How to Choose a Useful Range and Spacing of Test Conditions in a PSLE Science Investigation

Wait, What? Five Test Conditions Can Still Be Five Bad Test Conditions

A learner wants to investigate how the amount of light affects a measured plant outcome. They choose five light conditions:

20, 21, 22, 23 and 24 units.

The investigation is careful. The same type of plant is used. Other relevant conditions are kept comparable. The outcome is measured properly. Yet the results look almost the same.

The learner concludes: “Light does not affect the outcome.”

That conclusion may be too strong. The five conditions may simply be too close together to reveal a meaningful difference.

Now imagine the learner uses only two conditions: 0 and 100 units. A large difference appears. Have they mapped the relationship well?

Not necessarily. The two extreme points might show that the factor matters, but they could miss a threshold, plateau, turning point or best-performing region between them.

A good PSLE Science investigation does not merely have “more test conditions”. The chosen conditions must cover a useful scientific range and be spaced in a way that can reveal the relationship the question is trying to investigate.

This is the learner job of test-condition design. It is different from deciding how often to measure over time. It is different from repeating the same trial. It is different from interpreting a graph after the data have already been collected. It happens earlier: when you decide which values of the changed factor should actually be tested.

Quick Answer

When choosing test conditions in a PSLE Science investigation, first identify the deliberately changed factor and the measured outcome. Then choose a range that is broad enough to reveal a scientifically meaningful change but still appropriate and safe for the investigation. Within that range, choose enough test values to show the pattern without collecting many nearly identical conditions that add little information.

A useful design usually asks:

What factor am I changing?
What outcome am I measuring?
What range of the changed factor is scientifically appropriate?
Are my values spread widely enough to reveal a difference?
Are there enough values to distinguish a trend from one comparison?
Could a threshold, plateau or turning region lie between my chosen values?
Is the spacing smaller than the measuring method can meaningfully distinguish?
Have I kept the original scientific question unchanged?

DEFINE THE CHANGED FACTOR → DEFINE THE MEASURED OUTCOME → CHOOSE A SCIENTIFICALLY APPROPRIATE RANGE → CHOOSE INFORMATIVE SPACING → KEEP OTHER RELEVANT CONDITIONS COMPARABLE → COLLECT THE EVIDENCE → INSPECT THE PATTERN → REFINE THE RANGE OR SPACING ONLY IF THE QUESTION NEEDS MORE DETAIL.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one learner job: how a Primary 5 or Primary 6 learner chooses a useful range and spacing of values for the deliberately changed condition in a PSLE Science investigation so the resulting evidence can reveal the relationship being tested.

It does not replace the guide on variables and fair tests. It does not replace the guide on whether an investigation needs more repeats or more test conditions. It does not replace the guide on choosing measurement intervals over time. It does not replace the page on finding the best tested condition.

This page owns the design decision that comes after you know which factor should change but before you collect the data:

Which values of that factor should I actually test so the investigation has a fair chance of revealing the scientific relationship?

Why This Matters in the 2026 PSLE Science Frame

For examination from 2026, PSLE Science assesses attainment in the 2023 Primary Science syllabus. The official assessment objectives include knowledge with understanding, application of scientific facts, concepts and principles, and scientific inquiry involving prediction or hypothesis, interpretation and analysis of information, evaluation of observations, information and methods, and communication of explanations and reasoning.

Choosing useful test conditions belongs to scientific inquiry because the method determines what evidence the investigation is capable of producing. A weak set of conditions can hide a relationship even when the learner controls the variables carefully.

First Distinction: Test Conditions Are Values of the Changed Factor

Suppose the scientific question is:

How does the height of a ramp affect the distance travelled by a toy car?

The ramp height is the deliberately changed factor. The actual heights chosen—perhaps 5 cm, 10 cm, 15 cm, 20 cm and 25 cm—are the test conditions.

The distance travelled is the measured outcome.

This matters because learners sometimes confuse three different decisions:

Decision	Example	Different learner job
Which factor changes?	Ramp height	Variable identification
Which values of that factor are tested?	5, 10, 15, 20, 25 cm	This guide: range and spacing
How often is the outcome measured?	Every 30 seconds	Measurement-interval design
How many times is each condition repeated?	Three trials at each height	Repetition / reliability

Keep these jobs separate before combining them.

What “Range” Means

The range of test conditions is the span from the lowest condition tested to the highest condition tested.

If you test 10°C, 20°C, 30°C, 40°C and 50°C, the tested range is 10°C to 50°C.

If you test 20°C, 21°C, 22°C, 23°C and 24°C, the range is only 20°C to 24°C even though you still have five test conditions.

Number of test conditions and width of tested range are different properties of an investigation.

What “Spacing” Means

The spacing is the gap between neighbouring test conditions.

For 10, 20, 30, 40 and 50, the spacing is 10 units.

For 10, 15, 25, 40 and 50, the spacing is uneven. Uneven spacing is not automatically wrong. It may be useful if the investigator deliberately tests more closely near a region where something important happens.

Equal spacing is often easy to interpret, but it is not a universal scientific law.

The Evidence Mechanism: Test Conditions Sample a Relationship

Imagine a scientific relationship as a road stretching across a landscape. Every test condition is one place where you stop and take a measurement.

If all your stops are clustered within the same 100 metres, you know that small region well but almost nothing about the rest of the road.

If you stop only at the two ends, you know the endpoints but not what happens between them.

A useful investigation places enough “stops” across the scientifically relevant region to answer the question being asked.

Worked Example 1 — Too Narrow a Range Can Hide a Relationship

Original practice setup: a learner investigates how water temperature affects the time taken for a stated change to occur.

They test 30°C, 31°C, 32°C, 33°C and 34°C.

The recorded times are 52 s, 51 s, 51 s, 50 s and 50 s.

The small differences may be comparable with ordinary measurement variation. The learner cannot confidently map the broader relationship from this narrow range alone.

If scientifically appropriate within the school investigation, a wider set such as 20°C, 30°C, 40°C, 50°C and 60°C may reveal a clearer pattern.

The point is not that “wider is always better”. The point is that the chosen values must be capable of producing informative contrasts.

Worked Example 2 — Two Extreme Conditions Can Miss the Shape

A learner tests only Condition 0 and Condition 100. The outcome is 10 at Condition 0 and 20 at Condition 100.

Can the learner conclude that the outcome increases steadily throughout the whole range?

No. The values in between were not measured.

The true relationship could:

increase steadily;
rise sharply at first and then level off;
remain flat until a threshold;
rise to a peak and then fall before rising again;
change irregularly for reasons the simple design does not capture.

Additional intermediate conditions help reveal the pattern.

Worked Example 3 — Equal Spacing Reveals a Trend

A toy-car investigation uses ramp heights of 5, 10, 15, 20 and 25 cm.

Original practice results:

Ramp height / cm	Distance travelled / cm
5	42
10	58
15	75
20	91
25	107

The equal 5 cm spacing makes it easy to compare how the measured outcome changes across equal steps in the tested factor.

But do not assume equal input steps must produce equal output changes. The data decide.

Worked Example 4 — Unequal Spacing Can Be Purposeful

A first investigation suggests that an important change happens somewhere between Conditions 30 and 40.

A follow-up might test 30, 32, 34, 36, 38 and 40 rather than repeating the entire broad range with the same coarse spacing.

The spacing is now tighter because the scientific question has become more precise: where in this region does the change occur?

This is refinement, not inconsistency.

Worked Example 5 — Finding a Threshold Region

Suppose a measured response is:

Condition	Response
10	No detected change
20	No detected change
30	Small change
40	Large change
50	Large change

The broad investigation suggests an important transition between 20 and 40.

If the aim is to locate that transition more precisely, a follow-up can test additional values such as 25, 30 and 35 while preserving the same changed factor, outcome and fair-test conditions.

Do not call one observed cut-off a universal threshold beyond the tested conditions. The improved spacing only refines the evidence.

Worked Example 6 — Finding a Turning Region

Suppose the outcome across Conditions 10, 20, 30, 40 and 50 is 4, 8, 12, 9 and 5.

The response rises and then falls. Condition 30 gives the highest recorded result among those tested.

If the investigation wants to locate the best-performing region more precisely, values around 30—perhaps 25, 28, 30, 32 and 35—may be more informative than adding new tests far away.

This is a good example of a two-stage design:

BROAD SCAN → IDENTIFY IMPORTANT REGION → FINE-SCALE FOLLOW-UP.

Worked Example 7 — A Plateau Region

Outcomes are 3, 7, 10, 10 and 10 across increasing conditions.

The broad pattern appears to level off. Before concluding that the scientific process has truly stopped changing, check:

Is the measuring method still capable of detecting further changes?
Are the values at the upper limit of the instrument?
Would closer spacing around the start of the plateau reveal where levelling begins?
Could another limiting factor explain the plateau?

Spacing helps map the region; it does not explain the mechanism by itself.

Worked Example 8 — Conditions So Close That the Instrument Cannot Distinguish Them

A measuring method records only whole centimetres. The learner chooses changed conditions expected to produce differences of about 0.1 cm.

The test-condition spacing may be scientifically real, but the measurement method is too coarse to reveal the resulting differences.

This is why test-condition design and measurement design must fit each other.

If neighbouring conditions produce effects smaller than the method can distinguish, adding more closely spaced conditions may create more data points without creating more usable evidence.

Range, Spacing and Resolution Must Work Together

Design feature	Main question	Failure if weak
Range	Did we test across a broad enough scientifically relevant span?	Relationship may remain hidden
Spacing	Are neighbouring conditions positioned to reveal the pattern?	Thresholds or turning regions may be missed
Number of conditions	Do we have enough points to distinguish a pattern from one comparison?	Overgeneralisation from too few points
Measurement resolution	Can the method distinguish the resulting outcome differences?	Different conditions look identical
Repeats / specimens	Can we judge whether the observed pattern is reasonably stable?	One unusual result may dominate

Do Not Confuse Test-Condition Spacing With Measurement Interval

This confusion is common because both use words such as “interval” and “spacing”.

Test-condition spacing asks:

Which values of the changed factor do I test?

Measurement interval asks:

How often do I measure the changing outcome over time?

Example:

Temperatures tested: 20°C, 30°C, 40°C, 50°C — test-condition spacing.
Temperature reading taken every 2 minutes during one cooling trial — measurement interval.

The scientific roles are different.

Do Not Confuse More Conditions With More Repeats

Testing 10, 20, 30, 40 and 50 tells you how the outcome behaves across different values of the changed factor.

Repeating Condition 30 five times tells you how consistently that one condition produces a result.

Different jobs:

MORE CONDITIONS → MAP THE RELATIONSHIP.
MORE REPEATS → CHECK STABILITY / VARIATION AT A CONDITION.

How Wide Should the Range Be?

There is no universal PSLE number. The useful range depends on the scientific context and the values the question provides or permits.

A good range should be:

relevant to the scientific question;
wide enough to have a fair chance of revealing a change;
within the safe and appropriate conditions of the investigation;
measurable with the available method;
not so extreme that a different process or failure mode replaces the relationship you intended to test.

The learner should never invent dangerous or unrealistic experimental extremes simply to “make the range wider”.

Equal Spacing: Useful, Not Sacred

Equal spacing can make a table or graph easy to interpret because each input step is comparable.

However, unequal spacing can be scientifically sensible when:

the first broad test has already identified an important region;
a threshold appears to lie within a narrow band;
a turning point or best-performing region needs refinement;
the available apparatus only permits certain practical values;
the scientific variable is categorical rather than numerical.

Do not invent the rule “all fair tests must use equal intervals”. Fairness is about controlling competing causes, not about forcing every set of values into equal arithmetic gaps.

Numeric Conditions Versus Categorical Conditions

Not every changed factor has numerical spacing.

Examples of categorical conditions include:

material type;
surface type;
presence or absence of a cover;
different shapes;
different plant structures.

You cannot say the “spacing” between plastic, glass and metal is equal or unequal in the same numerical sense.

For categorical factors, the design question becomes: have I selected comparison categories that meaningfully test the scientific question?

Too Many Conditions Can Also Be Weak

More conditions are not free. They require time, specimens, measurement and control.

Testing 50 almost-identical conditions may:

consume resources without improving the conclusion;
reduce the number of useful repeats available at each condition;
make procedures harder to keep consistent;
create apparent precision that the measuring method cannot support.

Scientific design is about informative conditions, not maximum condition count.

The Broad-Then-Refine Strategy

When little is known about the relationship, a sensible design often begins with a broad exploratory set of conditions.

Test several appropriately spread values.
Plot or compare the outcomes.
Identify an important region: threshold, plateau start, turning point or rapid-change zone.
Design a follow-up using closer spacing in that region.
Keep the scientific question and relevant controls intact.

This is stronger than using ultra-fine spacing everywhere from the beginning.

Range Design and the “Best Tested” Problem

If the best recorded result occurs at the highest tested condition, the range may not extend far enough to bracket the best region.

Example:

Condition	Outcome
10	4
20	7
30	11

30 is best among those tested, but the outcome is still rising at the boundary. If scientifically appropriate, extending the range can test whether the response continues to rise, levels off or turns.

This does not mean every investigation should chase an optimum. Extend the range only when the scientific question requires it.

Range Design and Extrapolation

Untested values remain untested.

A well-chosen range reduces the amount of guessing needed inside the region of interest. It does not authorise unlimited prediction outside the range.

If a question asks what might happen beyond the tested values, distinguish a prediction from an observation and state the model limits.

Range Design and Fair Tests

Good test-condition spacing cannot rescue a confounded experiment.

If ramp height increases and car mass also changes at each condition, a beautiful five-point graph still cannot isolate which factor caused the difference.

The hierarchy is:

RIGHT QUESTION → RIGHT CHANGED FACTOR → RIGHT MEASURED OUTCOME → FAIR COMPARISON → USEFUL RANGE AND SPACING → SUITABLE MEASUREMENT → REPEATS / SAMPLING AS NEEDED.

Range Design and Natural Variation

When living things are used, natural variation may be large enough to hide small differences between neighbouring conditions.

If plant heights naturally vary by several centimetres, changing a condition by a tiny amount that affects growth by only a fraction of a centimetre may be difficult to detect with one plant per condition.

The repair may involve:

a more informative spacing of conditions;
more suitable similar specimens;
repeated trials where appropriate;
a more suitable measuring method.

Do not assume spacing alone solves every evidence problem.

Range Design and Measurement Limits

The chosen test range should keep the outcome inside a useful measuring range where possible.

If every high condition makes a sensor hit its maximum display, then the test-condition range may extend into a region where the instrument cannot distinguish the outcomes.

Possible repairs include:

using a suitable measuring method with a wider range;
choosing conditions that keep the outcome within the measurable window if that still answers the scientific question;
stating the measurement limitation honestly if the method cannot be changed.

How to Choose Conditions for a Relationship Question

Suppose the question asks generally how X affects Y.

A strong starting plan is:

Choose a lower value of X that is scientifically appropriate.
Choose an upper value that is meaningfully different but still relevant and safe.
Add intermediate values across the range.
Prefer interpretable spacing unless there is a reason to concentrate values.
Check that the measuring method can distinguish changes in Y.
Keep other relevant factors comparable.

How to Choose Conditions for a Threshold Question

First bracket the change using a broad range. Then add closer values around the region where the response first appears or changes sharply.

Do not call the first observed change an exact universal threshold unless the evidence justifies that precision and scope.

How to Choose Conditions for a Plateau Question

Include values below, around and above the apparent levelling region. Make sure the instrument has not simply reached its maximum. If the response remains similar while the method can still detect further change, evidence for a plateau becomes stronger.

How to Choose Conditions for a Turning-Point Question

Use values on both sides of the suspected turning region. If the first broad scan shows the outcome rising then falling, refine the spacing near the reversal.

How to Choose Conditions for a Simple Comparison Question

Sometimes only two conditions are needed because the scientific question itself asks a specific A-versus-B comparison.

Do not add extra conditions merely because “more data is better”. The method should fit the question.

The Test-Condition Design Table

Design	What it may reveal	Risk
Two widely separated conditions	Clear A–B difference	Cannot map what happens between
Many tightly clustered conditions	Fine detail in a narrow region	May miss broader relationship
Several evenly spread conditions	General trend across range	May miss fine threshold/turning detail
Broad scan followed by close spacing	General pattern then refined region	Requires a second stage
Conditions beyond measurement range	Little extra usable evidence	Instrument saturation or non-detection

The Earliest-Weak-Link Diagnostic

Failure signature	Earliest weak link	Repair
“I tested five values, so the investigation is strong.”	Condition count confused with informative range	Inspect the span and scientific usefulness of the values
“20, 21, 22, 23, 24 showed little change, so X has no effect.”	Range too narrow	Use an appropriately wider range before generalising
“0 and 100 are enough to prove a steady trend.”	Endpoints treated as full relationship	Add intermediate conditions
“All conditions must be equally spaced.”	Convenient graphing rule treated as scientific law	Use equal spacing when useful; refine unevenly when scientifically justified
“Measure every minute to create more test conditions.”	Time sampling confused with changed-factor values	Separate measurement interval from test-condition spacing
“Use the widest possible range.”	Range divorced from relevance/safety/method	Choose an appropriate range that answers the question
“Use tiny spacing for maximum accuracy.”	Spacing smaller than useful measurement resolution	Match condition spacing to detectable outcome differences

Misconception Repair 1 — “More Conditions Always Mean Better Science”

Compare five conditions clustered within 1 unit with four conditions spread across the scientifically relevant range. Ask which design is more likely to reveal the relationship and why.

Misconception Repair 2 — “Equal Spacing Is Part of a Fair Test”

A fair test controls competing causes. Equal spacing is a data-design choice. It can help interpretation but does not create fairness by itself.

Misconception Repair 3 — “Wider Is Always Better”

Use a hypothetical investigation where extreme conditions trigger a completely different failure process or exceed the measuring range. The useful range is the one that answers the intended question, not the most extreme one imaginable.

Misconception Repair 4 — “Fine Spacing Gives More Precise Science Automatically”

If the outcome differences between neighbouring conditions are smaller than the measurement resolution or natural variation, finer spacing can produce more rows without more discriminating evidence.

Misconception Repair 5 — “You Need the Same Number of Conditions in Every Investigation”

A two-condition comparison, a five-point trend investigation and a follow-up threshold refinement can all be scientifically reasonable because they answer different questions.

The PSLE Science Range-and-Spacing Protocol

Write the scientific question.
Name the deliberately changed factor.
Name the measured outcome.
Identify what values of the factor are scientifically possible and relevant.
Choose a lower and upper test condition that create a meaningful range.
Add enough intermediate conditions to reveal the expected pattern.
Use equal spacing if it helps interpretation; use closer spacing where evidence justifies refinement.
Check that the measurement method can distinguish the resulting outcomes.
Keep other relevant factors comparable.
Decide whether repeats or more similar specimens are also needed.
After collecting data, inspect whether an important region needs a focused follow-up.
Keep conclusions within the tested evidence.

How This Appears in Multiple-Choice Questions

A PSLE-style practice item may ask which proposed set of conditions gives a better investigation of a relationship.

Do not automatically choose the option with:

the most values;
the widest numbers;
equal intervals;
the most equipment.

Instead ask:

Does this set actually vary the intended factor?
Is the range relevant?
Is it broad enough to reveal the relationship?
Are there enough intermediate points?
Can the measuring method distinguish the outcomes?
Does the option preserve the original scientific question?

How This Appears in Structured Inquiry Answers

A useful reasoning shape is:

The current test conditions cover only ______ to ______, so they may not reveal how ______ affects ______ across a useful range. Test additional values such as ______ while keeping ______ comparable. If a change appears around ______, use more closely spaced values in that region to refine the pattern.

This is a thinking scaffold, not an official compulsory answer format.

Original Practice Set 1 — Which Design Maps the Relationship Better?

Question: How does Condition X affect measured Outcome Y?

Design A: X = 20, 21, 22, 23, 24
Design B: X = 10, 20, 30, 40, 50

If the scientifically relevant range is 10–50 and the measuring method can distinguish the changes, Design B provides a better first broad scan. Design A may later be useful if evidence identifies 20–24 as an important region.

Original Practice Set 2 — Which Design Finds the Turning Region Better?

Previous data suggest the highest outcome lies near Condition 30.

Design A: 0, 25, 50, 75, 100
Design B: 24, 27, 30, 33, 36

For a follow-up aimed specifically at refining the turning region, Design B is more informative. For a first exploration of the whole relationship, Design A may have a different purpose.

Original Practice Set 3 — When Fine Spacing Adds Little

A measuring method can only distinguish outcome differences of about 5 units. Testing conditions 30.0, 30.1, 30.2 and 30.3 is unlikely to be useful if those small condition changes produce outcome differences far below 5 units.

The investigation needs a better match between the test-condition spacing and the evidence the measurement can resolve.

Original Practice Set 4 — Categories Instead of Numbers

Question: How does surface material affect the distance travelled by the same toy car from the same release condition?

The test conditions are wood, rubber and tile. Numerical spacing is meaningless here. The learner should instead ask whether those categories are suitable comparisons and whether other relevant factors—slope, starting point and car—are kept comparable.

Practice Sequence — From Coarse Mapping to Scientific Refinement

Stage 1: Give the learner three candidate ranges and ask which is scientifically useful.
Stage 2: Keep the range fixed but change the spacing.
Stage 3: Show an emerging threshold and ask where closer spacing belongs.
Stage 4: Show a turning point and design a focused follow-up.
Stage 5: Add a measurement-resolution constraint.
Stage 6: Add natural variation and ask whether repeats or more specimens are also needed.
Stage 7: Mix numerical and categorical changed factors.
Stage 8: Return after several days with an unfamiliar investigation and no checklist.

Unfamiliar Transfer Challenge

A mystery process depends on Condition Z. You know only this:

Z can be tested appropriately from 0 to 80 units.
The instrument can distinguish outcome differences of 2 units.
A first test at Z = 0 and Z = 80 shows a large difference.

What should the next investigation do?

Add intermediate values across the 0–80 range so the relationship can be mapped. A possible broad design could use 0, 20, 40, 60 and 80. After collecting those results, refine any region where a threshold, plateau or turning behaviour appears.

What should you not conclude from the two endpoints alone? You cannot assume the relationship is linear, steadily increasing, or free of turning points between them.

Delayed Independent Return

Three to five days later, take a fresh investigation and answer without notes:

What is the scientific question?
What factor is deliberately changed?
What outcome is measured?
What is the tested range?
How are the conditions spaced?
Is the range broad enough for the learner job?
Are there enough intermediate conditions?
Could a threshold, plateau or turning region be missed?
Can the measurement method distinguish the resulting changes?
Are repeats or more specimens also needed?
Would a broad scan or a focused refinement be more useful now?
Does the improved design still answer the same scientific question?

The Answer-Checking Receipt

Did I identify the changed factor correctly?
Did I separate test-condition spacing from measurement timing?
Did I distinguish range width from number of conditions?
Did I choose a scientifically appropriate range rather than the widest imaginable range?
Did I include enough intermediate conditions for the question?
Did I avoid assuming equal spacing is compulsory?
Did I check whether the measurement method can resolve the expected differences?
Did I preserve fair-test conditions?
Did I distinguish more conditions from more repeats?
Did I keep the original scientific question unchanged?
Did I keep conclusions inside the tested evidence?

Evidence and Model Limits

At higher levels, experimental design can use formal sampling theory, optimisation, statistical power and mathematical models. Primary Science does not need that machinery here.

The durable learner principle is simpler: the values you choose to test determine which parts of a scientific relationship you can actually observe.

No set of conditions guarantees a perfect result. Unexpected variation, hidden limitations and model failure can still occur. Good Science treats the first investigation as evidence, not as a promise that the pattern must look the way the learner expected.

Useful Internal Routes

Previous and Next Learning Routes

Previous: How to Decide Whether a PSLE Science Investigation Needs More Repeats or More Test Conditions.

Next: How to Choose Measurement Intervals in a PSLE Science Investigation Without Missing the Pattern.

Parent and Tutor Teaching Guide

Start with a simple challenge: “You can test five values. Which five will you choose?” Do not reveal the answer immediately.

Give three candidate designs:

20, 21, 22, 23, 24;
0, 25, 50, 75, 100;
0, 10, 50, 90, 100.

Ask what each design can reveal and what it might miss. The aim is to make the learner see test conditions as a way of sampling a relationship, not as numbers chosen to make a neat table.

Next, show a broad data set with a possible turning point. Ask the child to design a follow-up. If they simply add more values everywhere, ask whether concentrating the new values around the turning region would produce more useful evidence.

Then introduce a measuring instrument with coarse resolution. Ask whether extremely fine condition spacing still helps. This connects design to what the apparatus can actually detect.

Finally, use a categorical changed factor such as material type. Ask why numerical spacing is no longer the right idea. The learner is ready when they choose test conditions based on the scientific question rather than on one memorised pattern such as “always use five equally spaced values”.

Authoritative and Research References

The research references support broader scientific inquiry and measurement-learning principles. They do not prescribe one fixed PSLE number of test conditions, one compulsory spacing pattern or one official marking phrase.

The Quiet Ending

An investigation does not see a whole relationship at once.

It sees the places you choose to look.

Choose those places badly and the pattern may hide.

Choose them well, and a handful of carefully selected conditions can turn a vague question into visible scientific evidence.

Small Group Tutorials