How to perform in the new G2 SEC Science examination includes recognising that the order of trials can matter. A first measurement can heat an apparatus, use up a reactant, leave residue, alter a biological sample, change a battery, adapt a sensor or simply make the experimenter faster at the procedure. If the next trial begins from a changed state, the data may reflect both the intended variable and a hidden carryover from what happened before.
This hundred-and-third Learner’s Guide focuses on sequence and carryover effects. The central rule is: before comparing later trials, ask whether the earlier trial changed the starting state. Repeating a procedure is not enough if the apparatus, material, sample or operator is no longer comparable.
For 2027, SEAB lists G2 Science as K223 Science (Physics, Chemistry), K224 Science (Physics, Biology) and K225 Science (Chemistry, Biology). Use the official G2 syllabus directory for current subject documents. The examples below are original eduKateSengkang teaching cases, not official specimen questions or instructions to carry out experiments at home.
Sequence effect
A sequence effect occurs when the order in which conditions are tested influences the measurements.
The condition tested second may begin from a different state from the condition tested first.
Carryover effect
Carryover is the influence of an earlier trial on a later one.
Residue, heat, charge, depletion, fatigue, learning, adaptation or contamination can carry information or physical state forward.
Why this matters
If the intended comparison is A versus B but trial B inherits something from trial A, the comparison is no longer only A versus B.
The learner should identify the inherited factor as a possible alternative explanation.
Same apparatus does not mean same condition
Using the same apparatus can improve comparability, but only if the apparatus is restored to a comparable starting state.
A warm beaker, partially discharged battery or wet filter may behave differently from its original state.
Same sample does not mean independent trial
Repeated measurements on the same biological or material sample may be influenced by the previous treatment.
The sample itself may have changed.
Same operator does not mean neutral sequence
The person carrying out the method can learn, tire, anticipate the endpoint or change timing with practice.
Operator state can therefore become part of the sequence effect.
Physics case: warming apparatus
A device is tested first at high power and immediately afterwards at low power.
If the apparatus remains warm, the second measurement may start from a different thermal state. The order could affect the comparison.
Physics case: battery discharge
A circuit is tested under several loads using the same battery in sequence.
If the battery state changes over time, later trials may not begin with the same source condition.
Physics case: sensor lag
A temperature sensor moved from hot water to cooler water may take time to stabilise.
A reading taken too soon can carry the previous temperature state into the next measurement.
Chemistry case: residual substance
An apparatus used for one solution may retain a small amount before the next trial.
If not appropriately cleaned or reset, the second condition can be contaminated by the first.
Chemistry case: reactant depletion
If the same reacting system is used repeatedly, a reactant may be partly consumed.
Later rates or yields may differ because starting amounts are no longer comparable.
Chemistry case: wet apparatus
A container rinsed with water but not restored appropriately may dilute the next solution.
The effect depends on the method and the quantity being measured. The learner should identify the actual consequence rather than writing “contamination” generically.
Biology case: plant response
A leaf or organism exposed to one condition may need time to return to baseline before another condition is tested.
The first exposure can influence the second response.
Biology case: fatigue
Repeated physical performance measurements on the same organism may decline because of fatigue rather than the intended experimental variable.
Rest time or separate comparable samples may be needed depending on the question.
Biology case: acclimatisation
An organism exposed to a condition may adapt temporarily.
A later trial can therefore reflect both the new condition and the organism’s prior experience.
The starting-state question
Before each trial, ask: Is the system in the same relevant starting state as before?
This question is more precise than simply asking whether the same apparatus is being used.
The reset question
If the earlier trial changes the system, what reset is needed before the next trial?
- cool to a defined starting temperature;
- recharge or replace a source;
- rinse and prepare apparatus appropriately;
- replace consumed materials;
- allow sensor stabilisation;
- provide rest or recovery time;
- use a fresh comparable sample.
The correct reset depends on the mechanism of carryover.
The order-randomisation idea
If order itself could bias results and several conditions are being compared, varying or randomising the order can help separate condition effects from sequence effects.
At G2 level, the important principle is not sophisticated experimental design terminology. It is recognising that always testing A before B can confound condition with order.
Counterbalancing idea
A simple design can test some samples in A→B order and others in B→A order.
If results change with order, sequence may be influencing the outcome.
Fresh-sample design
Using a fresh comparable sample for each condition can avoid some carryover effects.
This can introduce sample variation, so comparability and sample size still matter.
Reset-and-reuse design
Reusing the same system can reduce sample differences if the system can genuinely be returned to baseline.
The method should explain how that reset is achieved.
The order-confounding pattern
If all low values are measured early and all high values late, time and condition change together.
A trend may then reflect drift over time rather than the intended independent variable.
Instrument drift
An instrument can change calibration or response over time.
If all conditions are measured in one fixed sequence, instrument drift can mimic a condition effect.
Environmental drift
Room temperature, light or humidity may change during a long session.
Later trials can differ because the environment changed, even if the written method stayed the same.
Operator learning
The experimenter may become faster or more consistent with practice.
If one condition is always tested first, it may receive less practised technique than later conditions.
Operator fatigue
The opposite can also occur: later trials may be less careful because the operator becomes tired.
Sequence effects can therefore work in more than one direction.
The order-versus-condition question
When two trials differ, ask whether the difference aligns with the intended condition or simply with which trial came first or last.
This is the core diagnostic question.
The sequence-effect evidence
One clue is that reversing order changes the result.
Another is that later trials drift consistently regardless of condition.
The carryover signature
Carryover often produces a later result that depends on the previous state.
For example, the same condition may give different results when preceded by A versus preceded by B.
Do not invent carryover
Not every repeated experiment has a carryover problem.
The learner should identify a plausible mechanism linking earlier trial to later measurement.
The mechanism requirement
A strong evaluation explains how the earlier state affects the measured outcome.
“There may be carryover” is weaker than “the sensor remains warm, so the next temperature reading may begin too high”.
The consequence requirement
State the likely direction or uncertainty created where possible.
Residue may raise or lower concentration; a warm apparatus may raise a later temperature; fatigue may reduce later performance.
The targeted-improvement requirement
The improvement should remove or measure the sequence effect.
Reset, replace, rest, randomise order or use fresh samples according to the actual mechanism.
Practice clinic one: warm sensor
A temperature probe is used in hot water and immediately moved into cooler water. The first reading in the cooler water is higher than later readings.
Possible carryover: the probe itself remains warm. Improvement: allow the probe to reach a stable reading or use an appropriate reset procedure before recording.
Practice clinic two: reused solution container
A container holds solution A, is emptied, then receives solution B without appropriate cleaning.
Possible carryover: residue from A changes the composition of B. The improvement should address the residue rather than simply repeat the contaminated sequence.
Practice clinic three: battery sequence
A battery powers three circuit conditions in order. The third condition always gives a lower reading.
Possible sequence effect: battery state changes across time. Reverse or vary order, use a suitable fresh source or monitor source condition depending on the task.
Practice clinic four: fatigue
An organism performs the same physical task under three conditions without rest. Performance decreases across trials.
The decline may reflect fatigue as well as the intended condition. Rest periods, separate comparable samples or varied order can help distinguish the effects.
Practice clinic five: learning
A human participant performs a reaction task repeatedly and becomes faster.
Later conditions may benefit from practice. Sequence and condition are confounded if every participant experiences the same order.
Practice clinic six: adaptation
A plant sample is exposed to low light, then immediately to high light, and the second response is compared with a fresh high-light sample.
If the first exposure changes the sample’s state, the reused sample may not be directly comparable with the fresh one.
Practice clinic seven: residue in measuring apparatus
A measuring cylinder contains traces of a concentrated solution before a dilute solution is measured.
Residual material may change concentration. The consequence should be stated specifically where the direction can be inferred.
Practice clinic eight: wet apparatus
A dry solid is weighed in a container that is still wet from cleaning.
Water adds mass and may also interact with the sample depending on context. The starting state matters.
Practice clinic nine: heating sequence
A metal sample is tested at several temperatures by increasing temperature step by step without returning to baseline.
If a previous high-temperature exposure changes the material or apparatus, later readings can inherit that change. Whether this matters depends on the process.
Practice clinic ten: order and time drift
A set of samples is tested A, B, C, D over one hour while room temperature rises gradually.
Condition and time are aligned. If outcomes also rise, room-temperature drift becomes an alternative explanation.
The fixed-order trap
Testing every condition in the same order makes order inseparable from condition.
A different order across repeats can reveal whether the sequence itself matters.
The reset-is-not-a-word rule
Do not write “reset the apparatus” without saying what reset means.
Cool, rinse, dry, recharge, replace, rest or recalibrate according to the mechanism.
The fresh-sample trade-off
Fresh samples can reduce carryover but may introduce more sample-to-sample variation.
The design should balance these risks according to the scientific question.
The reuse-sample trade-off
Reusing the same sample controls some between-sample differences but increases the risk of history effects.
The learner should recognise both sides rather than treating one design as always superior.
The washout-period idea
Some systems need time between conditions for the earlier effect to fade.
At G2 level, the key idea is that a waiting period may restore a comparable baseline if the process is reversible.
The baseline check
Measure or verify a baseline before each trial where appropriate.
If the baseline has shifted, the new trial should not be treated as starting from the same condition.
The time-order table
During practice, record trial number, condition, starting state and result.
A pattern linked to trial number rather than condition is a clue that sequence may matter.
The order-reversal drill
Run a hypothetical A→B sequence and B→A sequence.
Ask whether the result attributed to B changes depending on what came before. If it does, carryover is plausible.
The fresh-versus-reused drill
Compare data from fresh samples with data from the same sample reused across conditions.
Identify which design better answers the intended question and what new limitation each introduces.
The reset-quality drill
Provide several reset suggestions. The learner chooses which actually restores the relevant starting state.
For example, wiping a warm sensor dry does not cool it. Waiting for stabilisation addresses a different problem from cleaning.
The sequence-confound drill
Give a data table where condition A is always first and B always second.
Ask what alternative explanation exists if all second-trial values are higher.
The drift-versus-condition drill
Give measurements from several conditions taken over time while an environmental factor changes.
Ask how varying the condition order could help separate condition effect from time drift.
The carryover-mechanism drill
For every proposed carryover effect, require the chain: earlier trial → changed starting state → altered later measurement.
If the learner cannot state the chain, the carryover claim may be speculative.
The carryover error ledger
- same apparatus assumed to mean same state;
- same sample treated as independent repeat;
- fixed order confounds condition with time;
- reset suggested without mechanism;
- fresh sample used without considering sample variation;
- operator practice or fatigue ignored;
- sensor lag ignored;
- residue or depletion ignored;
- carryover invented without plausible pathway.
These categories give experiment evaluation a precise target.
Sequence effects and control conditions
Use Vol 0099. Order can be another alternative explanation that a control strategy must address.
Sequence effects and repeatability
Use Vol 0091. Repeating a fixed biased order can reproduce the same sequence effect consistently.
Sequence effects and data ownership
Use Vol 0075. Every measurement belongs not only to a condition but also to a place in the sequence and a starting state.
Sequence effects and model-versus-measurement
Use Vol 0044. A sequence effect can create systematic deviations that look like model failure if order is ignored.
The sequence-audit checklist
- What condition came before this trial?
- Could the previous trial change apparatus, sample, source, sensor or operator?
- Was a baseline restored?
- Was order varied or fixed?
- Could time drift explain the trend?
- Would reversing order change the interpretation?
These six questions cover most high-value sequence problems.
The 20-minute sequence session
- Five minutes: identify possible carryover mechanisms.
- Five minutes: design resets.
- Five minutes: compare fixed and varied order.
- Five minutes: write bounded conclusions.
This can be adapted to the learner’s actual G2 Science combination.
A four-week sequence-effect build
Week 1 — recognise changed starting states
Heat, charge, residue, depletion, fatigue and adaptation.
Week 2 — reset and fresh samples
Match the reset to the mechanism and compare design trade-offs.
Week 3 — order variation
Use A→B, B→A and time-drift scenarios.
Week 4 — timed experiment evaluation
Identify sequence as an alternative explanation and propose a targeted improvement.
Use the Science index
If the learner cannot judge the carryover because the underlying process is missing, return to the Complete Science Index.
The PSLE bridge
The earlier rule Evidence Before Explanation remains central.
At G2, the learner also asks whether the evidence itself was shaped by what happened before it was measured.
Use Examination Craft
For timed structured questions, continue through the Examination Craft hub.
Final rule
A later trial is not automatically a fresh trial.
Ask what the previous trial changed. Restore the baseline when possible, vary order when order can bias results, and use fresh samples when reuse creates history effects. A fair comparison begins not only with the same written method, but with a genuinely comparable starting state.
Carryover clinic: thermal memory
A metal block is heated for Trial A and then used immediately in Trial B, which is meant to begin at room temperature.
The block carries thermal history into Trial B. A valid reset would verify the starting temperature rather than assume enough time has passed.
Carryover clinic: chemical residue
A pipette transfers concentrated solution, then is used for a dilute solution without suitable preparation.
Residual concentrated solution can alter the second sample. The effect is not simply “dirty apparatus”; the carried substance changes composition.
Carryover clinic: depletion
The same source material is used across several reaction-rate trials.
If material is consumed, later trials may begin with less reactant. Repetition without replacement changes the starting quantity.
Carryover clinic: sensor saturation
A sensor exposed to an extreme condition takes time to return to its baseline response.
A later reading taken too soon may reflect sensor recovery as well as the new condition.
Carryover clinic: participant practice
A participant performs a response task under three conditions. Accuracy improves simply because the task becomes familiar.
If every participant experiences conditions in the same order, practice effect can be mistaken for a condition effect.
Carryover clinic: participant fatigue
The same participant performs repeated trials with increasing physical effort.
Later performance may fall because of fatigue. Rest or varied order can help separate fatigue from treatment.
Carryover clinic: biological adaptation
A sample is exposed to a strong stimulus and then tested under a weaker stimulus.
If the first exposure changes responsiveness, the second measurement is not a clean response to the weaker condition alone.
Carryover clinic: contamination by cleaning
Cleaning can itself create carryover if a cleaning fluid remains and changes the next condition.
A reset must remove both the original material and any reset-related influence relevant to the measurement.
Carryover clinic: calibration drift
An instrument is calibrated once at the beginning of a long sequence and slowly drifts.
Later conditions are systematically measured with a different instrument state. Checking calibration during the sequence may be necessary depending on the task.
Carryover clinic: environmental drift
A room warms over an afternoon while conditions are tested in fixed order.
Temperature becomes linked to trial number. If outcomes also rise, sequence and environment are confounded.
The baseline definition
A baseline is the relevant starting state before a trial.
It may be temperature, charge, concentration, resting performance, instrument zero or another condition. The right baseline depends on what previous trials can change.
Baseline verification
Do not assume the baseline has returned just because the clock advanced.
Where possible, verify the relevant quantity directly or use a justified waiting/reset procedure.
The state-variable idea
Carryover is easiest to understand by naming the hidden state that changes across trials.
- temperature;
- battery charge;
- residual concentration;
- sample fatigue;
- sensor response;
- operator experience;
- environmental condition.
The hidden state creates the pathway from previous trial to later result.
The order-effect table
During practice, add columns for trial number, condition, baseline state and result.
If results track trial number more closely than condition, order deserves investigation.
The sequence-as-variable idea
If all A trials occur before all B trials, sequence itself has become a hidden variable.
Varying order across repeats can make that variable visible.
The AB/BA design
A simple counterbalanced comparison uses some sequences A→B and others B→A.
If B performs differently depending on whether it follows A, carryover is supported as a possible explanation.
The washout logic
A washout or recovery period is useful only if the previous effect can reasonably fade with time.
If the previous trial permanently changes the sample, waiting is not enough; a fresh sample may be required.
The fresh-sample logic
Fresh samples remove some history effects but create between-sample variation.
The learner should evaluate which source of variation is more important for the question.
The same-sample logic
Reusing the same sample controls some sample-to-sample differences.
It works well only when the sample can return to a comparable starting state.
The random-order idea
Randomising order can prevent a systematic relationship between condition and time/order.
It does not guarantee every other variable is controlled, but it reduces one predictable source of bias.
The order-balanced idea
Balanced order deliberately ensures that conditions appear in different sequence positions across trials or samples.
The teaching point is to separate condition from sequence, not to memorise a particular design name.
The carryover-versus-drift distinction
Carryover comes from the previous trial affecting the next. Drift can occur gradually over time even without direct transfer from one trial.
Both can create sequence patterns but may need different fixes.
The carryover-versus-contamination distinction
Contamination is one specific carryover mechanism involving transferred material.
Carryover can also be thermal, electrical, biological or behavioural.
The carryover-versus-learning distinction
Learning or practice is a behavioural sequence effect.
The participant’s internal state changes even when apparatus and environment are unchanged.
The carryover-versus-fatigue distinction
Fatigue changes later performance because effort accumulates.
Practice may improve later performance; fatigue may reduce it. Both can coexist.
The plausible-pathway rule
Do not write “order effect” merely because one condition came later.
State the pathway: previous trial changes X, X affects later measurement Y.
The direction-of-effect rule
Where possible, predict how carryover should shift the later measurement.
A warm sensor may bias a later cool reading high; a depleted battery may lower later output; fatigue may lower performance. Direction makes evaluation stronger.
The no-direction caution
If the direction is unclear, state uncertainty rather than inventing one.
A valid limitation can still be identified even when the sign of the bias is unknown.
The sequence-check drill
Give a table of results ordered A, B, C, A, B, C. Ask whether drift could affect later cycles.
Then reorder hypothetically and predict what evidence would distinguish drift from condition.
The baseline-reset drill
Provide several reset methods and ask which state variable each restores.
Learners should reject resets that do not target the mechanism.
The fresh-sample trade-off drill
Compare a reused-sample design with a fresh-sample design.
List one advantage and one limitation of each before deciding which better answers the aim.
The carryover evidence drill
Give results from A→B and B→A sequences.
If B changes depending on order, ask what additional evidence would strengthen the carryover explanation.
The method-rewrite drill
Take a flawed fixed-order method and rewrite only the part needed to control sequence.
This teaches targeted improvement rather than replacing the entire procedure.
The sequence-effect error ledger
- baseline assumed rather than checked;
- reset does not restore relevant state;
- fresh sample introduced without considering sample variation;
- fixed order ignored;
- drift confused with carryover;
- carryover claimed without mechanism;
- operator practice/fatigue ignored;
- direction of bias invented without evidence.
These categories make evaluation more precise.
The final sequence-effect checklist
- What was tested immediately before?
- What state could it have changed?
- Was that state reset or replaced?
- Was order fixed or varied?
- Could time drift be responsible?
- Would reversing order change the interpretation?
Six questions are usually enough to expose a hidden history effect.
Advanced standard
An advanced G2 Science learner knows that every measurement has a history.
They can distinguish a genuinely fresh trial from a repeated procedure whose starting state has changed. They evaluate not only what the method says will be controlled, but whether the system can realistically return to the intended baseline.
Final perspective
Sequence is part of experimental design whenever the past can affect the present.
A good comparison controls not only the listed variables but also the history carried from one trial to the next. Reset what changed, vary order when order can bias results, and use fresh samples when the previous condition leaves a lasting state.
Final sequence-effect practice: separate condition from history
Take a three-condition investigation and imagine all trials are run in A→B→C order. Now list every state that might drift with trial number: temperature, battery charge, residue, operator practice, fatigue, sensor baseline or environmental conditions. Then ask which of those states could plausibly affect the dependent measurement.
Next, imagine the order is reversed or varied. If the result attributed to B changes depending on whether B comes first or second, history has become part of the evidence. The correct conclusion may need to distinguish condition effect from sequence effect rather than choosing one automatically.
Carryover after an apparently successful reset
A reset should be evaluated by the state it is meant to restore. Rinsing can remove residue but may leave water. Waiting can reduce thermal carryover but cannot replace a consumed reactant. Rest can reduce fatigue but may not reverse long-term adaptation. “Reset” is therefore not one universal action.
During evaluation, state the state variable and the evidence that it has returned to baseline. This makes the improvement testable rather than verbal.
Order effects and replication
Repeating the same fixed order can reproduce an order bias very consistently. Meaningful replication should therefore consider whether order is part of the hidden structure. If A is always first and C always last, apparent condition differences may contain sequence information.
Varying or balancing order across trials can reveal whether the pattern belongs to the condition itself or to where that condition appears in the sequence.
The final carryover audit
- What changed in the previous trial?
- Could that state persist?
- How would persistence affect the next measurement?
- Was the state actually restored?
- Was order varied enough to detect a history effect?
- Would a fresh sample reduce carryover more effectively?
These questions keep sequence effects connected to a plausible mechanism.
The advanced standard
An advanced learner treats each trial as having a starting state and a history. They recognise that a written method can look identical while the physical, chemical, biological or operator state has changed underneath it.
Good experimental reasoning therefore asks not only “what condition is this?” but also “what happened immediately before it?”
One last sequence-effect rule
When a later result looks unusual, compare it with trial order before blaming the condition itself. If the same shift appears whenever a condition occurs late—regardless of which condition it is—time, drift, fatigue or apparatus history becomes a stronger alternative explanation.
Likewise, if reversing the order changes the apparent effect, the learner should not average the two orders blindly and declare the problem solved. The order dependence is itself evidence that history matters and should be explained or controlled.
The final standard is a genuinely comparable starting state. Same written instructions are not enough if the system entering Trial 2 is physically, chemically, biologically or behaviourally different from the system that entered Trial 1.
Final calibration: when order evidence is strong enough
One difference between first and second trials does not automatically establish a sequence effect. Stronger evidence appears when the same condition changes systematically with order, when reversing order changes the result, or when the proposed carryover mechanism can be linked to a measurable starting-state difference.
Likewise, absence of an order difference in one small comparison does not prove carryover is impossible. The method may simply not be sensitive enough to detect it. Sequence reasoning should remain proportional to the evidence.
In evaluation questions, state the mechanism, the likely consequence and the design change that would test or reduce it. This moves the answer beyond “order may matter” into a scientific explanation that can be checked.
The final sequence-effect habit is to record the starting state explicitly whenever history could matter. A simple note such as “cooled to room temperature”, “fresh sample”, “battery replaced”, or “participant rested” makes the intended reset visible and testable.
That one step helps separate true condition effects from changes caused by what happened earlier in the sequence.
Sequence control is complete when the learner can describe the intended condition and the inherited state separately. Only then can a later measurement be interpreted as evidence about the condition rather than evidence about what happened earlier.
A final comparison is trustworthy only when the system enters each trial from a comparable relevant state. If that state cannot be restored, the design should acknowledge history explicitly rather than pretending the later trial is fresh.
History matters.
