearer?

No. You need to find out whether your method can produce useful measurements before you commit to it.

That’s what a pilot study does. In an HSC Science investigation, you use a pilot study to run a small preliminary trial, test whether your measurement range and equipment are suitable, identify practical problems, and make justified changes before collecting your final data.

The goal isn’t to prove your hypothesis early. It’s to make sure your experiment can test it properly.

01What should a pilot study actually test?

Imagine trying to weigh a paperclip using kitchen scales that only display measurements in 10 g increments. The scales might work perfectly, but they aren’t suitable for such a small object.

Scientific equipment has similar limitations.

A pilot study helps you identify three things before your full investigation.

These are connected, but they’re not interchangeable.

A measurement can fit within an instrument’s range and still be too imprecise to be useful. An experiment can be quick and repeatable but measure the wrong thing.

For example, a gas syringe might hold up to 100 mL, but if its markings are 5 mL apart, it may be unsuitable for measuring changes of just 1 mL.

A useful pilot study therefore asks two separate range questions:

• Independent-variable range: What values will I deliberately test, such as mass, concentration, temperature, or distance?
• Dependent-variable measurement range: Can my instrument measure the resulting extension, gas volume, temperature change, or time accurately enough?

Your independent variable is what you change. Your dependent variable is what you measure in response.

The pilot helps you choose a workable combination of the two.

02How to run a small pilot study

Return to the spring investigation.

The question is: How does the mass suspended from a spring affect its extension?

Step 1: Build the apparatus and identify the measurements

Clamp a spring to a retort stand so that it hangs vertically. Attach a lightweight mass hanger to the bottom.

Fix a ruler beside the spring, parallel to its length. A pointer attached to the bottom of the spring indicates its position on the ruler.

First, record the pointer position with no added mass.

Then add a known mass, allow the spring to stop moving, and record the new pointer position.

The extension is the change in pointer position:

\[ x=L-L_0 \]

Here:

• \(x\) is the spring’s extension, measured in millimetres (mm).
• \(L\) is the pointer position with the added mass, in mm.
• \(L_0\) is the original unloaded pointer position, in mm.

Because the spring is hanging vertically, adding mass increases the downward gravitational force on it.

Keep the spring, ruler alignment, pointer, and measurement procedure consistent. Read the ruler at eye level to avoid parallax error, where the apparent position changes with viewing angle.

Step 2: Predict the likely measurement problem

Suppose you plan to test masses between 30 g and 150 g.

You might predict that the smallest mass will give a very small extension and that the largest mass will be easier to measure.

But does a bigger extension automatically make the experiment better?

Not necessarily. A sufficiently large load may stretch the spring beyond the region where it returns to its original length. The measurement becomes easier to read, but the spring itself may change.

This is why the pilot needs to examine both ends of the proposed range.

03Use the pilot results to revise the method

A pilot becomes useful when you turn an observation into a specific decision.

For the spring experiment, imagine the following preliminary results.

These numbers are hypothetical teaching data, not results from an actual experiment.

Assume the ruler and stand remained fixed.

At first glance, the student might simply discard the 30 g condition and use heavier masses. But the last observation changes that decision.

04What if the pilot problem is in a Chemistry investigation?

The same reasoning applies even when you’re measuring a chemical reaction rather than a moving object.

Suppose your investigation measures gas produced by a reaction.

A reaction flask is sealed with a bung. A delivery tube runs from the flask to a gas syringe. Gas produced inside the flask travels through the tube and pushes the syringe plunger outwards.

You read the collected gas volume from the syringe markings.

Before collecting the full dataset, a small pilot can reveal several problems:

Notice the difference between changing a method to fix a measurement problem and changing it to get a preferred result.

If a reaction genuinely produces very little gas under one condition, that might be scientifically important. You shouldn’t remove that condition simply because it doesn’t support your prediction.

First establish whether the apparatus could have detected the gas.

Also, changing from gas volume after 30 seconds to time required to produce 10 mL changes the dependent variable. Both may be useful, but they answer different versions of the question.

A pilot helps you make that decision before the final experiment.

05How do you know when the pilot is finished?

You don’t need to complete a miniature version of every planned trial.

Choose a few conditions that are likely to expose measurement and practical problems. Often that means testing the low and high ends of your proposed range, a middle condition, and at least one repeat or control check.

The precise amount depends on the investigation.

Use this decision process.

There isn’t a universal number of pilot trials that makes every investigation reliable.

One preliminary result can reveal an obvious equipment failure. Several measurements may be needed to establish whether an apparent problem is consistent.

You also need to test how long the procedure actually takes.

Suppose five conditions each require three repeat measurements, and each measurement takes two minutes.

The measurement time alone is:

\[ 5\times3\times2=30\,\mathrm{min} \]

That excludes setup, cleaning, resetting equipment, and recording observations.

If your allocated laboratory time is 20 minutes, the proposed procedure isn’t feasible as written.

The solution might involve adjusting the number of conditions, changing the measurement technique, or arranging an approved longer session. Simply rushing through the same procedure risks introducing new errors.

06How should you record the pilot study?

Keep the original observations, including the unsuccessful ones.

A concise record should let another person understand what you tested, what went wrong, and why you changed your method.

For the spring example, a useful record would contain:

Separate the raw observations from your explanation.

For example, “The unloaded pointer moved from 84 mm to 86 mm after removing the largest mass” is an observation.

“The spring may have undergone permanent deformation” is an interpretation.

The distinction matters because the pointer shift might alternatively have resulted from a loose clamp or an unstable measuring reference. Those possibilities should be checked before drawing a firm conclusion.

You should also keep the pilot measurements separate from the final dataset when the procedure has changed.

A measurement using one spring and a different measurement procedure is not automatically an interchangeable repeat of a measurement using another spring.

If you intend to include preliminary measurements that used an unchanged method, check your assessment instructions and clearly identify their origin. Do not quietly combine data collected under different conditions.

07Is a pilot study required by NESA?

NESA’s HSC Science syllabuses include Working Scientifically outcomes concerned with planning, conducting, analysing, and evaluating scientific investigations.

A pilot study is a practical way to demonstrate those skills, but you should not assume that every NSW HSC Science task requires an identical pilot-study format, sample size, or written report.

Your school determines the specific assessment task instructions within the applicable NESA requirements.

Check your teacher’s instructions for whether preliminary trials are required, what can be changed after approval, which equipment may be used, and how the pilot should be documented. School laboratory safety procedures also apply.

There is an important syllabus transition to be aware of as of November 2026:

• NESA Physics 11-12 (2025) begins implementation with Year 11 in 2027, with its first HSC examination in 2028.
• NESA Chemistry 11-12 (2025) begins implementation with Year 11 in 2028, with its first HSC examination in 2029.

The earlier syllabus continues to apply to relevant cohorts during the transition. Check the syllabus your class is actually studying rather than assuming Physics and Chemistry change in the same year.

The official NESA ACE rules provide the broader assessment and certification framework.

If your pilot forms part of a larger investigation, see What Is a Depth Study in HSC Science?. The broader investigation skills are explained in What Is Working Scientifically in HSC Physics and Chemistry?.

For the pilot itself, the useful evidence is simple: what you tried, what the measurements revealed, what you changed, and how you confirmed that the revised method was suitable.

08Questions and solutions

These questions focus on the decisions a pilot study should help you make. The early questions deal with measurement limitations. Later questions involve competing explanations, changing conditions, and whether a measurement really supports the intended conclusion.

All numerical datasets are hypothetical.

Question 1

A student wants to investigate how sodium chloride concentration affects electrical conductivity.

A conductivity sensor measures from 0 to 2000 \(\mu\mathrm{S/cm}\), where \(\mu\mathrm{S/cm}\) means microsiemens per centimetre.

During a pilot:

• At 0.10 mol/L, the sensor reads 1800 \(\mu\mathrm{S/cm}\).
• At 0.20 mol/L, the display shows “OL”, indicating that the measurement exceeds the instrument’s range.

The final investigation must include 0.20 mol/L.

Explain what the student should change and why entering 2000 \(\mu\mathrm{S/cm}\) as the second result would be incorrect.

Solution 1

The student should use a suitably calibrated conductivity meter with a higher measurement range and repeat the pilot, because the current sensor cannot measure the required 0.20 mol/L condition.

“OL” indicates that the actual conductivity is beyond the instrument’s measurable range. It does not mean the conductivity equals 2000 \(\mu\mathrm{S/cm}\).

Plotting it as exactly 2000 would invent a numerical measurement and could distort the relationship between concentration and conductivity.

The tempting option is to record the largest number the instrument can display. But the sensor has supplied a limit, not a measured value.

Because the 0.20 mol/L condition is required, simply excluding it would change the investigation’s scope. A suitable higher-range instrument is the appropriate revision.

Question 2

A Physics student investigates how pendulum length affects the period of oscillation.

The period is the time taken for one complete oscillation.

A pilot gives an average period of approximately 1.00 s for a 25 cm pendulum and 1.09 s for a 30 cm pendulum.

The student estimates that hand-starting and stopping the stopwatch produces a maximum combined timing uncertainty of \(\pm 0.20\) s per timed interval.

A student suggests timing ten complete oscillations instead of one.

Assume the pendulum continues oscillating regularly and both ten-oscillation measurements take less than the available 18 seconds.

Evaluate the proposed change quantitatively. Explain one limitation that would still need checking before final data collection.

Solution 2

Timing ten oscillations is a useful revision because it makes the timing uncertainty a much smaller fraction of the measured interval.

For one oscillation, the difference between the periods is:

\[ \Delta T=1.09-1.00=0.09\,\mathrm{s} \]

Here, \(\Delta T\) is the difference in period.

The difference of 0.09 s is smaller than the estimated \(\pm 0.20\) s timing uncertainty for a single interval. The pilot therefore cannot distinguish the two periods confidently using individual hand timings alone.

For ten oscillations, the expected total times are:

\[ \begin{aligned} t_{25}&=10(1.00)=10.0\,\mathrm{s}\\ t_{30}&=10(1.09)=10.9\,\mathrm{s} \end{aligned} \]

Here, \(t_{25}\) and \(t_{30}\) represent the total times for ten oscillations at the two lengths.

The expected difference becomes:

\[ \Delta t=10.9-10.0=0.9\,\mathrm{s} \]

The estimated worst-case uncertainty in the difference between two separately timed intervals is:

\[ 0.20+0.20=0.40\,\mathrm{s} \]

A 0.9 s difference is now larger than this conservative combined uncertainty bound.

The period is calculated using:

\[ T=\frac{t}{N} \]

where \(T\) is the period in seconds, \(t\) is the total time in seconds, and \(N\) is the number of complete oscillations.

For ten oscillations, the estimated maximum timing uncertainty per period becomes:

\[ \frac{0.20}{10}=0.020\,\mathrm{s} \]

This assumes the combined start-and-stop timing uncertainty remains approximately the same when measuring ten oscillations.

However, the student must still check whether oscillations are counted correctly and whether the pendulum is released consistently at a small angle.

Timing more oscillations improves the relative timing precision, but it does not automatically control every source of experimental variation.

A short repeat pilot using ten oscillations would be justified before collecting the final results.

Question 3

A Chemistry student wants to investigate how reactant concentration affects the volume of gas collected after 30 seconds.

Gas is collected in a 50 mL gas syringe.

The pilot shows:

Further short trials establish that both conditions can produce at least 10 mL of gas.

The low-concentration reaction first reaches 10 mL after 42 s. The high-concentration reaction reaches 10 mL after 6 s.

The student cannot obtain a larger gas syringe.

They propose changing the dependent variable to time required to produce the first 10 mL of gas, while keeping the original investigation question unchanged.

Evaluate the proposal. Calculate the average gas production rates up to 10 mL, and explain what the student must do before final data collection.

Assume the reaction apparatus is gas-tight and the gas volumes are measured under comparable conditions.

Solution 3

Measuring the time to collect the first 10 mL is a potentially feasible alternative, but it changes the dependent variable and therefore requires the investigation question and method to be revised.

The original investigation measures gas volume at a fixed time of 30 s.

The revised investigation measures time to reach a fixed gas volume.

Those are different measurements.

Calculate the average rates.

The average gas production rate up to the 10 mL endpoint is:

\[ r=\frac{V}{t} \]

Here, \(r\) is average volume produced per second in mL/s, \(V\) is the collected volume in mL, and \(t\) is the elapsed time in seconds.

For the low-concentration condition:

\[ \begin{aligned} r_{\mathrm{low}}&=\frac{10}{42}\\ &\approx0.238\,\mathrm{mL/s} \end{aligned} \]

For the high-concentration condition:

\[ \begin{aligned} r_{\mathrm{high}}&=\frac{10}{6}\\ &\approx1.67\,\mathrm{mL/s} \end{aligned} \]

The high-concentration condition has a greater average gas production rate over the first 10 mL.

However, these values are averages over different time intervals. They do not establish the instantaneous reaction rates at a particular moment or the total gas each reaction would eventually produce.

Evaluate feasibility.

Both reactions reach the 10 mL threshold, so the endpoint is physically achievable using the existing syringe.

But the high-concentration reaction reaches the endpoint in only 6 s.

If the stopwatch timing uncertainty were approximately \(\pm 0.5\) s, the corresponding relative timing uncertainty would be:

\[ \frac{0.5}{6}\times100\%\approx8.3\% \]

The student should check whether this is acceptable for the intended comparison and whether a more precise, approved timing method is available.

Revise the investigation.

The student should:

• Rewrite the question to investigate the effect of concentration on the time taken to collect a fixed gas volume, or on the average gas production rate up to that volume.
• Keep the reaction quantities, apparatus, starting conditions, and endpoint definition consistent.
• Perform a short trial to verify timing repeatability at both concentrations.
• Obtain any required teacher approval for the modified investigation.

The tempting mistake is to treat the revised endpoint as just an equipment adjustment. It solves the syringe-capacity problem by changing what is measured, so that change must be acknowledged.

Question 4

A student investigates how temperature affects gas production by yeast in a sugar solution.

A gas syringe records the volume produced over ten minutes.

The pilot gives:

All trials use separate yeast samples and the same quantities of yeast and sugar.

The student predicts that gas production will increase as temperature increases, so they want to remove the 50°C condition and restrict the investigation to 20-35°C.

Another student argues that the near-zero result at 50°C is scientifically important.

Explain whether either conclusion is justified from the pilot alone. Propose a short follow-up pilot that can distinguish an apparatus problem from a genuine low response.

Solution 4

The 50°C condition should not be discarded merely because it contradicts the predicted trend, but the pilot does not yet establish that gas production genuinely stops at that temperature.

The recorded result means the gas volume was below the instrument’s readable limit. It does not prove that exactly zero gas was produced.

Several explanations remain possible.

The yeast may have very low metabolic activity under those conditions. Alternatively, a leak, an unsuccessful preparation, or a measurement limitation could produce a similar observation.

The first student’s proposed range change is tempting because the remaining data show a clear increase from 20°C to 35°C.

But choosing a range simply to preserve an expected trend would remove potentially meaningful evidence.

The second student is right that the result deserves investigation, but cannot yet claim it demonstrates complete biological inactivity.

A useful follow-up pilot would include these checks:

1. Verify the gas-collection system by introducing a known, safe test volume into the apparatus using an approved procedure and checking whether it is recovered.
2. Repeat the 50°C condition using a fresh yeast preparation and consistent timing.
3. Repeat the 35°C condition as a positive comparison to establish that the preparation can produce measurable gas under the experimental conditions.
4. Confirm that the yeast mixture actually reaches and remains near the intended temperature, rather than assuming the water-bath setting equals the sample temperature.

If the apparatus passes its checks and repeated 50°C trials remain below the detection limit, the student has stronger evidence that gas production is very low under that condition.

The investigation may then retain 50°C as a scientifically meaningful condition, reporting the result as below the readable limit rather than as an exact zero.

If the apparatus fails its checks, it must be corrected and the affected measurements repeated.

The deeper point is that an inconvenient result and an invalid measurement are not the same thing. A pilot helps distinguish them before the final experiment.

Question 5

A student investigates whether changing the concentration of a coloured solution affects the amount of light transmitted through it.

A light source shines through a transparent cuvette containing the solution. A sensor on the opposite side measures transmitted light in arbitrary units.

The sensor’s usable range is 0-100 units, with a resolution of 1 unit.

The solution also becomes darker as a slow reaction proceeds after preparation. This means the sensor reading may depend on both concentration and the age of the sample.

Every sample is freshly prepared, each is measured once, and temperature, cuvette geometry, and light-source conditions are controlled.

The first pilot measures:

• Dilute solution at 1 minute after preparation: 80 units.
• Concentrated solution at 4 minutes after preparation: 40 units.

The student concludes that increasing concentration reduces the sensor reading by 40 units.

Another student recognises the timing problem and conducts two additional pilot measurements:

• Dilute solution at 4 minutes: 60 units.
• Concentrated solution at 1 minute: 60 units.

The second student now argues that concentration has no effect, because the two additional readings are equal.

Tasks:

(a) Determine whether either student’s conclusion is supported. Quantify the effects of concentration and sample age at the tested conditions.

(b) Explain why comparing the two additional readings directly gives the wrong conclusion.

(c) The student plans to measure final samples exactly 2 minutes after preparation. Can the existing pilot data establish the exact expected readings at 2 minutes?

(d) Propose the most important final pilot check and explain why simply randomising the order of measurements would not necessarily solve the problem.

Solution 5

Neither student’s conclusion is correct. The four pilot readings show a concentration effect of 20 sensor units and an ageing effect of 20 sensor units across the tested conditions. They do not establish the exact reading at 2 minutes.

The difficulty is that the original trial changed two things together: concentration and elapsed time.

Part (a): Separate the two effects.

Arrange the pilot readings so that concentration can be compared at the same sample age.

First compare concentrations at the same time.

At 1 minute:

\[ 60-80=-20\ \text{units} \]

At 4 minutes:

\[ 40-60=-20\ \text{units} \]

Increasing concentration reduces transmitted-light readings by 20 units at both tested ages.

Now compare sample ages at the same concentration.

For the dilute solution:

\[ 60-80=-20\ \text{units} \]

For the concentrated solution:

\[ 40-60=-20\ \text{units} \]

From 1 minute to 4 minutes, the sensor reading decreases by 20 units at both tested concentrations.

The results are consistent with both factors affecting the reading.

The original 40-unit difference contained both effects.

Part (b): Explain the apparently contradictory result.

The additional measurements were:

• Dilute at 4 minutes: 60 units.
• Concentrated at 1 minute: 60 units.

The equal readings are not evidence that concentration has no effect.

The concentrated sample was measured earlier, when samples generally transmit more light. The dilute sample was measured later, when samples transmit less light.

Under the tested conditions, these two influences offset one another.

This is the conceptual trap: two measurements can be equal even when the independent variable genuinely affects the outcome.

To isolate the effect of concentration, sample age must be held constant.

The student’s second comparison changes concentration and age in opposite directions, so it cannot isolate either factor.

Part (c): Can the student calculate the exact reading at 2 minutes?

No.

The measurements establish differences at 1 minute and 4 minutes, but they do not establish how the solution changes at every intermediate time.

For example, the colour might change approximately steadily throughout the interval. Alternatively, most of the change might occur in the first minute or just before the fourth minute.

Both patterns could produce the same measured endpoints.

It would be tempting to assume a straight-line change and interpolate between the readings. But the pilot has not established that the time dependence is linear.

Even the equal 20-unit concentration differences at the measured endpoints do not prove the effect must remain identical at every intermediate age.

The student therefore cannot claim an exact expected reading at 2 minutes without an additional assumption or measurement.

Part (d): Design the final pilot check.

The most important next step is to prepare fresh dilute and concentrated samples and measure both at the intended final age of exactly 2 minutes.

The student should repeat these measurements to check whether the readings are consistent and remain comfortably within the sensor’s usable range.

The final procedure should standardise:

• The elapsed time between preparation and measurement.
• The preparation method and solution quantities.
• The cuvette orientation and light path.
• The sensor’s baseline and measurement procedure.
• The temperature and other relevant experimental conditions.

Randomising measurement order can reduce certain sequence-related biases, but it does not guarantee equal sample ages.

If the concentrated samples are sometimes measured 1 minute after preparation and sometimes after 4 minutes, the time-dependent colour change remains a source of variation.

Random order alone does not remove that problem.

The strongest revision is to control sample age directly, or explicitly measure and account for it using a validated method.

The pilot’s most important finding is therefore not the apparent 40-unit difference, or even the two separate 20-unit differences.

It is that the original measurement procedure could not isolate the variable being investigated.

Once the student controls measurement timing and verifies that the sensor works over the chosen range, the final investigation can compare concentration effects without confusing them with changes occurring as the sample ages.