Top 100 Secondary 1 Vocabulary List | Science, Observation and Investigation for Middle School Students

Lane: EDKSG-SEC1-VOCAB-WORLD-110

This Secondary 1 science vocabulary list is a world-facing Grade 7 science vocabulary guide for students learning the language of the scientific method, scientific inquiry, observation, inference, measurement, hypothesis, variables, controlled experiments, data, evidence, models, analysis and conclusion. It is designed for middle school students who need to read Science questions accurately, plan investigations, describe what they observed, separate observation from inference, control variables, record quantitative and qualitative data, interpret graphs and explain what the evidence actually supports.

Students searching for 7th grade science vocabulary words, middle school scientific method vocabulary, nature of science vocabulary, scientific investigation terms, observation and inference, independent and dependent variables, control group, empirical evidence, accuracy and precision, qualitative and quantitative data, hypothesis, experiment, measurement, trial, outlier, scientific model, theory and law often meet short definition lists. This article builds a working vocabulary system instead: each word is connected to scientific reasoning, experimental design, evidence quality, explanation and transfer across Biology, Chemistry, Physics, Earth Science and everyday investigation.

The wider eduKateSG route begins with the Vocabulary Learning Hub and connects to What Is Science?, What is Science | The Replaceable Logic, How Investigable Questions Work in School Science, Why Students Need to Know What Counts as Evidence, How Model-Based Reasoning Works and Claim–Evidence–Reasoning. Those pages own the deeper mechanisms. This page owns the Secondary 1 vocabulary job: the language students need to observe, investigate, test, measure, analyse, explain and revise scientifically.

How Maren, Iona and Leonie Use Science Vocabulary

Maren uses the vocabulary to build explanations: what changed, what mechanism could explain the change, and how the conclusion should be written so it does not claim more than the evidence supports. Iona uses the vocabulary to interrogate evidence: what was directly observed, what was inferred, what was measured, what uncertainty remains, and what alternative explanations still fit the data. Leonie uses the vocabulary to run investigations reliably: define the question, choose instruments, control variables, repeat trials, record results, check safety and make the procedure reproducible. Together they turn Science from a list of facts into a disciplined way of finding out.

Part I — Observation and Measurement: Words 1–20

1. Science

Meaning: systematic investigation of the natural world using observation, measurement, evidence, models, testing and explanation. Collocations: scientific inquiry, science investigation, scientific evidence, science education. Precision: science is not merely a collection of facts. Its claims remain connected to methods and evidence that can be inspected, challenged and improved. Example: “Iona treated the textbook statement as scientific knowledge because it rests on a wider evidence base, not because the sentence appeared in print.” Science move: ask what observation, measurement or test would make the claim stronger or weaker.

2. Observation

Meaning: information noticed or recorded directly using the senses or scientific instruments. Collocations: make an observation, careful observation, observational data, direct observation. Precision: “the liquid turned blue” can be an observation; “a chemical reaction occurred because a new substance formed” adds interpretation and inference. Instruments extend observation beyond unaided senses. Example: “Maren recorded that bubbles appeared within ten seconds without yet claiming what gas had formed.” Science move: write what was detected before writing what you think it means.

3. Inference

Meaning: an interpretation or conclusion drawn from observations, evidence and prior knowledge. Collocations: make an inference, infer from evidence, reasonable inference, scientific inference. Precision: inference goes beyond direct observation. It can be strong or weak depending on how well the evidence supports it. Example: “Leonie observed condensation outside the cold cup and inferred that water vapour from the air had condensed on the surface.” Science move: state which observations support the inference and identify what additional evidence could test it.

4. Question

Meaning: a clearly stated problem or uncertainty that an investigation aims to answer. Collocations: research question, investigable question, ask a question, scientific question. Precision: not every interesting question is directly investigable with available evidence. “Which fertiliser produces the greatest increase in plant height over four weeks?” is more testable than “Which fertiliser is best?” Example: “Iona rewrote the vague question so the variable, measurement and time period were visible.” Science move: ask whether the question tells you what evidence would count as an answer.

5. Investigation

Meaning: a planned process for answering a scientific question by collecting and analysing evidence. Collocations: scientific investigation, conduct an investigation, investigation design, investigation results. Precision: an investigation can be experimental, observational, comparative, field-based or model-based; not every scientific investigation is a controlled laboratory experiment. Example: “Maren investigated pond biodiversity by sampling and recording organisms rather than manipulating the pond.” Science move: choose the investigation type that can actually produce the evidence your question needs.

6. Inquiry

Meaning: disciplined questioning and evidence-seeking used to develop or test understanding. Collocations: scientific inquiry, inquiry process, inquiry-based science, guided inquiry. Precision: inquiry is broader than one experiment. It includes asking questions, searching existing knowledge, designing tests, interpreting evidence and revising explanations. Example: “Leonie continued the inquiry after the first experiment because the unexpected result created a new question.” Science move: treat an answer as a possible starting point for the next evidence-seeking question.

7. Phenomenon

Meaning: an observable event, process or pattern in the natural world that can be described or explained scientifically. Collocations: natural phenomenon, observed phenomenon, explain a phenomenon, physical phenomenon. Precision: a phenomenon is what happens; an explanation proposes why or how it happens. Example: “The shadow changing length through the day was the phenomenon; Earth’s rotation and the changing angle of sunlight were part of the explanation.” Science move: describe the phenomenon accurately before choosing a mechanism to explain it.

8. Evidence

Meaning: observations, measurements, data or results that bear on a scientific claim or explanation. Collocations: empirical evidence, supporting evidence, evidence from experiments, evaluate evidence. Precision: data become evidence relative to a claim. The same measurement may support one claim, weaken another and be irrelevant to a third. Example: “Iona asked which result actually supported the claim that temperature affected dissolving rate.” Science move: connect every conclusion to the specific evidence that justifies it.

9. Empirical

Meaning: based on observation, measurement or experiment rather than on speculation alone. Collocations: empirical evidence, empirical data, empirical test, empirical result. Precision: an empirical claim is tied to what can be observed or measured, but empirical evidence still needs good methods and interpretation. Example: “Maren preferred empirical measurements of cooling time to a guess about which material felt colder.” Science move: ask what observable evidence would distinguish competing explanations.

10. Measurement

Meaning: determination of a quantity by comparing it with an agreed unit or scale. Collocations: take a measurement, measurement error, accurate measurement, repeated measurement. Precision: measurement is more than “looking carefully”; it assigns a value using an instrument, unit and procedure. Example: “Leonie measured the liquid volume with a graduated cylinder rather than estimating by eye from a beaker.” Science move: choose an instrument and scale suitable for the size and precision of the quantity.

11. Unit

Meaning: an agreed standard quantity used to express a measurement. Collocations: unit of measurement, SI unit, standard unit, convert units. Precision: a number without an appropriate unit can be scientifically incomplete or ambiguous. “5” does not tell us whether the value is five metres, five grams or five seconds. Example: “Iona wrote 12.4 cm rather than 12.4 because the unit is part of the measurement.” Science move: record the unit whenever a measured quantity requires one.

12. SI

Meaning: the International System of Units, the globally standardised measurement system used in science and many technical fields. Collocations: SI unit, SI system, base unit, derived unit. Precision: using standard units allows measurements from different people and places to be compared meaningfully. Example: “Maren converted the measurement into metres so the results followed the expected SI convention.” Science move: know which standard unit is appropriate for the quantity and convert consistently when comparing datasets.

13. Accuracy

Meaning: closeness of a measurement or result to the true, accepted or reference value. Collocations: measurement accuracy, improve accuracy, accurate result, accuracy of an instrument. Precision: accuracy is different from precision. Several measurements can be tightly clustered yet all be far from the true value. Example: “Leonie’s balance repeatedly read 2 g too high, so the readings were precise but not accurate.” Science move: compare with a trusted reference or calibrated standard when accuracy matters.

14. Precision

Meaning: closeness of repeated measurements to one another and, in another sense, the level of detail allowed by the measurement scale. Collocations: measurement precision, precise readings, high precision, precision instrument. Precision: precise does not automatically mean accurate. Example: “Iona obtained 20.1, 20.1 and 20.2 cm, showing strong repeatability even before checking the ruler’s calibration.” Science move: repeat measurements and report only the level of detail the instrument genuinely supports.

15. Estimate

Meaning: an approximate value based on available information when exact measurement is unnecessary, unavailable or impractical. Collocations: estimate a value, rough estimate, reasonable estimate, estimated quantity. Precision: estimating is not careless guessing. A scientific estimate should use relevant scale, prior knowledge or partial measurement. Example: “Maren estimated the tree height using shadows before comparing the result with a more formal measurement.” Science move: state that a value is estimated and explain the basis of the approximation.

16. Instrument

Meaning: a device used to observe, detect or measure a quantity or condition. Collocations: measuring instrument, laboratory instrument, instrument reading, calibrate an instrument. Precision: an instrument extends observation but also introduces limits such as range, resolution and calibration error. Example: “Leonie chose a thermometer whose range covered the expected temperatures.” Science move: know what the instrument measures, its scale, its limits and the correct way to read it.

17. Qualitative

Meaning: describing qualities or characteristics without expressing the main observation as a numerical quantity. Collocations: qualitative observation, qualitative data, qualitative description, qualitative change. Precision: colour, texture, shape and smell can be qualitative observations. They are not automatically less scientific than numbers; usefulness depends on the question and method. Example: “Iona recorded the precipitate as pale yellow and powdery.” Science move: use precise descriptive language and agreed categories where possible.

18. Quantitative

Meaning: expressed using numbers obtained through counting or measurement. Collocations: quantitative data, quantitative measurement, quantitative result, quantitative comparison. Precision: numerical data are not automatically high quality. A poorly measured number can be less useful than a careful qualitative observation. Example: “Maren recorded that the mass decreased by 1.8 g rather than writing only that it became lighter.” Science move: attach units, methods and appropriate precision to quantitative values.

19. Record

Meaning: to preserve observations, measurements, procedures or results in a form that can later be inspected. Collocations: record data, record observations, written record, experimental record. Precision: memory is not a scientific record. Good records preserve enough detail to reconstruct what happened. Example: “Leonie recorded the temperature immediately after each reading instead of filling the table from memory later.” Science move: record data at the time of observation and do not silently alter unexpected values.

20. Table

Meaning: an organised arrangement of values or observations in rows and columns so patterns and comparisons are easier to inspect. Collocations: data table, results table, table heading, tabulate data. Precision: a table should make variables, units and trials clear. Decorative formatting cannot rescue unclear labels. Example: “Iona placed the independent variable in the first column and labelled the dependent measurement with its unit.” Science move: design the table before collecting data so every required result has a clear place to go.

Checkpoint 1 — Separate What You Observed From What You Think It Means

Choose one simple phenomenon, such as ice melting or a plant bending toward light. Write three direct observations, then three inferences. Turn the curiosity into an investigable question and decide what type of investigation or inquiry could collect useful empirical evidence. Identify the measurements, units and instruments you would use. Distinguish qualitative from quantitative information, decide whether any values are merely estimates, then record the planned observations in a clearly labelled table. Finish by explaining how accuracy and precision could fail in different ways.

Part II — Experimental Design: Words 21–40

21. Hypothesis

Meaning: a testable proposed explanation or relationship that can be investigated with evidence. Collocations: test a hypothesis, formulate a hypothesis, support a hypothesis, hypothesis testing. Precision: a hypothesis is not merely an “educated guess.” A useful scientific hypothesis states something that evidence could support, weaken or contradict. Example: “Maren hypothesised that increasing water temperature would increase the rate at which sugar dissolved.” Science move: write the hypothesis so the variables and expected relationship are clear enough to test.

22. Prediction

Meaning: a statement about an expected result if an explanation or pattern is correct. Collocations: make a prediction, predicted result, test a prediction, prediction based on evidence. Precision: a hypothesis proposes an explanation or relationship; a prediction states what should be observed if that hypothesis is correct. Example: “Iona predicted that the sugar would dissolve fastest in the warmest water.” Science move: make predictions specific enough that the actual result can be compared with them.

23. Variable

Meaning: a factor or quantity that can take different values in an investigation. Collocations: experimental variable, change a variable, measure a variable, control variables. Precision: variables include what is deliberately changed, what is measured and what must be kept sufficiently constant. Example: “Leonie listed temperature, stirring rate and amount of sugar as variables that could affect dissolving.” Science move: identify every factor that could change the outcome before deciding which one the investigation will manipulate.

24. Independent Variable

Meaning: the variable deliberately changed or selected by the investigator to test its relationship with another variable. Collocations: independent variable, manipulated variable, vary the independent variable, levels of the independent variable. Precision: “independent” here does not mean unrelated; it means the investigator treats this variable as the input or explanatory factor being changed. Example: “Maren used water temperature as the independent variable.” Science move: change the independent variable systematically rather than letting it drift accidentally.

25. Dependent Variable

Meaning: the measured or observed outcome expected to change in response to the independent variable. Collocations: dependent variable, responding variable, measure the dependent variable, dependent outcome. Precision: the dependent variable is the result being measured, not simply “the thing that changes” because many uncontrolled factors can also change. Example: “Leonie measured dissolving time as the dependent variable.” Science move: define exactly how the dependent variable will be measured so different trials remain comparable.

26. Controlled Variable

Meaning: a factor deliberately kept sufficiently constant so it does not provide an alternative explanation for the result. Collocations: controlled variable, keep constant, control conditions, control a variable. Precision: controlled variables are not the same as a control group. One refers to conditions kept stable; the other refers to a comparison group. Example: “Iona kept sugar mass and stirring rate constant while changing only water temperature.” Science move: control factors that could otherwise influence the dependent variable.

27. Constant

Meaning: a condition or value kept unchanged during a defined part of an investigation. Collocations: keep constant, experimental constant, constant condition, constant value. Precision: in school experiments, constants are often controlled variables. In broader science and mathematics, “constant” can also mean a fixed quantity. Example: “Maren kept the volume of water constant at 100 mL for every trial.” Science move: state the actual fixed value whenever possible instead of merely writing “keep the same.”

28. Control Group

Meaning: a comparison group that does not receive the experimental treatment or receives the standard condition. Collocations: control group, compare with control, untreated control, experimental control. Precision: not every investigation requires a control group. It is especially useful when comparing treatment against a baseline. Example: “Leonie compared fertilised plants with a control group receiving no fertiliser.” Science move: use the control group to estimate what would happen without the treatment.

29. Experimental Group

Meaning: the group that receives the treatment, condition or manipulation being tested. Collocations: experimental group, treatment group, compare groups, experimental condition. Precision: the experimental group is meaningful only relative to a comparison condition and a clearly defined treatment. Example: “Maren’s experimental group received a measured amount of fertiliser while the control group did not.” Science move: change only the intended treatment while keeping other relevant conditions as similar as possible.

30. Fair Test

Meaning: an investigation designed so that differences in outcome can reasonably be linked to the intended variable rather than obvious uncontrolled differences. Collocations: fair test, conduct a fair test, fair comparison, test fairly. Precision: “fair” is methodological, not moral. It refers to comparability and control of relevant factors. Example: “Iona used the same soil mass, pot size and light exposure so the fertiliser comparison was fairer.” Science move: ask which uncontrolled difference could create the result even if the hypothesis were wrong.

31. Procedure

Meaning: an ordered set of steps describing how an investigation is carried out. Collocations: experimental procedure, follow a procedure, step-by-step procedure, procedure error. Precision: a procedure should be detailed enough for another person to repeat the investigation meaningfully. Example: “Leonie specified the amount of water, stirring time and temperature rather than writing ‘mix the solution.’” Science move: write operational details, measurements and sequence clearly enough that the method is reproducible.

32. Method

Meaning: the overall approach or technique used to collect evidence or answer a scientific question. Collocations: experimental method, sampling method, measurement method, scientific method. Precision: method is broader than procedure. Two investigations may use the same method but different step-by-step procedures. Example: “Maren used repeated timed trials as the method for comparing dissolving rates.” Science move: choose a method capable of producing evidence relevant to the question, not merely one that is easy to perform.

33. Trial

Meaning: one complete run or measurement sequence within an investigation. Collocations: repeated trials, trial result, first trial, conduct a trial. Precision: one trial can be affected by random variation or an unnoticed mistake. Multiple trials help show whether the pattern is stable. Example: “Iona timed five trials at each temperature rather than relying on a single measurement.” Science move: record each trial separately before calculating summaries such as a mean.

34. Repetition

Meaning: performing the same measurement or procedure multiple times within an investigation to assess variation and reduce the influence of random error. Collocations: repeated measurements, repetition of trials, repeat the experiment, repeated observation. Precision: repetition usually occurs within the same investigation or team; replication is independent repetition by another study or team. Example: “Leonie repeated each timing five times to see whether one unusual value was typical.” Science move: repeat enough times to reveal the spread of results rather than hiding it.

35. Replication

Meaning: independent repetition of an investigation or study to see whether similar findings occur again. Collocations: replicate a study, independent replication, replication result, reproducible finding. Precision: replication is stronger than merely rerunning the same apparatus immediately because independent work can reveal hidden dependencies in the original setup. Example: “A second class replicated the investigation using the same design and found a similar trend.” Science move: treat a result as more trustworthy when it survives independent attempts to reproduce it.

36. Sample

Meaning: a selected subset of a larger population, material or set of possible observations used for investigation. Collocations: sample size, representative sample, collect a sample, random sample. Precision: conclusions about a population depend on how the sample was chosen and whether it represents relevant variation. Example: “Maren sampled leaves from several parts of the tree rather than taking all measurements from one branch.” Science move: choose a sample large and representative enough for the claim you want to make.

37. Random

Meaning: selected or assigned using a process in which individual outcomes are not deliberately chosen, helping reduce systematic selection effects. Collocations: random sample, random assignment, random variation, choose randomly. Precision: random does not mean careless. A random method is deliberately designed to avoid predictable selection. Example: “Iona used randomly generated positions to decide where to sample grass in the field.” Science move: use randomisation when deliberate choice could unintentionally favour certain cases.

38. Bias

Meaning: a systematic tendency in sampling, measurement, procedure or interpretation that shifts results in a particular direction. Collocations: sampling bias, measurement bias, observer bias, reduce bias. Precision: bias is different from random error. Random error creates scatter; bias can move many results consistently away from the truth. Example: “Leonie recognised sampling bias when all water-quality samples were taken beside the cleanest-looking bank.” Science move: ask whether the method could systematically favour one result.

39. Safety

Meaning: practices and conditions used to reduce unacceptable risk of injury, exposure or damage during scientific work. Collocations: laboratory safety, safety goggles, safety procedure, safety precaution. Precision: safety is part of experimental design, not an extra note added after the method is chosen. Example: “Maren replaced an open flame with a safer heating method because the investigation did not require combustion.” Science move: identify hazards before beginning and choose controls proportionate to the risk.

40. Risk

Meaning: the combination of the possibility of harm and the seriousness of its consequences. Collocations: risk assessment, reduce risk, safety risk, level of risk. Precision: hazard and risk are related but different. A hazard is something capable of causing harm; risk depends on exposure and likelihood as well as severity. Example: “Iona judged the risk from dilute solution differently from the risk from the same chemical in concentrated form.” Science move: consider hazard, exposure, likelihood and consequence when planning controls.

Checkpoint 2 — Design a Test That Can Actually Answer the Question

Choose a simple question and write a testable hypothesis plus a specific prediction. Identify all important variables, then label the independent variable, dependent variable and controlled variables. Decide whether a control group and experimental group are needed. Write the method and detailed procedure, plan multiple trials and distinguish repetition from later replication. If sampling is needed, state the sample and whether random selection would reduce bias. Finish with a safety check and a brief risk assessment. Then ask the hardest design question: if the predicted result appears, is there still another uncontrolled explanation that could have produced it?

Part III — Data, Patterns and Analysis: Words 41–60

41. Data

Meaning: recorded observations, measurements, counts or coded information collected during an investigation. Collocations: collect data, raw data, experimental data, data set. Precision: data are not automatically evidence until they are connected to a claim or question. Data quality depends on how they were collected. Example: “Maren kept the individual timing measurements as raw data before calculating any average.” Science move: preserve raw values so later summaries can be checked against the original record.

42. Result

Meaning: an outcome produced by an observation, measurement, trial or analysis. Collocations: experimental result, result of a trial, unexpected result, report results. Precision: a result is what happened; a conclusion is what the investigator thinks the result means in relation to the question. Example: “Iona reported that the warm-water trial averaged 42 seconds before discussing whether this supported the hypothesis.” Science move: separate the results section from the interpretation whenever possible.

43. Pattern

Meaning: a recurring or structured relationship visible across observations or data. Collocations: data pattern, recurring pattern, identify a pattern, pattern in results. Precision: a pattern may be real, random or caused by a hidden factor. Recognising it is the start of explanation, not the end. Example: “Maren noticed that dissolving time tended to fall as temperature increased.” Science move: describe the pattern before proposing why it occurred.

44. Trend

Meaning: an overall direction of change across ordered data, such as increasing, decreasing or remaining approximately stable. Collocations: upward trend, downward trend, overall trend, trend over time. Precision: a trend can exist even when individual values fluctuate. Example: “Leonie saw an overall downward trend in dissolving time despite one slower trial at 60°C.” Science move: describe the broad direction without pretending every point fits perfectly.

45. Relationship

Meaning: an association or connection between two or more variables or quantities. Collocations: relationship between variables, direct relationship, inverse relationship, investigate a relationship. Precision: a relationship may be causal, correlational or produced by another factor. Example: “Iona described a relationship between temperature and dissolving time without yet claiming temperature was the only cause.” Science move: state what changes with what, then ask what evidence is needed to identify the mechanism.

46. Correlation

Meaning: a statistical or observable association in which two variables tend to vary together. Collocations: positive correlation, negative correlation, strong correlation, correlation between variables. Precision: correlation does not by itself prove that one variable causes the other. A third variable, reverse causation or coincidence may explain the association. Example: “Maren found a correlation between sunny days and ice-cream sales but did not claim ice cream causes sunshine.” Science move: ask what alternative causal structures could generate the same pattern.

47. Cause

Meaning: a factor that helps produce a change or outcome. Collocations: cause an effect, causal factor, cause of change, establish causation. Precision: scientific causation usually requires stronger evidence than simple association. Causes can be multiple, conditional or indirect. Example: “Leonie treated temperature as a plausible cause only after the investigation controlled other major variables.” Science move: look for controlled comparisons, mechanisms and evidence that alternative causes were reduced.

48. Effect

Meaning: a change or outcome produced in response to one or more causal factors. Collocations: effect of temperature, measurable effect, observe an effect, causal effect. Precision: an effect can be large, small, delayed, variable or absent under different conditions. Example: “Iona measured the effect of fertiliser on plant growth rather than merely asking whether fertiliser was present.” Science move: define how the effect will be measured and over what time period.

49. Graph

Meaning: a visual representation of data used to show patterns, comparisons or relationships. Collocations: line graph, bar graph, plot a graph, graph the results. Precision: the graph type should fit the data and question. A misleading scale or inappropriate graph can distort interpretation. Example: “Maren used a line graph because the independent variable changed across ordered numerical values.” Science move: choose a graph that makes the relevant relationship visible without exaggerating it.

50. Axis

Meaning: a reference line on a graph along which a variable and its scale are displayed. Collocations: horizontal axis, vertical axis, x-axis, y-axis. Precision: axes need variable names and units where appropriate. Example: “Leonie placed water temperature on the horizontal axis and dissolving time on the vertical axis.” Science move: label each axis so the reader knows exactly what quantity is represented and in which unit.

51. Scale

Meaning: the set of numerical intervals used to represent values on an instrument or graph. Collocations: graph scale, measurement scale, scale interval, choose a scale. Precision: scale determines how much detail and apparent variation the display shows. Example: “Iona used equal intervals of 10°C on the axis rather than uneven spacing that could mislead the reader.” Science move: choose a scale that uses the available space well while preserving proportional meaning.

52. Mean

Meaning: the arithmetic average calculated by adding values and dividing by the number of values. Collocations: calculate the mean, mean value, mean result, average measurement. Precision: the mean is useful for repeated quantitative results but can be strongly affected by extreme values. Example: “Maren calculated the mean of five trial times after checking the individual results for anomalies.” Science move: do not report the mean alone when the spread of repeated results matters.

53. Median

Meaning: the middle value when a set of numerical values is arranged in order. Collocations: median value, calculate the median, median result, middle value. Precision: the median is less affected by extreme values than the mean and can be useful for skewed data. Example: “Iona compared mean and median when one unusually large value pulled the average upward.” Science move: choose the summary measure that fits the data distribution rather than using one automatically.

54. Range

Meaning: the difference between the largest and smallest values in a dataset or repeated set of measurements. Collocations: data range, measurement range, wide range, calculate the range. Precision: range is a simple measure of spread and can be strongly influenced by one extreme value. Example: “Leonie reported the mean timing and the range so readers could see how variable the trials were.” Science move: use the range as a quick check of consistency, not as a complete description of variability.

55. Outlier

Meaning: a value that lies unusually far from the main pattern of the data. Collocations: identify an outlier, possible outlier, extreme value, investigate an outlier. Precision: an outlier is not automatically a mistake and should not be deleted simply because it is inconvenient. Example: “Maren checked the apparatus and notes before deciding how to interpret the unusually long trial.” Science move: investigate the cause, report the value transparently and justify any exclusion.

56. Variation

Meaning: differences among repeated measurements, individuals, samples or conditions. Collocations: natural variation, measurement variation, variation between trials, biological variation. Precision: variation is often expected rather than a sign that science has failed. Living systems in particular may vary even under similar conditions. Example: “Iona expected some variation in plant height because individual plants were not identical.” Science move: measure and describe variation instead of pretending every repeated value should be the same.

57. Error

Meaning: the difference between a measured or observed value and the true or intended value, or a flaw introduced by measurement or procedure. Collocations: measurement error, random error, systematic error, experimental error. Precision: scientific error does not always mean a careless mistake. Random and systematic errors can remain even in careful work. Example: “Leonie distinguished reaction-time variation from a wrongly zeroed balance.” Science move: identify whether the error source is random, systematic or procedural because the remedy differs.

58. Uncertainty

Meaning: the limited confidence or range associated with a measurement, estimate, model or conclusion because evidence and instruments are imperfect. Collocations: measurement uncertainty, uncertainty range, scientific uncertainty, reduce uncertainty. Precision: uncertainty is not the same as ignorance. Science often quantifies or bounds what is not known exactly. Example: “Maren reported the thermometer reading with appropriate uncertainty instead of pretending the value was exact.” Science move: communicate uncertainty honestly and avoid conclusions more precise than the data allow.

59. Analyze

Meaning: to examine data or a problem systematically by separating parts, comparing values and looking for relationships or patterns. Collocations: analyze data, analyze results, data analysis, analytical method. Precision: analysis is more than calculation. It includes deciding what comparisons matter and whether the apparent pattern is meaningful. Example: “Iona calculated averages, checked outliers and compared the trend across temperatures.” Science move: move from raw values to structured comparisons without discarding inconvenient evidence.

60. Interpret

Meaning: to explain what data or results mean in relation to the scientific question, hypothesis and wider knowledge. Collocations: interpret results, interpret a graph, data interpretation, interpret evidence. Precision: interpretation adds meaning and therefore requires judgment. Different interpretations may fit the same limited data, so alternatives should sometimes remain open. Example: “Leonie interpreted the downward trend as consistent with faster dissolving at higher temperatures while noting that the experiment did not explain the molecular mechanism directly.” Science move: state what the data support, what they do not establish and what further evidence would discriminate alternatives.

Checkpoint 3 — Turn Raw Results Into an Evidence-Based Interpretation

Take a small set of data from repeated trials. Report the individual results, then identify any pattern, trend or relationship. Decide whether you have merely a correlation or stronger evidence for cause and effect. Plot an appropriate graph with correctly labelled axes and a sensible scale. Calculate the mean, consider whether the median would also be useful, and report the range. Investigate every outlier, describe variation, identify possible error and state relevant uncertainty. Finally, analyze the evidence and interpret it in a sentence that is strong enough to be useful but cautious enough to remain true.

Part IV — Scientific Explanation and Models: Words 61–80

61. Conclusion

Meaning: a statement that answers the investigation question by interpreting the results in relation to the hypothesis and evidence. Collocations: draw a conclusion, scientific conclusion, support a conclusion, conclusion from data. Precision: a conclusion should not merely repeat the result. It should state what the result means and how strongly the evidence supports the claim. Example: “Maren concluded that higher water temperature was associated with shorter dissolving time under the tested conditions.” Science move: make the conclusion no broader than the method and evidence justify.

62. Claim

Meaning: a statement asserting what is believed to be true about a phenomenon, relationship or explanation. Collocations: scientific claim, make a claim, support a claim, evidence for a claim. Precision: a claim can be correct or incorrect; what makes it scientific is that it can be connected to evidence and reasoning. Example: “Iona wrote the claim that temperature affects dissolving rate, then asked which data actually supported it.” Science move: separate the claim from the evidence used to justify it.

63. Reasoning

Meaning: the logical connection explaining why particular evidence supports or weakens a claim. Collocations: scientific reasoning, evidence-based reasoning, causal reasoning, reasoning chain. Precision: evidence does not speak for itself. Reasoning connects the observed result to scientific principles and the claim. Example: “Leonie explained that faster particle motion at higher temperature could increase collision frequency, linking the measured trend to the claim.” Science move: write the bridge between data and conclusion instead of placing them side by side without explanation.

64. Explanation

Meaning: an account of how or why a phenomenon occurs, supported by evidence and scientific ideas. Collocations: scientific explanation, explain a phenomenon, explanatory model, evidence-based explanation. Precision: description tells what happened; explanation proposes the mechanism or relationships producing it. Example: “Maren moved beyond ‘the ice melted’ to explain energy transfer from the warmer surroundings.” Science move: include both the observed pattern and the mechanism that could produce it.

65. Model

Meaning: a simplified physical, conceptual, mathematical or computational representation used to describe, explain or predict a system or phenomenon. Collocations: scientific model, model system, model prediction, revise a model. Precision: a model is deliberately incomplete. Its value depends on whether the simplifications preserve the features needed for the task. Example: “Iona used a particle model to explain diffusion while remembering that the diagram was not a literal picture of molecules.” Science move: ask what the model represents well, what it leaves out and what prediction it makes.

66. Theory

Meaning: a broad, well-supported scientific explanation that integrates many observations, findings and tested relationships. Collocations: scientific theory, theory of evolution, theory explains, evidence for a theory. Precision: in science, theory does not mean a casual guess. A scientific theory is a high-level explanatory framework supported by substantial evidence. Example: “Leonie explained that cell theory is a foundational scientific framework rather than an untested idea.” Science move: distinguish everyday use of “theory” from its scientific meaning.

67. Law

Meaning: a concise statement or mathematical relationship describing a consistent pattern in nature under specified conditions. Collocations: scientific law, law of motion, physical law, law describes. Precision: laws and theories do different jobs. A law describes a regular relationship; a theory explains broader mechanisms and patterns. One does not simply “become” the other. Example: “Maren treated a law as a description of regular behaviour rather than a superior stage of a theory.” Science move: ask whether the statement mainly describes what happens or explains why it happens.

68. Mechanism

Meaning: the process, interaction or sequence of causes through which a phenomenon is produced. Collocations: causal mechanism, biological mechanism, reaction mechanism, mechanism of action. Precision: a mechanism explains how one state leads to another rather than merely noting that the two are associated. Example: “Iona asked what mechanism links increased temperature to faster dissolving.” Science move: look for the intermediate steps connecting cause and effect.

69. System

Meaning: a set of interacting components considered together because their relationships produce behaviour or outcomes. Collocations: biological system, physical system, system boundary, interacting system. Precision: understanding one component does not automatically explain the whole system because interactions matter. Example: “Leonie described the digestive system as interacting organs and processes rather than a list of separate parts.” Science move: identify components, interactions, inputs, outputs and boundaries.

70. Structure

Meaning: the arrangement, organisation or form of parts within an object or system. Collocations: cell structure, molecular structure, structural feature, structure and function. Precision: structure concerns how parts are organised; function concerns what they do. Example: “Maren connected the thin walls of an alveolus to the short diffusion distance they create.” Science move: describe the relevant structural feature before explaining its functional consequence.

71. Function

Meaning: the role, operation or contribution of a component within a system. Collocations: biological function, function of a structure, functional role, structure–function relationship. Precision: function is not simply “what it looks like”; it describes what the part does in the system. Example: “Iona explained the function of red blood cells in transporting oxygen rather than describing only their shape.” Science move: connect function to evidence about how the component affects system performance.

72. Process

Meaning: a sequence of changes, interactions or operations through which a system moves from one state to another. Collocations: biological process, physical process, chemical process, process occurs. Precision: a process unfolds through time and often contains several stages. Example: “Leonie described photosynthesis as a process involving energy capture and chemical transformation rather than as one single event.” Science move: identify the sequence, inputs, transformations and outputs.

73. Matter

Meaning: physical substance that has mass and occupies space. Collocations: state of matter, matter cycle, conservation of matter, particles of matter. Precision: matter and energy are related in physical science but are not interchangeable vocabulary in ordinary school explanations. Example: “Maren tracked matter moving through a food chain rather than saying energy and matter ‘cycle’ in exactly the same way.” Science move: identify what material substance is present and how it moves or changes.

74. Energy

Meaning: a physical quantity associated with the capacity for change, motion, heating or other transformations. Collocations: energy transfer, energy store, energy transformation, conservation of energy. Precision: energy is not a material substance that gets “used up” in the same way as fuel. It is transferred and transformed while total energy is conserved in a closed system. Example: “Iona described energy transfer from the warmer water to the ice.” Science move: track where energy comes from, where it goes and how the form changes.

75. Force

Meaning: an interaction that can change an object’s motion or shape. Collocations: applied force, resultant force, gravitational force, force acts. Precision: force is not the same as motion. An object can move with no net force, and a force can act without producing visible motion if other forces balance it. Example: “Leonie distinguished the downward gravitational force from the object’s downward motion.” Science move: identify the interactions and consider all forces acting, not just the most obvious one.

76. Change

Meaning: a difference in state, value, structure, composition or behaviour over time or between conditions. Collocations: physical change, chemical change, change over time, rate of change. Precision: scientific explanation asks what changed, how much, under what conditions and through which mechanism. Example: “Maren measured the change in mass rather than simply writing that the sample became smaller.” Science move: define the before-and-after states and measure the relevant difference.

77. Cycle

Meaning: a recurring sequence in which materials, states or processes return through related stages. Collocations: water cycle, carbon cycle, life cycle, cyclic process. Precision: not every repeating diagram represents identical repetition; quantities and conditions can vary from cycle to cycle. Example: “Iona traced water through evaporation, condensation and precipitation while tracking where matter moved.” Science move: follow the pathway of the quantity or material through each stage instead of memorising arrows alone.

78. Equilibrium

Meaning: a condition in which opposing processes or effects balance so that the overall state remains approximately stable. Collocations: dynamic equilibrium, thermal equilibrium, equilibrium state, reach equilibrium. Precision: equilibrium does not always mean nothing is happening. In dynamic equilibrium, opposing processes can continue at equal rates. Example: “Leonie explained that evaporation and condensation can continue even when the amount of liquid remains stable.” Science move: ask which opposing processes balance and what would disturb that balance.

79. Classification

Meaning: the process of grouping objects, organisms or observations according to shared characteristics or rules. Collocations: scientific classification, classify organisms, classification system, classification criteria. Precision: classification depends on chosen features and can change when new evidence changes our understanding of relationships. Example: “Maren classified samples using observable properties, then compared that practical grouping with a formal scientific system.” Science move: state the criteria used and test whether ambiguous cases fit consistently.

80. Taxonomy

Meaning: an organised system for naming and classifying entities, especially organisms, according to defined categories and relationships. Collocations: biological taxonomy, taxonomic group, classification hierarchy, taxonomic evidence. Precision: taxonomy is a formal classification framework; ordinary classroom sorting is classification but not necessarily taxonomy. Example: “Iona used taxonomy to place organisms within nested biological groups.” Science move: understand the hierarchy and the evidence behind the grouping rather than memorising labels without relationships.

Checkpoint 4 — Move From Result to Scientific Explanation

Take one set of results and write a cautious conclusion. Turn that conclusion into a precise claim, then add the reasoning that connects the evidence to the claim. Build an explanation using an appropriate model and identify whether the wider knowledge involved is a theory, law or more local mechanism. Map the relevant system, structure, function and process. Track any matter, energy or force, define the important change, and identify whether a cycle or equilibrium is involved. If the investigation includes grouping, distinguish ordinary classification from a formal taxonomy. Then ask what part of your explanation is directly supported and what remains model-based or inferential.

Part V — Scientific Quality, Communication and Knowledge: Words 81–100

81. Validity

Meaning: the extent to which an investigation or measurement actually supports the intended conclusion or measures what it claims to measure. Collocations: experimental validity, valid conclusion, validity of a method, improve validity. Precision: a precise result can still be invalid if the method measures the wrong thing or uncontrolled factors explain the outcome. Example: “Iona questioned the validity of using leaf colour alone as a measure of plant growth.” Science move: ask whether the method really tests the stated question and whether alternative explanations remain.

82. Reliability

Meaning: the degree to which a method or measurement produces sufficiently consistent results under similar conditions. Collocations: reliable measurement, reliability of results, improve reliability, reliable method. Precision: reliability concerns consistency, not necessarily correctness. A badly calibrated instrument can give reliably wrong readings. Example: “Leonie improved reliability by repeating measurements and standardising the procedure.” Science move: examine whether repeated trials and independent attempts give a similar pattern.

83. Reproducibility

Meaning: the extent to which comparable results can be obtained again when the method, data or analysis is repeated or independently checked. Collocations: reproducible result, reproducibility crisis, reproduce findings, reproducible method. Precision: reproducibility is broader than simply repeating one measurement; it concerns whether the finding survives another attempt to obtain it. Example: “Maren judged the finding more credible after another group reproduced the trend.” Science move: record methods and data well enough that others can test the result.

84. Calibration

Meaning: checking and adjusting an instrument against a known reference so its readings remain appropriately accurate. Collocations: calibrate an instrument, calibration standard, calibration error, calibration curve. Precision: repeated readings can be precise while systematically wrong if calibration is poor. Example: “Iona calibrated the balance with a known mass before collecting experimental data.” Science move: verify instruments before trusting a suspiciously consistent dataset.

85. Standard

Meaning: an agreed reference, rule or benchmark used to make measurements, methods or judgments comparable. Collocations: standard unit, measurement standard, standard procedure, reference standard. Precision: standards create comparability; they do not guarantee that every use of the standard is correct. Example: “Leonie compared the thermometer against a reference standard rather than assuming the printed scale was accurate.” Science move: identify the reference against which a measurement or procedure is judged.

86. Peer Review

Meaning: evaluation of scientific work by other knowledgeable researchers before or after publication to assess methods, reasoning and evidence. Collocations: peer-reviewed article, peer review process, reviewer comment, scientific review. Precision: peer review is a quality-control process, not a guarantee that a study is correct forever. Example: “Maren understood that peer review can identify weak methods or reasoning that the original authors missed.” Science move: treat review as one layer in a larger error-correction system.

87. Research

Meaning: systematic work undertaken to develop, test or extend knowledge using evidence and appropriate methods. Collocations: scientific research, conduct research, research question, research findings. Precision: research can include experiments, field studies, observations, modelling, data analysis and synthesis of existing studies. Example: “Iona distinguished a classroom investigation from the broader body of research on the same phenomenon.” Science move: connect one study to what is already known rather than treating it as an isolated answer.

88. Literature

Meaning: the body of published scientific work relevant to a topic or research question. Collocations: scientific literature, research literature, review the literature, literature search. Precision: “literature” in science means published research and scholarly discussion, not fiction. Example: “Leonie checked the literature to see whether similar experiments had already been performed.” Science move: compare your result with a wider evidence base before assuming it is unique or definitive.

89. Source

Meaning: the origin from which scientific information, data or evidence is obtained. Collocations: primary source, reliable source, data source, source quality. Precision: sources differ in authority, methods, relevance and recency. A textbook, news article, research paper and raw dataset do different jobs. Example: “Maren traced the claim to the original study instead of relying only on a social-media summary.” Science move: ask who produced the information, how, with what method and for what purpose.

90. Citation

Meaning: a reference that identifies the source of a scientific claim, quotation, method or evidence. Collocations: cite a source, citation format, reference list, scientific citation. Precision: a citation is useful only if it points to a real and relevant source that actually supports the statement. Example: “Iona checked whether the cited paper contained the claimed result rather than trusting the citation itself.” Science move: verify the relationship between claim and source.

91. Communication

Meaning: clear sharing of scientific questions, methods, results, reasoning and uncertainty so others can understand and evaluate the work. Collocations: scientific communication, communicate results, science communication, communicate uncertainty. Precision: communication is part of science because evidence must be inspectable by others. Example: “Leonie explained the method and limitations so another group could understand why the conclusion was cautious.” Science move: make methods and uncertainty visible instead of presenting only a polished final answer.

92. Report

Meaning: a structured account of an investigation containing enough information about purpose, method, results and interpretation for others to evaluate it. Collocations: laboratory report, scientific report, report findings, report format. Precision: a report records what was done and discovered; it should not hide unexpected results or method weaknesses. Example: “Maren organised the report into question, method, results, discussion and conclusion.” Science move: write so a reader can trace the path from question to evidence to conclusion.

93. Discussion

Meaning: the part of a scientific report where results are interpreted, compared with expectations and evaluated for strengths, weaknesses and implications. Collocations: discussion section, discuss results, scientific discussion, discussion of limitations. Precision: discussion is not casual conversation; it is structured interpretation of what the results mean. Example: “Iona used the discussion to explain the outlier and consider whether uncontrolled temperature variation affected the result.” Science move: include alternative explanations and method limitations rather than only restating the graph.

94. Limitation

Meaning: a constraint or weakness that reduces the scope, accuracy or strength of an investigation or conclusion. Collocations: method limitation, study limitation, limitation of evidence, acknowledge limitations. Precision: a limitation does not automatically invalidate a study. It tells us where confidence or generalisation should be reduced. Example: “Leonie noted that the small temperature range limited how broadly the conclusion could be applied.” Science move: state how the limitation could affect the result, not merely that a limitation exists.

95. Assumption

Meaning: a condition or idea treated as true for the purpose of a model, calculation or investigation even though it may not be directly tested within that work. Collocations: simplifying assumption, underlying assumption, make an assumption, test an assumption. Precision: assumptions are not automatically mistakes. They become problematic when hidden, unrealistic or crucial to the conclusion. Example: “Maren assumed the room temperature remained approximately constant during the short experiment.” Science move: make important assumptions explicit and ask whether the result would change if they were false.

96. Alternative Explanation

Meaning: another plausible account of the same observation or result that differs from the preferred explanation. Collocations: alternative explanation, competing explanation, rule out an explanation, compare explanations. Precision: strong scientific reasoning looks for alternatives rather than protecting the first idea that fits. Example: “Iona considered whether increased stirring rather than temperature could explain faster dissolving.” Science move: ask what evidence would distinguish the preferred explanation from the strongest alternative.

97. Revision

Meaning: deliberate modification of a hypothesis, model, method or explanation when new evidence or criticism shows improvement is needed. Collocations: revise a model, revise a hypothesis, model revision, revise the method. Precision: revision is a strength of science because explanations are expected to improve when evidence changes. Example: “Leonie revised the model after the new observation contradicted one of its predictions.” Science move: identify exactly which part of the explanation failed and change that part rather than discarding everything blindly.

98. Falsifiable

Meaning: capable, in principle, of being shown wrong by some possible observation or test. Collocations: falsifiable hypothesis, testable claim, falsifiable prediction, scientific falsifiability. Precision: falsifiable does not mean false. It means the claim risks being contradicted by evidence rather than being protected from every possible result. Example: “Maren preferred a hypothesis that made a measurable prediction over a statement that could explain any outcome after the fact.” Science move: ask what result would make you change your mind.

99. Consensus

Meaning: broad agreement among relevant experts that develops when multiple lines of evidence and repeated scrutiny support a shared conclusion. Collocations: scientific consensus, expert consensus, consensus view, reach consensus. Precision: consensus is not decided by popularity alone and can change when strong new evidence arrives. Example: “Iona understood consensus as the outcome of accumulated evidence and criticism rather than as one authority declaring the answer.” Science move: examine why experts agree and what evidence supports the agreement.

100. Knowledge

Meaning: organised understanding built from evidence, reasoning, models, tested explanations and accumulated scientific work. Collocations: scientific knowledge, knowledge base, build knowledge, revise knowledge. Precision: scientific knowledge can be dependable without being absolutely final. Its strength comes from evidence, reproducibility, criticism and capacity for revision. Example: “Maren treated current scientific knowledge as the best-supported model available, open to improvement if better evidence appears.” Science move: hold confidence in proportion to evidence and keep the route to revision open.

Checkpoint 5 — Decide How Much Confidence the Evidence Deserves

Take one investigation and evaluate its validity, reliability and reproducibility. Check whether instruments need calibration against a known standard. Imagine what peer review might criticise and compare the finding with the wider research literature. Trace important sources and citations. Plan how the result should be communicated in a scientific report and what belongs in the discussion. State every important limitation and assumption, then generate the strongest alternative explanation. Ask whether the claim is genuinely falsifiable and what evidence would force revision. Finally, distinguish one study’s result from broader scientific consensus and from the larger body of scientific knowledge.

Part VI — Scientific Investigation Laboratories: Turn Vocabulary Into Method

The scientific vocabulary in this article matters because investigations fail at the boundaries between words. Students may know what “variable” means but change three things at once. They may calculate a mean but delete an inconvenient outlier. They may observe bubbles and immediately write a causal explanation as though it were a direct observation. The laboratories below force the vocabulary to work under real conditions.

Laboratory 1 — Does Stirring Make Sugar Dissolve Faster?

Maren begins with a familiar phenomenon: sugar appears to disappear faster when tea is stirred. A weak investigation simply compares one stirred cup with one unstirred cup and declares victory. A strong investigation first turns the observation into an investigable question: “How does stirring rate affect the time taken for 10 g of sugar to dissolve in 100 mL of water at 25°C?”

The independent variable is stirring rate. The dependent variable is dissolving time. Water volume, water temperature, sugar mass, sugar crystal size, cup shape and starting procedure become possible controlled variables. The point is not to control absolutely everything in the universe. It is to control the factors that could plausibly change the measured outcome enough to confuse the comparison.

Leonie writes the procedure. Each trial uses the same sugar mass and water volume. The water is checked at the same starting temperature. Stirring is performed at a defined rate, perhaps using a metronome or mechanical stirrer rather than “slow,” “medium” and “fast” by feel. Timing begins at a stated moment and ends according to an operational definition of “dissolved.”

This last detail is more important than it looks. If one student stops timing when no sugar is visible and another waits until every crystal has vanished from the bottom, the measurement method changes. The dependent variable may appear numerical, but the criterion defining the endpoint contains judgment. Good science makes that judgment visible.

Iona asks for repeated trials. One timing at each stirring rate can be distorted by reaction time, one unusually large crystal or a small temperature difference. Five trials reveal variation. The team records every value in a table, calculates the mean, reports the range, and marks any possible outlier without deleting it automatically.

Suppose the results are 95, 92, 94, 93 and 145 seconds at one stirring rate. The 145-second value deserves investigation. Did stirring stop? Was the water cooler? Was the timer started late? If no reason can be identified, the value remains part of the evidence. Repetition helps the group see whether it is representative.

Maren then plots a graph. Stirring rate belongs on the horizontal axis; mean dissolving time belongs on the vertical axis. The team looks for a trend. If dissolving time falls as stirring rate increases, there is a relationship. Because the investigation deliberately manipulates stirring rate while controlling major alternatives, the design provides stronger causal evidence than a simple observation of tea at home.

But the conclusion must stay within the evidence. “Stirring causes sugar to dissolve faster under all conditions” is too broad. “Under the tested conditions, higher stirring rates reduced the measured dissolving time” is better. The experiment has not tested every temperature, every solute or every stirring pattern.

Iona then asks for a mechanism. Stirring moves unsaturated solution away from the crystal surface and brings fresh solvent into contact with it. That mechanistic explanation connects the measured trend to particle-level reasoning. It is different from the result itself.

Your task: redesign the experiment for temperature instead of stirring. State the question, hypothesis, prediction, independent variable, dependent variable, controlled variables, procedure, number of trials and safety considerations. Then write one conclusion that would be justified if the predicted trend appears and one alternative explanation you would still check.

The laboratory teaches a core principle: a fair test is not created by writing “keep everything else the same.” It is created by identifying the causal alternatives that matter and building the method so evidence can discriminate among them.

Laboratory 2 — The Plant Grew Taller: Did the Light Colour Cause It?

Iona places two plants on different windowsills. One receives blue light from a lamp; the other receives ordinary white light. After two weeks the blue-light plant is taller. The tempting conclusion is immediate: blue light causes faster growth. The evidence, however, is much weaker than the sentence sounds.

The first issue is sample size. One plant per condition gives no information about ordinary biological variation. Perhaps the taller plant had a larger seed, healthier roots or a different genetic tendency. Increasing the sample allows the group to see whether the difference is consistent across individuals.

The second issue is controlled variables. Were both plants the same species and age? Did they receive equal water, soil, pot size, temperature and total light intensity? A blue lamp can also produce a different amount of heat or total light from a white lamp. If several conditions differ, colour is only one possible cause.

Maren rewrites the design. Twenty similar seedlings are randomly assigned to two groups. Both groups receive equal soil, pot size, water, temperature and exposure duration. Light intensity is measured rather than guessed. Plant height is measured from the same reference point every two days. This does not create perfection, but it greatly improves validity.

Leonie asks whether height is the best dependent variable. A taller plant is not automatically a healthier or more productive plant. Stem elongation can occur under stress. The scientific question determines what outcome should be measured. If the question is biomass production, dry mass might be more valid than height. If the question is leaf development, leaf area or number may matter.

This distinction between “easy to measure” and “valid measure” is crucial. Science students often choose the convenient number. Strong investigation design asks whether the number actually represents the construct named in the question.

Suppose the improved experiment still shows greater mean height under blue light. The group can now make a stronger claim, but it should still report variation and uncertainty. The mean difference might be large relative to the spread, or it might be small and inconsistent. The pattern matters, not just which average is larger.

Iona distinguishes correlation from causation. In the original windowsill observation, light colour and height differed together, but many other factors also differed. In the controlled experiment, deliberate manipulation and controlled conditions make causal interpretation more credible. Experimental design changes what kind of inference the data can support.

Maren then asks about replication. If another class repeats the study with a different batch of seedlings and obtains the same general result, confidence increases. If the second group gets no difference, the disagreement becomes scientifically useful. The method, sample, light intensity and biological conditions can be compared.

Your task: take the vague claim “music helps plants grow.” Design an investigation that could test one precise version of the claim. Define what kind of music, how exposure is measured, which outcome represents growth, what variables must be controlled, how many plants are needed and what result would count against the hypothesis.

The lesson is that causation is not a stronger word chosen at the end of a report. It is an inference earned—or not earned—by the design of the investigation.

Laboratory 3 — The Outlier Everyone Wants to Delete

Leonie times a pendulum swinging ten times. Most results are between 14.8 and 15.2 seconds. One result is 18.9 seconds. A classmate says, “Delete it. It ruins the average.” This is exactly the moment scientific vocabulary matters.

The 18.9-second value is a possible outlier. It is unusual relative to the other measurements, but unusual does not mean false. The group must investigate. Did the pendulum strike something? Did the timer start late? Was the swing count wrong? Did the string length change? Is the value a recording error, a procedural event or genuine variation?

Iona checks the laboratory notes. One line says, “Paused because chair moved.” That provides a plausible procedural reason for the long time. The group can report the value, explain the issue and justify excluding that trial from the main summary while preserving the raw record. Transparency matters.

Now change the scenario. Imagine there is no note and no obvious event. Deleting the point simply because it is inconvenient would create bias. The proper response may be to repeat additional trials and see whether similar long times recur. One unexpected value can reveal an error, but it can also reveal a phenomenon the original model ignored.

Maren examines how the value changes the mean and median. The mean may shift substantially; the median may barely move. This does not mean the median is always better. It means different summary measures respond differently to extremes. The choice should fit the scientific question and data structure.

Leonie then separates random error from systematic error. Human timing introduces random variation because reaction time differs from trial to trial. A stopwatch that consistently runs slow would produce systematic error. Repeating trials helps reduce uncertainty from random variation but does not automatically fix a systematic shift.

This explains why calibration matters. An instrument can produce tightly clustered values and still be wrong. Precision without accuracy can fool students because consistent numbers feel trustworthy. Calibration compares the instrument against a known reference.

Iona adds uncertainty. Every measured value has limits. The timer resolution, reaction time and method contribute uncertainty. Reporting 15.000000 seconds would create false precision if the method cannot support that detail. The number of decimal places should match the measurement process.

The group now writes the discussion. Instead of “there were no errors,” they explain likely error sources, how those errors might affect results, and which changes would improve reliability or validity. A perfect-sounding report is less scientific than an honest one that identifies weaknesses.

Your task: create a dataset of eight repeated measurements containing one outlier. Calculate the mean, median and range with and without the outlier. Write two possible explanations for the unusual value and state what further evidence would determine whether exclusion is justified.

The larger lesson is that inconvenient data are not enemies of science. They are invitations to diagnose the method, the instrument or the model.

Laboratory 4 — Observation Is Not Explanation

Maren takes a cold metal can from a refrigerator and places it on the table. A few minutes later, droplets appear on the outside. A student writes: “Water leaked through the can.” Another writes: “Condensation happened because the cold surface cooled water vapour in the air.” The two sentences do different scientific jobs.

The direct observation is that droplets appeared on the external surface. Perhaps the can’s mass can be measured before and after. Perhaps the droplets can be collected. “Water leaked through the can” is an inference or explanation, not an observation. The distinction matters because the explanation can be tested.

Iona proposes an alternative explanation: water vapour from the air condensed on the cold surface. What evidence would distinguish the two accounts? If the can contains coloured water and the external droplets are colourless, that would count against simple leakage. If an empty sealed cold can still develops external droplets, the liquid inside is not required.

This is scientific reasoning as comparison among explanations. The preferred explanation becomes stronger not merely because it sounds familiar but because it predicts observations that differ from the alternative.

Leonie designs a procedure. Use two sealed metal cans, one chilled and one at room temperature. Keep external conditions similar. Record surface temperature and appearance of droplets over time. The independent variable is can temperature; the dependent variable could be mass of external water collected after a fixed period.

Maren connects the result to a model of matter. Water molecules in the air move and interact with the colder surface. Cooling changes the balance between evaporation and condensation. The model explains why droplets can form without liquid crossing the metal.

Iona asks what the model predicts. If air humidity is lower, less external water should appear under otherwise similar conditions. If the surface is not sufficiently cold, condensation should decrease. Predictions make the explanation testable.

The team also checks vocabulary around certainty. One classroom investigation does not prove every detail of the molecular mechanism directly. It provides evidence consistent with the condensation explanation. Strong scientific writing can be confident without claiming more than was measured.

Your task: observe a familiar phenomenon—fogging glasses, melting ice, rusting metal or a shadow changing. Write three observations and three inferences. Choose one inference and generate an alternative explanation. Then design one measurement or comparison that would distinguish them.

The lesson is foundational: science advances when students learn to put a small distance between “I saw” and “therefore I know why.” That distance is where evidence and testing live.

Part VI — Scientific Investigation Laboratories: Calibration, Sampling, Models and Consensus

Laboratory 5 — Precise, Consistent and Wrong

Iona measures a 100 g reference mass five times with a digital balance. The readings are 102.0, 102.1, 102.0, 102.1 and 102.0 g. The results are beautifully consistent. They are also wrong. This is the cleanest possible demonstration that precision and accuracy are different.

The narrow spread shows high precision. The average is close to 102.0 g. But the accepted reference is 100.0 g, so the instrument appears systematically high. Repeating the measurement has not repaired the problem because the same systematic error affects every trial.

Maren asks what calibration means. The balance should be compared against known standards and adjusted if necessary. Calibration creates a relationship between the instrument reading and a reference value. It turns “the screen says 102.0” into a question about whether 102.0 corresponds to the real mass.

Leonie then checks zeroing. Perhaps the balance has a 2 g offset because it was not tared correctly. This is a procedural issue rather than a random fluctuation. Once corrected, repeated measurements can be taken again.

Iona introduces validity. Suppose the investigation asks whether a material loses mass during heating. A balance that is consistently offset by 2 g might still detect the change accurately if the offset remains constant before and after. The absolute measurement is inaccurate, but the comparison might still answer the question. This shows why validity depends on the scientific job.

Maren then changes the instrument. A spring balance with coarse markings gives readings of 100, 105, 100, 100 and 105 g. The average may be close to the true value, but the measurements are less precise. A result can be accurate on average while individual measurements remain noisy.

The class now has four concepts to separate: accuracy, precision, reliability and validity. Accuracy asks closeness to truth or reference. Precision asks closeness among repeated measurements. Reliability asks consistency under similar conditions. Validity asks whether the method supports the intended conclusion.

Your task: create four imaginary instruments: accurate and precise, accurate but imprecise, precise but inaccurate, and neither. Give each five measurements of the same 50.0-unit standard. Explain which problem repeated trials can reduce and which requires calibration or method change.

The lesson is that tidy numbers deserve inspection, not automatic trust. Consistency is evidence about one property of a measurement system, not proof of overall correctness.

Laboratory 6 — The Stream That Looked Healthy Because We Sampled the Wrong Place

Leonie wants to investigate biodiversity in a stream. She chooses a clear, shallow section near the path because it is easy to reach and counts many organisms. The result looks encouraging. Iona asks a harder question: does this sample represent the stream?

A sample is useful only relative to the population or area about which a claim will be made. If the stream contains shaded sections, deep pools, fast-flowing areas and locations near a drain, one convenient site may miss important variation. Convenience sampling can create bias.

Maren suggests a random or stratified approach. The stream can be divided into sections, with sampling points selected across different habitat types. Random selection prevents the investigator from consciously or unconsciously choosing only attractive or accessible locations.

Leonie now defines the method. At each location, use the same sampling area and observation time. Record water temperature, flow condition and organism counts. Photographs or field notes can preserve qualitative observations. If possible, repeat sampling on different days to reveal temporal variation.

Iona asks about classification. Identifying organisms requires a consistent classification system. If one team records “small insect” while another records “mayfly nymph,” the categories may not be comparable. A field key or taxonomy guide improves consistency.

The group then confronts measurement uncertainty. Some organisms may be missed, counted twice or misidentified. Turbid water reduces visibility. Fast flow changes what can be captured. These limitations should be recorded rather than hidden.

Maren plots organism richness against location characteristics. A pattern may appear, but causation is harder to establish because this is mainly an observational field investigation. Low biodiversity near a drain may be associated with water quality, flow, habitat structure or several interacting factors.

This is where students learn that controlled experiments are not the only route to science. Some systems cannot be manipulated ethically or practically. Careful sampling, repeated observation and multiple lines of evidence become the main tools.

Your task: design a field-sampling plan for insects in a school garden. Define the population, sample size, selection method, observation unit, classification method and one likely source of sampling bias. Explain what conclusion your design could support and what claim would be too broad.

The larger lesson is that evidence can be perfectly recorded and still misleading if the sample was chosen poorly. Scientific quality begins before the first measurement.

Laboratory 7 — When the Model Predicts the Wrong Thing

Maren uses a simple particle model to explain diffusion. Particles move randomly and spread from regions of higher concentration to lower concentration. The model works well for many classroom examples. Then the class studies a membrane that allows one substance through but blocks another. The original model is incomplete.

A weak response is to say the model was “wrong” and discard it entirely. A stronger scientific response asks which part of the model remains useful and which new structure or mechanism must be added. This is revision.

Iona lists the model’s assumptions: particles move, collisions occur, concentration differences matter, and the boundary does not initially distinguish particle types. The membrane result contradicts the last assumption. The model must now represent selective permeability.

Leonie adds a diagram showing membrane channels or pores. The revised model predicts that particles with certain properties can cross more easily than others. This creates new predictions that can be tested.

Maren explains why models are useful despite being incomplete. A model simplifies reality so important relationships become visible. A model containing every possible detail would often be unusable. The scientific question is whether the simplification preserves the mechanism relevant to the task.

Iona then asks about falsifiability. A model that can be adjusted after every result without ever risking contradiction is scientifically weak. Strong models make predictions that could fail. Failed predictions reveal where revision is needed.

The group connects this to theory. Scientific theories are not single classroom diagrams; they integrate broad evidence and explanatory structures. Individual models can represent parts of a theory for specific questions.

Leonie also distinguishes model failure from measurement failure. If the prediction fails, first check whether the experiment and instruments worked. Only then decide how strongly the evidence challenges the model. Science is a loop among model, prediction, observation and revision.

Your task: choose a simple school science model—particle model, food web, circuit diagram or ray model. Write three assumptions, two useful predictions and one situation where the model is limited. Propose a revision and state one new prediction created by the revised model.

The lesson is that models are not pictures to memorise. They are working representations that earn their value by helping us explain, predict and revise.

Laboratory 8 — From One Experiment to Scientific Consensus

Iona reads a headline: “New study proves food X improves memory.” The article describes one small experiment with thirty participants. The result is statistically interesting, but Maren asks whether one study is enough to create scientific knowledge.

One study contributes evidence. Its strength depends on validity, sample size, method, analysis, transparency and limitations. If the method is weak, the finding may not survive independent scrutiny. If the effect is real but small, another sample may produce a different-looking result by chance.

Leonie looks for peer review. Publication in a reviewed journal is useful because other researchers have inspected the work, but peer review is not a certification of eternal truth. Reviewers can miss problems, and later evidence can revise the interpretation.

Iona then checks the wider literature. Are there earlier studies? Do they use similar methods? Do results agree? A single surprising study may become less convincing when larger studies fail to replicate it, or more convincing when independent teams repeatedly reproduce the effect.

Maren distinguishes evidence accumulation from voting. Scientific consensus is not simply “most scientists believe X.” It usually reflects convergence of evidence across methods, datasets, laboratories and theoretical understanding. Consensus can be stronger when different methods point toward the same conclusion.

The group also examines source quality. A news story may simplify the paper. A social-media post may simplify the news story. By the time the claim reaches students, important uncertainty can disappear. Tracing back to the original source and citation restores context.

Leonie writes a confidence ladder. One unreplicated small study: interesting. Several consistent studies: stronger. Independent replications across contexts: stronger still. Converging mechanisms and large evidence syntheses: stronger again. At every stage, future evidence can still modify details.

Maren then adds the word limitation. Even a strong consensus can have boundaries. A conclusion may apply to adults but not children, laboratory conditions but not field conditions, one species but not another. Scientific knowledge becomes powerful when its scope is clear.

Your task: take a science headline and build an evidence chain: media claim → original source → study design → sample → key result → limitations → later or related research. State your confidence at each stage and explain what new evidence would move that confidence up or down.

The final lesson is that science is neither “one experiment proves it” nor “nothing can ever be known.” Reliable knowledge emerges from repeated testing, transparent methods, criticism, replication, synthesis and willingness to revise.

What the Eight Investigation Laboratories Reveal

The laboratories reveal a recurring pattern: the quality of a scientific conclusion is constrained by the quality of the question, method, measurement, sample, analysis and explanation that precede it. A polished conclusion cannot rescue a weak design.

They also show why scientific vocabulary is operational. “Validity” changes which outcome you measure. “Bias” changes how you sample. “Calibration” changes whether you trust the instrument. “Alternative explanation” changes whether you stop after the first plausible story. “Replication” changes how much confidence one study deserves.

Most importantly, science becomes a closed loop: observe → question → model → predict → investigate → measure → analyse → explain → criticise → revise → test again. Vocabulary names the parts of that loop so students can diagnose where reasoning failed.

Part VII — Precision Clinics: Science Terms That Must Not Collapse Into One Another

Strong science answers often depend on distinctions that look small on paper but completely change the reasoning. These clinics make the boundaries explicit.

Clinic 1 — Observation vs Inference

An observation records what was directly detected or measured. An inference interprets what that observation may mean. “The liquid became cloudy” is observation. “A precipitate formed because an insoluble product was produced” includes inference and explanation.

The distinction matters because observations can often be checked directly, while inferences must compete with alternatives. Good scientific writing shows the evidence before the interpretation.

Clinic 2 — Hypothesis vs Prediction

A hypothesis proposes a testable explanation or relationship. A prediction states what result should occur if that hypothesis is correct. “Increasing temperature increases dissolving rate” is a hypothesis. “Sugar will dissolve fastest at 60°C” is a prediction derived from it.

Students often write predictions and call them hypotheses. The repair is simple: ask what broader idea the predicted result is testing.

Clinic 3 — Independent vs Dependent vs Controlled Variable

The independent variable is deliberately changed. The dependent variable is measured as the outcome. Controlled variables are kept sufficiently stable so they do not provide obvious alternative explanations.

For temperature and dissolving time: temperature is independent, dissolving time is dependent, while sugar mass, water volume and stirring rate are controlled. Naming the variables should reveal the causal structure of the test.

Clinic 4 — Control Group vs Controlled Variable

A control group is a comparison group that does not receive the experimental treatment or receives the standard condition. A controlled variable is a factor kept similar across groups or trials.

A fertiliser experiment may have a no-fertiliser control group while still controlling water, soil, light and pot size. The words sound related because both help comparison, but they refer to different parts of the design.

Clinic 5 — Accuracy vs Precision

Accuracy means closeness to a true or accepted value. Precision means repeated values are close to one another or that a measurement is reported at fine resolution. Measurements can be precise but inaccurate when an instrument has a systematic offset.

The practical lesson is that repeating a biased instrument does not create accuracy. Repetition helps reveal random variation; calibration helps detect systematic shift.

Clinic 6 — Repetition vs Replication

Repetition usually means taking repeated measurements or repeating trials within the same investigation. Replication means another independent attempt to reproduce the result, often by another team or study.

Repetition helps estimate variation inside one setup. Replication tests whether the finding survives a different implementation and therefore provides a stronger challenge to hidden dependencies.

Clinic 7 — Correlation vs Causation

Correlation means variables vary together. Causation means changing one factor helps produce a change in another. Correlation can arise through a third factor, reverse direction or coincidence.

Controlled manipulation, temporal order, dose–response patterns, mechanisms and converging evidence can strengthen causal inference. A graph showing two rising lines is not enough by itself.

Clinic 8 — Result vs Conclusion

A result reports what happened: measurements, averages, patterns and observations. A conclusion interprets those results in relation to the question or hypothesis.

“Mean dissolving time fell from 80 s to 40 s” is result. “Higher temperature reduced dissolving time under these conditions” is conclusion. Keeping them separate makes it easier to check whether the interpretation is justified.

Clinic 9 — Theory vs Law

A scientific theory is a broad explanatory framework supported by substantial evidence. A scientific law describes a consistent relationship or pattern, often mathematically. A theory does not “graduate” into a law.

Theory and law perform different scientific jobs: explanation and description. Both can be powerful and well supported.

Clinic 10 — Validity vs Reliability vs Reproducibility

Validity asks whether the method supports the intended conclusion. Reliability asks whether results are sufficiently consistent under similar conditions. Reproducibility asks whether comparable findings can be obtained again when the work is repeated or independently checked.

A method can be reliable but invalid: it consistently measures the wrong outcome. A valid idea tested with a noisy method may lack reliability. A finding that repeatedly fails independent reproduction deserves reduced confidence.

The Precision Principle for Science

Scientific language is useful because it prevents reasoning from sliding across boundaries unnoticed. Observation is not inference. A pattern is not automatically a cause. A mean is not the whole dataset. A model is not reality. A peer-reviewed study is not a final consensus. When the words stay precise, the logic becomes easier to inspect and improve.

Part VIII — A 30-Day Secondary 1 Science, Observation and Investigation Curriculum

The purpose of this 30-day route is not to make students memorise one hundred definitions in isolation. It is to turn the words into a working scientific language. Each day combines retrieval with an observable task, a design decision, a data move or an explanation move. By Day 30, students should be able to use the vocabulary to plan an investigation, diagnose a weak method, interpret uncertain evidence and explain why a conclusion deserves exactly the amount of confidence they give it.

Days 1–5 — Observation, Measurement and the Discipline of Noticing

Day 1 — Science, observation, inference and question: choose one ordinary phenomenon: condensation on a cold cup, a shadow changing length, bread becoming mouldy, a metal spoon warming in soup, or a plant leaning toward light. Write five observations that another person could check. Then write five inferences. Circle every sentence where you accidentally mixed observation and inference. Turn one inference into an investigable question by naming what could be measured or compared.

Day 2 — Investigation, inquiry, phenomenon and evidence: take yesterday’s question and decide what kind of investigation fits it. Does it need a controlled experiment, a field observation, a comparison, repeated measurement or a model? Write what evidence would support the question and what evidence would count against your preferred explanation. This prevents the common mistake of collecting interesting data that never actually answer the question.

Day 3 — Measurement, unit, SI and estimate: measure five objects using appropriate units. Before measuring, estimate each value. Record estimate, measurement and difference. Convert at least two values into another suitable SI unit. Then explain why “12.4” is scientifically incomplete if the quantity requires a unit. Finish by deciding when an estimate is useful and when exact measurement is necessary.

Day 4 — Accuracy, precision and instrument: imagine three measuring devices: one accurate and precise, one precise but inaccurate, and one noisy. Create five readings for each device around a known reference. Decide which problem could improve through repetition and which would require calibration. Then inspect one real instrument and note its smallest scale interval, useful range and one possible reading error.

Day 5 — Qualitative, quantitative, record and table: observe one material or biological specimen. Record five qualitative features and five quantitative measurements or counts if appropriate. Design a table before collecting the second set of observations. Label headings and units clearly. Compare the usefulness of descriptive and numerical data for the same scientific question; neither type should be treated as automatically superior.

Days 6–10 — Build an Investigation That Can Survive Questions

Day 6 — Hypothesis, prediction and variable: choose a relationship that could be investigated safely, such as temperature and dissolving time, surface area and reaction rate, or light intensity and seedling growth. Write one hypothesis and two predictions that follow from it. List every variable that could influence the result. Check that the hypothesis can be challenged by evidence rather than adjusted to fit every possible outcome.

Day 7 — Independent, dependent and controlled variables: mark one independent variable, one dependent variable and at least four controlled variables for your investigation. For each controlled variable, explain why variation could confuse the result. Do not write “keep everything else the same.” Name actual quantities or procedures: 100 mL water, 10 g solute, 30 seconds of stirring, identical container type.

Day 8 — Control group, experimental group and fair test: decide whether your investigation needs a control group. If it does, state the baseline condition and experimental treatment. If it does not, explain why a series of controlled levels of an independent variable is sufficient. Then identify one unfair comparison and repair it. The task is to understand that “fair” means methodologically comparable, not morally fair.

Day 9 — Method, procedure and trial: write the overall method in one sentence, then write a procedure another student could follow without asking you what you meant. Include quantities, instruments, sequence, endpoint and timing. Mark the boundary of one trial. Ask a classmate to identify any ambiguous instruction. If “stir gently” or “heat for a while” appears, replace it with something operational.

Day 10 — Repetition, replication, sample and random: design repeated trials within your own investigation, then describe how another class could independently replicate it. If the investigation uses living organisms, field locations or people, define the population and sample. Decide whether random selection could reduce selection bias. Explain why repetition and replication answer different reliability questions.

Days 11–15 — Data Become Useful Only When You Read Them Properly

Day 11 — Data, result, pattern and trend: create or use a small dataset with at least five levels of an independent variable and repeated results. First describe individual results without interpretation. Then identify any recurring pattern and overall trend. Write one sentence that reports the result and a second sentence that interprets it. Keep the two jobs separate.

Day 12 — Relationship, correlation, cause and effect: examine three paired variables: rainfall and umbrella sales, study time and test score, temperature and dissolving time. Decide which examples show only correlation and which could support causal reasoning if investigated experimentally. For every correlation, generate at least one third-variable explanation. This is practice in resisting causal language that the design has not earned.

Day 13 — Graph, axis and scale: turn a data table into an appropriate graph. Label both axes with variables and units. Choose equal intervals and a scale that uses the graph space effectively without exaggerating differences. Then deliberately make a misleading version by truncating or distorting the scale. Compare how the same data can look different when visual design is poor.

Day 14 — Mean, median and range: calculate all three for at least two datasets: one tightly clustered and one containing an extreme value. Explain how the outlier changes the mean, median and range differently. Decide which summary measure gives the most useful picture for each dataset and why. The goal is not to learn that one statistic is “best” but to match the summary to the data.

Day 15 — Outlier, variation, error and uncertainty: inspect an unusual value. Generate three possible explanations: genuine variation, random measurement error and procedural error. State what evidence would distinguish them. Then write a measurement with a sensible level of uncertainty. Finish by explaining why reporting many unsupported decimal places can create false precision.

Days 16–20 — From Data to Explanation

Day 16 — Analyze, interpret, conclusion and claim: take yesterday’s dataset and analyse it systematically: averages, spread, trend, anomaly. Interpret what the pattern means, then write a conclusion that answers the original question. Convert the conclusion into a precise claim. Underline every word that makes the claim broader than the tested conditions and remove anything the evidence cannot support.

Day 17 — Reasoning, explanation and mechanism: choose one claim and write the evidence beneath it. Then add the reasoning that connects the evidence to the claim. Finally add a mechanism: the intermediate process that explains how the cause could produce the effect. Compare this with a weak answer that simply repeats the data. Strong Science writing builds a bridge rather than stacking disconnected sentences.

Day 18 — Model, theory and law: choose one school Science topic and identify a model used to represent it: particle model, ray diagram, food web, circuit model or cell model. Write what the model includes and leaves out. Then explain the difference between a model, a scientific theory and a scientific law. Make sure your explanation does not say that theories become laws when “proven.”

Day 19 — System, structure, function and process: select a biological or physical system. Draw its components and interactions. For one component, describe its structure and function. Then trace one process through the system from input to output. The task is to move from naming parts to explaining relationships among parts.

Day 20 — Matter, energy, force, change, cycle and equilibrium: take one phenomenon and track the relevant quantities. For melting ice, track matter and energy. For a moving object, distinguish motion from force. For the water cycle, track matter through stages. For thermal equilibrium, identify the opposing transfers that eventually produce no net temperature change. Avoid treating these words as interchangeable “science stuff.”

Days 21–25 — Quality Control: How Science Protects Itself From Easy Mistakes

Day 21 — Validity, reliability and reproducibility: evaluate one investigation three separate ways. Does it measure what the question requires? Are the repeated results consistent? Could another team follow the method and obtain comparable findings? Propose one improvement for each category. A single improvement should not be assumed to solve all three.

Day 22 — Calibration and standard: choose an instrument and identify a suitable reference standard. Explain how calibration would reveal a systematic offset. Then imagine the instrument is precise but inaccurate. Decide whether repeated trials, averaging or calibration provides the correct repair. Link the action to the type of error.

Day 23 — Peer review, research and literature: trace a scientific claim beyond one webpage. Find or imagine the original study, then place it within a wider body of research. Ask what peer reviewers might examine: question, method, sample, analysis, interpretation, limitation. Explain why peer review improves quality control without guaranteeing perfection.

Day 24 — Source, citation, communication and report: take three scientific claims and trace each to a source. Check whether the citation actually supports the wording. Then design a short investigation report with Question, Method, Results, Discussion and Conclusion. Write for a reader who did not watch the experiment and therefore needs enough information to evaluate it independently.

Day 25 — Limitation, assumption and alternative explanation: audit one investigation. List three limitations and explain how each could affect the result. Identify two assumptions that were treated as true. Generate the strongest alternative explanation for the result. Then propose one new test that would discriminate between the preferred and alternative explanations.

Days 26–30 — Integrate the Full Scientific Loop

Day 26 — Falsifiability and revision: write a claim that is testable and one that is protected from every possible result. Explain why the second claim is scientifically weak. For the testable claim, state a result that would force revision. Then practise revising only the failed part of a model or explanation instead of abandoning everything.

Day 27 — Consensus and knowledge: build a confidence ladder from one classroom result to a wider scientific conclusion. Include repetition, independent replication, multiple studies, converging methods, review and synthesis. Explain why scientific consensus is stronger than one authority’s opinion but still open to revision when powerful new evidence arrives.

Day 28 — Full investigation audit: take any experiment from a textbook and audit it using at least thirty target terms. Identify the question, hypothesis, prediction, variables, controls, procedure, trials, measurements, data, graph, uncertainty, possible errors, conclusion, mechanism and limitations. Rewrite one weak part of the method.

Day 29 — Cross-domain transfer: use the same scientific vocabulary in four contexts: a Biology field sample, a Chemistry reaction-rate test, a Physics motion graph and an environmental water-quality study. Notice which terms keep the same logical job even when the subject matter changes. “Controlled variable,” “uncertainty” and “alternative explanation” are portable reasoning tools.

Day 30 — Teach the loop: explain Science to another student without using the phrase “memorise facts.” Your explanation must include observation, question, model, prediction, investigation, measurement, analysis, explanation, criticism, revision and replication. Finish with one sentence stating why scientific knowledge can be dependable without being unchangeable.

The 30-Day Route as a Closed Scientific Loop

The route is intentionally circular rather than linear. Better observations create better questions. Better questions create more valid methods. Better measurements create stronger data. Better analysis creates more disciplined explanations. Better criticism reveals limitations. Better limitations reveal what to test next. Scientific competence grows when students can move around this loop and identify the first weak link rather than merely repeating a memorised “scientific method” sequence.

Part IX — Cross-Subject Science Transfer Missions

Mission 1 — Biology: Sampling, Variation and Structure–Function

Biology makes natural variation impossible to ignore. Two leaves from the same plant are not identical; two seedlings can respond differently under similar conditions; two organisms of the same species may vary in size, behaviour or physiology. That means a biological investigation needs appropriate samples, repeated measurements and cautious conclusions.

Choose a school-garden question such as whether leaf size differs between shaded and sunny areas. Define the population, sampling method and measurement procedure. Use random or systematic sampling rather than selecting only the most convenient leaves. Report mean, median or range where useful, but preserve visible variation instead of forcing every organism into an “average organism” that does not exist.

Then move from measurement to explanation. Connect biological structure to function: thin leaves, root hairs, alveoli, red blood cells, xylem vessels. The important move is not “structure X has function Y” as two memorised facts. It is reasoning about how the structure changes transport, diffusion, strength, surface area or another mechanism.

Mission 2 — Chemistry: Reaction Evidence, Variables and Measurement

Chemistry trains students to separate visible evidence from molecular explanation. Colour change, temperature change, gas production or precipitate formation can be observations. Deciding that a new substance formed is an inference based on those observations and wider chemical knowledge.

Design a safe reaction-rate comparison. Choose one independent variable such as temperature, concentration or surface area. Define a dependent variable such as time to produce a fixed gas volume. Control other relevant factors. Repeat trials and record every result. If the rate changes, explain the mechanism using a particle or collision model while remembering that the model represents invisible processes rather than direct observation.

Chemistry also makes measurement discipline visible. Graduated cylinders, balances, thermometers and gas syringes each have limits. The quality of a chemical conclusion can be restricted by instrument resolution, calibration and the chosen endpoint.

Mission 3 — Physics: Forces, Energy, Graphs and Models

Physics rewards precise distinctions. Force is not motion. Energy is not matter. A graph’s gradient or trend is not the same thing as a causal explanation. Students should use the article’s vocabulary to keep measurement, representation and mechanism separate.

Take a toy car moving down a ramp. Measure travel time at several ramp heights. Plot a graph with correct axes and units. Describe the trend before explaining it. Identify controlled variables such as car, track length and release method. Discuss timing uncertainty and whether human reaction time creates random error.

Then build a model of energy transfer and forces. Ask what the model predicts if ramp height increases, surface friction changes or mass changes. A useful physics model does not merely label arrows; it should help predict a measurable outcome that could prove the model incomplete.

Mission 4 — Earth and Environmental Science: Field Evidence, Systems and Cycles

Earth and environmental science often investigate systems too large, slow or ethically impossible to manipulate in a classroom. Students therefore need observational evidence, sampling, long-term records, models and comparisons across locations or time.

Map a local water, soil, weather or biodiversity question. Define the system boundary. Identify inputs, outputs and cycles. Decide which quantities can be measured directly and which are inferred. Use sampling locations that represent relevant environmental variation instead of choosing only accessible sites.

When two environmental variables correlate, resist instant causal claims. Rainfall, land use, temperature, season and human activity can interact. Environmental explanations often require multiple lines of evidence and models because controlled experiments at ecosystem scale are limited.

Mission 5 — Mathematics and Data: Numbers Are Part of the Evidence, Not a Substitute for Reasoning

Science uses mathematics to summarise and represent evidence, but calculation should not become a ritual detached from the question. A mean, median, range, graph or trend is useful because it reveals something about variation or relationship.

Give students the same dataset and ask three different scientific questions. They may need different summaries or graphs for each. This makes clear that data analysis is goal-directed. The mathematically correct calculation can still be scientifically unhelpful if it answers the wrong question.

Students should also test sensitivity to outliers, scale and missing data. Change one extreme value and see how the mean moves. Change the graph scale and see how visual impression changes. Then return to the raw data. Mathematical representation should make evidence clearer rather than hide the evidence behind one number.

Mission 6 — English: Claim–Evidence–Reasoning as Scientific Writing

Scientific writing is also language work. A strong answer must distinguish claim, evidence and reasoning. The claim answers the question. The evidence identifies relevant observations or data. The reasoning explains why those data support the claim using scientific ideas.

Take a weak answer such as “The plant grew more because the fertiliser worked.” Rewrite it. State the measured difference, identify the controlled comparison, explain a plausible biological mechanism, acknowledge uncertainty and avoid claiming beyond the tested conditions. Scientific vocabulary improves sentence precision because it improves conceptual precision.

Source and citation language also transfers directly to English reading. A source can be relevant but weak; a citation can exist but fail to support the sentence. Students who learn to inspect claim-to-source relationships in Science become stronger critical readers everywhere.

Mission 7 — Technology and AI: Models, Bias, Evidence and Verification

The scientific method vocabulary transfers naturally into technology and AI literacy. An AI model makes outputs that can be evaluated. A dataset can contain sampling bias. A benchmark is a kind of standardised test. Repeated trials reveal variation. A confident output still requires verification.

Choose one AI or digital tool and treat it as an object of investigation. Define one measurable capability, create a test set, record outputs, classify errors and report uncertainty. Avoid changing prompts differently for successful and unsuccessful trials unless the experiment explicitly studies prompting. In other words, apply fair-test thinking to software.

Then distinguish model performance from causal claims about human learning or productivity. If students using a tool score higher, that is not automatically evidence that the tool caused the improvement unless the study design controls major alternatives. Scientific reasoning protects technology discussions from easy overclaiming.

Mastery Diagnostic — Five Levels of Secondary 1 Science Vocabulary Ownership

Level 1 — Recognition: the student can match terms such as observation, hypothesis, variable, outlier and model to broadly correct meanings when options are visible.

Level 2 — Retrieval: the student can define a term without looking, give an original scientific example, use a natural collocation and identify the term inside an investigation.

Level 3 — Distinction: the student can separate close concepts: observation versus inference, hypothesis versus prediction, control group versus controlled variable, accuracy versus precision, repetition versus replication, correlation versus causation, result versus conclusion, theory versus law, validity versus reliability.

Level 4 — Application: the student can use the vocabulary to redesign a weak investigation, diagnose a measurement problem, interpret a graph, justify treatment of an outlier, write a claim–evidence–reasoning explanation and state meaningful limitations.

Level 5 — Transfer and scientific regulation: the student can use the same concepts across Biology, Chemistry, Physics, Earth Science, fieldwork, data analysis and technology evaluation; can decide what evidence deserves trust; and can revise a model or conclusion when new evidence demands it.

The Ten Master Questions for Any Scientific Investigation

1. What exactly is the question? Is it specific enough that we know what evidence could answer it?
2. What was observed, and what was inferred? Keep direct evidence separate from interpretation.
3. Which variables matter? Identify what changes, what is measured and what must be controlled or compared.
4. Does the measurement validly represent the thing named in the question? Easy-to-measure is not automatically correct-to-measure.
5. Is the comparison fair enough to test the intended relationship? Look for uncontrolled alternatives, weak control conditions and inconsistent procedures.
6. Are there enough trials or a sufficiently representative sample? One result can be dominated by variation, chance or selection bias.
7. What do the raw data show before interpretation? Inspect pattern, spread, outliers, graph scale and uncertainty before writing the conclusion.
8. Does the evidence show correlation, causation or only a tentative association? Let the design determine the strength of the causal language.
9. What alternative explanation, error or limitation could still produce this result? Strong conclusions survive serious alternatives rather than ignoring them.
10. How much confidence should this result receive? Consider validity, reliability, reproducibility, replication, literature and the wider evidence base.

How This Article Connects to the eduKateSG Science Ecosystem

Use the Vocabulary Learning Hub for the wider vocabulary route. For the broad nature of scientific knowledge, continue through What Is Science? and What is Science | The Replaceable Logic. Those pages explore the larger scientific system while this page remains the Secondary 1 vocabulary owner.

For investigation design, continue to How Investigable Questions Work in School Science | Why the Question Determines What Evidence You Can Collect. For evidence and epistemic standards, use Why Students Need to Know What Counts as Evidence. These deeper pages explain why vocabulary such as question, evidence, validity and inference changes the structure of scientific reasoning.

For explanation and representation, continue to How Model-Based Reasoning Works | Why Science Learners Need Models They Can Test, Break and Revise. For writing the bridge from observations to explanation, use Claim–Evidence–Reasoning | How Observations Become Explanations.

The wider How X Works Hub can then take students into mechanisms across Biology, Chemistry, Physics, technology, civilisation and complex systems. The important transfer is that scientific vocabulary keeps its logical job even when the subject changes: variables still require control, evidence still has limits, models still simplify, and conclusions still deserve confidence only in proportion to the quality of the evidence.

Closing Principle — Science Becomes Teachable When Students Can Name Where Evidence Ends and Inference Begins

A student who knows many scientific facts can still struggle if the language of investigation remains vague. “It changed” is weaker than identifying the dependent variable. “The experiment worked” is weaker than explaining validity and reliability. “The graph proves it” is weaker than distinguishing a trend from a causal mechanism. “Everyone agrees” is weaker than explaining replication, literature and scientific consensus.

Precise vocabulary gives students a diagnostic map. If an experiment fails, they can ask whether the question was investigable, the variable was controlled, the instrument was calibrated, the sample was biased, the outlier was mishandled, the graph was misleading, the conclusion exceeded the evidence or the model needed revision. Each word names a place where scientific quality can be strengthened.

That is why this list is not really about one hundred isolated definitions. It is about learning the grammar of scientific thought: observe before inferring, measure before claiming, compare before attributing cause, model before explaining, test before trusting, criticise before generalising, and revise when evidence says the world is different from the story in your head.

When students can do that, Science stops being a collection of answers produced by other people. It becomes a method for asking the world better questions—and knowing what kind of evidence would make an answer worth believing.

Part X — The Secondary 1 Science Investigation Operating Manual

A vocabulary list becomes a scientific operating system only when a student can use the words to repair a weak investigation. This section turns the one hundred terms into six practical modules. Each module begins with a common school-science failure and follows the failure backward until the first weak link becomes visible. The purpose is not to add more terminology. It is to make the existing vocabulary do real work.

The operating principle is simple: Read → Diagnose → Prioritise → Repair → Practise → Connect → Perform → Review. Read what the investigation claims to do. Diagnose where evidence becomes weak. Prioritise the earliest weakness that can corrupt everything after it. Repair that point. Practise the repaired move. Connect it to the wider scientific system. Perform the investigation or explanation. Review what the evidence now allows you to say.

Module A — Repair the Question Before Repairing the Experiment

Many poor practicals begin with equipment instead of a question. Students see beakers, lamps, plants or ramps and immediately ask what they can do with them. Scientific investigation works better in the opposite direction. The question determines what evidence is relevant; the evidence determines what must be measured; the measurement determines what apparatus and procedure are needed.

Consider the question, “Does exercise make people healthier?” It sounds scientific because it mentions a possible cause and effect, but it is not yet operational. What counts as exercise—walking, sprinting, duration, frequency, intensity? What counts as healthier—resting heart rate, mood, body composition, sleep, blood pressure, endurance? Over what period? Which people? Without these decisions, the word healthier hides several dependent variables and the word exercise hides several independent variables.

A Secondary 1 repair is not to make the question complicated. It is to make the evidence pathway visible. “How does five minutes of step exercise at a fixed pace affect heart rate immediately after exercise in students of the same class?” is narrower. It identifies a treatment, a measurement and a time point. It still has limitations, but a reader can now imagine what evidence would answer it.

Now compare “Which material is the best insulator?” The word best could mean lowest heat transfer, lowest cost, lightest mass, safest handling or smallest thickness. A fair scientific test needs a defined performance criterion. “Which 5 mm material layer produces the smallest temperature drop in 100 mL of warm water over ten minutes?” turns an evaluative word into a measurable outcome.

The question also determines the permissible conclusion. If the test compares only cotton, aluminium foil and bubble wrap at one thickness, the conclusion should not become “bubble wrap is the world’s best insulator.” It can become “of the three materials tested at the stated thickness and conditions, bubble wrap produced the smallest temperature decrease.” Scope is part of scientific precision.

Iona uses a four-check question audit. Observable: can the relevant outcome be detected or measured? Variable: is the factor being compared or changed clear? Boundary: are population, conditions or time period defined enough? Discriminator: could different possible results favour different explanations? A question that fails these checks often produces a practical that looks busy but cannot support a strong conclusion.

Maren then turns the question into a hypothesis only after the variables are clear. The hypothesis should state the proposed relationship or explanation. The prediction should specify the expected observation if that relationship is correct. This order prevents the student from writing a prediction such as “Plant A will grow tallest” without stating what scientific idea makes that result meaningful.

Leonie adds feasibility and safety. A question can be scientifically interesting but unsuitable for a school investigation because the time, instrument, sample or hazard makes valid evidence impossible. Good investigation design includes knowing when not to run the test. The scientific alternative may be to use existing data, a simulation, a safer proxy measurement or a field observation.

Repair drill: rewrite each vague question into an investigable form: “Does music help concentration?”, “Which drink is healthiest?”, “Does sunlight help plants?”, “Are phones bad for sleep?”, “Which surface has the most friction?” For every repaired question, identify the outcome measure and one limitation introduced by narrowing the question. Narrowing is not a weakness; it is how a claim becomes testable.

Module B — Build a Measurement Architecture, Not Just a Results Table

Once the question is clear, the next weak link is often measurement. Students can identify the dependent variable correctly and still measure it badly. “Measure growth” is not a measurement architecture. The investigator must decide what quantity represents growth, which instrument measures it, where the reference point is, when readings are taken, how units are recorded and what level of precision the instrument supports.

Imagine an investigation of paper-towel absorbency. One group pours water on a sheet and writes “very absorbent.” Another measures the mass before and after water absorption. A third measures the maximum volume retained before dripping. These methods operationalise “absorbency” differently. None is automatically correct. The best method depends on the intended meaning of absorbency and the practical conditions.

This is a validity question before it is an accuracy question. A perfectly calibrated balance can measure mass accurately, but if the real question concerns the rate at which a towel absorbs a spill, final retained mass may not represent the target property well. Validity asks whether the measure corresponds to the scientific construct named in the question.

Accuracy and precision enter after the measure is chosen. Suppose a thermometer reads consistently 1.5°C high. The readings may be precise but inaccurate. Suppose another thermometer fluctuates by ±2°C around the correct temperature. Its average may be close to the reference while individual readings lack precision. The corrective actions differ: calibration for systematic offset; improved technique, more stable conditions or repeated measurements for random variability.

Leonie teaches the instrument–range–resolution–reference check. What quantity does the instrument measure? Does its range cover the expected values? What is the smallest meaningful division or displayed increment? Is there a reference or calibration check? A 30 cm ruler cannot directly measure a 0.02 mm thickness with credible precision simply because the student writes many decimal places.

Endpoint definition matters too. In a reaction-rate experiment, when exactly has the reaction “finished”? In dissolving, when exactly is a solute “dissolved”? In plant germination, what counts as “germinated”? If the endpoint remains subjective, different observers can generate systematic differences. Operational definitions turn vague events into reproducible decisions.

Iona adds uncertainty explicitly. Every measurement is constrained by instrument resolution, method and observer. Students do not always need advanced uncertainty calculations, but they should understand that a reading is not infinitely exact. A thermometer marked every 1°C does not justify 23.000°C. A stopwatch capable of milliseconds does not make a human reaction-time experiment accurate to the millisecond.

Maren connects measurement to communication. A results table should expose the architecture: variable names, units, trial structure and enough raw data to inspect variation. If the table contains only the mean, the reader cannot see outliers or spread. If it contains only raw data with no organisation, the pattern may remain hidden. Strong scientific communication preserves inspectability while making structure visible.

Repair drill: choose one vague dependent variable—fitness, cleanliness, reaction speed, plant health, water quality or brightness. Propose three possible operational measurements. For each, identify instrument, unit, likely uncertainty and one way the measure could be valid for one question but invalid for another.

Module C — Debug the Experiment Before Blaming the Data

Students often discover a messy graph and try to “fix the data.” Science asks a different question first: what in the design could have produced the mess? The goal is not to make the graph pretty. The goal is to understand whether variation reflects the natural system, random measurement error, systematic bias, uncontrolled variables or a genuinely wrong hypothesis.

Start with controlled variables. A controlled variable matters only if it could plausibly affect the dependent variable. In a cooling experiment, container material, water volume, starting temperature, surface area and room conditions may matter. Writing a long list of irrelevant constants does not improve validity. Controlling the first causal alternatives does.

Next examine sequence effects. Suppose students test low temperature first, then medium, then high, using the same apparatus. The equipment may warm over time. The order of trials becomes a hidden variable. Randomising order or allowing the apparatus to return to a standard state can reduce this problem. Experimental design is often about controlling time-dependent changes that are not visible in a simple variables table.

Then examine observer effects. If one student measures all treatment samples and another measures all controls, differences in technique can become confounded with the treatment. Using a standardised method, training observers or blinding labels in some contexts can reduce observer bias. At school level, the core idea is that who measures and how they know the condition can matter.

Sampling can corrupt a design before the apparatus is touched. A biodiversity survey taken only beside the path may overrepresent species that tolerate disturbance. A survey about sleep sent only to students who choose to answer may overrepresent particular habits. Random, systematic or stratified sampling can reduce some selection problems, but each requires the population and claim to be clear.

Repetition addresses random variation but not every bias. Ten repetitions of the same biased sample remain biased. Fifty readings from an uncalibrated instrument remain systematically wrong. This is why “do more trials” is not a universal method improvement. The repair must match the failure mechanism.

Iona uses an error taxonomy: random errors vary unpredictably among trials; systematic errors shift results consistently; procedural errors arise because the method was executed differently from its intended design; sampling errors arise because selected cases do not adequately represent the target; model errors arise when the conceptual explanation omits a relevant mechanism.

Outliers become diagnostic clues inside this taxonomy. An outlier can indicate a recording mistake, an unusual case, a procedural event or a phenomenon the model failed to anticipate. The scientific response is to investigate and document. Deleting first and asking later destroys information.

Leonie closes the debug loop by changing one thing at a time in the investigation design. If the suspected issue is reaction-time error, automate timing or measure a longer interval. If the suspected issue is temperature drift, monitor temperature continuously. If the suspected issue is sample bias, redesign sampling. Repair should target the earliest weak link that can explain the observed failure.

Repair drill: imagine an experiment with highly variable results. Generate one random-error explanation, one systematic-error explanation, one uncontrolled-variable explanation, one sampling explanation and one model explanation. For each, design a diagnostic test. The aim is to transform “our experiment went wrong” into a set of hypotheses about why.

Module D — Decide Whether the Evidence Deserves Causal Language

One of the most important scientific language decisions is whether to write “is associated with,” “affects,” “causes,” or “is consistent with.” These phrases are not stylistic decoration. They encode how much causal confidence the design deserves.

Observational data can reveal correlations. If students who sleep more tend to score higher, sleep and score are associated. But prior achievement, stress, exercise, family routines or study time may influence both. Reverse direction is also possible: students doing well may experience less stress and sleep better. The correlation alone cannot choose among these causal structures.

A controlled experiment strengthens causal inference because the investigator deliberately changes one factor and holds relevant alternatives stable. Yet even experiments have boundaries. A short laboratory test may demonstrate a causal effect under specific conditions without proving the same effect across every population or environment.

Mechanistic evidence strengthens the story in another way. If temperature changes reaction rate and a particle-collision model explains how temperature changes collision frequency or energy, the measured effect and mechanism support one another. But a plausible mechanism alone cannot replace empirical evidence. A story that “makes sense” can still be wrong.

Temporal order matters: causes precede their effects. Dose–response patterns can matter: stronger exposure sometimes produces stronger outcomes. Replication matters: effects that recur across independent tests deserve more confidence. Converging methods matter: a laboratory experiment, field observation and mechanistic model pointing in the same direction can be stronger than one method alone.

Maren uses a causal-language ladder. Observed together is weakest. Associated with identifies a relationship. Consistent with an effect indicates evidence that fits a causal interpretation but leaves alternatives. Affected or caused requires a design and evidence base capable of supporting stronger causal inference. Students should climb only as high as the evidence permits.

Iona adds counterfactual reasoning: to claim that X caused Y, we are implicitly comparing what happened with X to what would have happened under a similar condition without X. Control groups, matched conditions and experimental manipulation are methods for approximating that otherwise unobservable comparison.

Leonie then checks external validity. A causal result in one narrow setup may not generalise. Sugar dissolving faster at higher temperatures in water does not automatically establish the same magnitude of effect for every solvent and solute. Plant growth under one lamp system does not automatically transfer to field agriculture. Causation and generalisation are separate questions.

Repair drill: rewrite each claim at three confidence levels: “screen time causes poor sleep,” “fertiliser increases plant growth,” “higher temperature speeds dissolving,” “exercise improves memory.” For each, state what study design would justify moving from association to stronger causal language.

Module E — Stress-Test the Explanation, Not Just the Result

After a result appears, students often treat the explanation as finished. Scientific thinking becomes stronger when the explanation is treated as another object to test. A good explanation should connect observations to a mechanism, make predictions, survive alternative explanations and reveal what evidence could prove it incomplete.

Take rusting. Observation: reddish-brown material appears on iron exposed to certain conditions. Explanation: iron reacts with oxygen in the presence of water to form rust products. The explanation should make predictions. Iron kept dry should rust less. Removing oxygen should reduce rusting. Coatings that prevent contact should slow the process. Predictions turn an explanatory story into a testable model.

Alternative explanations matter. Suppose a coated nail rusts less. Is the coating chemically inhibiting the reaction, or simply blocking water and oxygen? Different mechanisms can produce the same result. A new investigation can separate them. This is why the strongest alternative explanation is useful: it tells us what experiment to run next.

Assumptions must also be visible. A model of projectile motion may assume air resistance is negligible. A food-web diagram may assume the important feeding relationships are adequately represented. A particle diagram may ignore intermolecular detail. Simplification is necessary, but hidden assumptions can make a model appear universally valid when it is not.

Falsifiability provides a discipline. Ask: “What observation would make us revise this explanation?” If no possible result can count against it, the explanation risks becoming immune to evidence. Scientific models earn trust partly by making risky predictions that could fail.

When a prediction fails, revision should be targeted. Perhaps the model’s boundary conditions were wrong. Perhaps one omitted variable matters. Perhaps the mechanism needs an extra stage. Perhaps the measurement failed instead. Good revision does not automatically throw away every useful part of the model.

Maren teaches a four-column model audit: Represents—what does the model include? Omits—what does it simplify or ignore? Predicts—what result should occur if it is useful? Breaks—under what conditions does it fail? This turns model-based reasoning into an inspectable practice.

Iona adds comparison among models. Two explanations may both fit current data. The next test should be chosen for diagnostic value: a result that one model predicts and the other does not. Science progresses faster when experiments are designed to discriminate among live alternatives rather than merely collect more examples consistent with the favourite idea.

Repair drill: choose a model used in Secondary 1 Science. Create the four-column audit. Then write one competing model or alternative explanation and design a decisive observation that would favour one over the other.

Module F — Write the Report as an Audit Trail of Reasoning

A strong science report is not a ceremonial format. It is an audit trail. A reader should be able to follow the chain from question to method to raw evidence to interpretation and see where uncertainty enters. If the report hides the chain, the reader cannot evaluate the claim independently.

The question defines the scientific job. The method explains how evidence was generated. The results preserve observations and measurements. The analysis organises patterns and variation. The discussion interprets those patterns, considers alternatives and limitations. The conclusion answers the question at a confidence level appropriate to the evidence.

Results should not be rewritten to match expectations. Unexpected values belong in the record. If a value is excluded, the criterion and reason should be stated. If a procedure changed mid-experiment, that change should be recorded. Transparency allows later readers to diagnose what the original investigator could not see.

The discussion should not become a list of generic errors: “human error,” “do more trials,” “use better equipment.” Each limitation should be causal. If reaction time adds random timing variation, use automated timing or measure a longer interval. If heat escapes differently because containers are unequal, standardise containers. If sampling is biased, redesign selection. Improvements should target specific failure mechanisms.

Source use follows the same discipline. A citation should support the sentence to which it is attached. A general website describing photosynthesis does not automatically support a numerical claim about a particular experiment. Primary research, reference sources, textbooks and reviews serve different purposes. Scientific literacy includes choosing a source appropriate to the claim.

Uncertainty language should be deliberate. “Proves” is rarely appropriate for one school investigation. “Supports,” “is consistent with,” “suggests,” “under the tested conditions,” and “within the measurement uncertainty” can make writing more accurate, not weaker. Confidence becomes credible when it is calibrated rather than absolute.

Peer review can be simulated in class. Exchange reports. The reviewer does not merely correct grammar. The reviewer asks: Can the question be answered by this method? Are the variables identifiable? Could I reproduce the procedure? Are raw data visible? Does the graph represent the data honestly? Does the conclusion overreach? Which alternative explanation remains?

Maren then asks the author to respond to review. A strong response is not defensive. It distinguishes misunderstandings caused by unclear writing from genuine weaknesses in the investigation. The report is revised accordingly. This is a small-scale version of scientific error correction.

Finally, the report should connect one investigation to the wider literature. Does the result agree with established knowledge? If not, is there a plausible method reason, a context difference or a genuinely interesting contradiction? One classroom experiment rarely overturns a large evidence base, but it can reveal exactly how scientific claims are connected to reproducible observations.

Repair drill: take a completed practical report and annotate every sentence with its scientific job: question, procedure, observation, result, analysis, inference, claim, evidence, reasoning, limitation or conclusion. Find any sentence that performs two incompatible jobs at once. Rewrite it so the evidence chain becomes visible.

The Operating Manual in One Page

• Question: define the measurable scientific job before choosing apparatus.
• Measurement: make the dependent variable valid, observable, unit-bearing and appropriately precise.
• Design: control the first plausible alternative causes, not an arbitrary list of details.
• Data: preserve raw results, variation and anomalies before summarising.
• Analysis: choose graphs and statistics that answer the question rather than decorate the report.
• Causation: let the design determine the strength of causal language.
• Explanation: connect evidence to a mechanism and test the strongest alternative.
• Model: make assumptions, predictions, boundaries and failure conditions visible.
• Quality: separate validity, reliability, accuracy, precision and reproducibility.
• Communication: write an audit trail that another person can inspect, repeat and criticise.
• Revision: change the model, method or claim when evidence demands it.

The deepest Secondary 1 Science habit is therefore not “follow the scientific method.” It is find the first weak link. A perfect graph cannot repair an invalid measurement. Repetition cannot repair systematic bias. A detailed explanation cannot repair missing evidence. A strong student learns to move upstream until the earliest failure becomes visible, repairs it, and then lets the rest of the investigation rebuild from that stronger foundation.