Teaching Scientific Graphs with an Original Dataset

Teach scientific graphs with an original dataset

A graph lesson should expose the decisions between a table and a conclusion. This worked teaching pack uses a simulated liquid-level display, a completed graph, a deliberately poor graph and a separate learner run. Learners choose axes and scales, plot coordinates, estimate within the measured interval and explain the limits of their conclusions. It is an optional worked pack within the wider Teacher Resource Packs 2027 collection; it does not describe every resource in that collection.

The dataset and its boundary

Imagine a virtual display that reports the height of a liquid surface above a fixed reference mark. Run A follows the supplied model h = 2 + 1.5t, where t is seconds and h is centimetres, only for 0 to 12 seconds. The model stops at 12 seconds and 20 cm. Nothing here establishes what would happen after that boundary. These numbers are constructed for teaching, not recorded laboratory results. Learners do not need to fill a container or operate equipment.

Time is the chosen horizontal variable. Liquid level is the vertical variable. We are graphing height above a reference, not volume, total liquid added or flow in cubic centimetres per second. A 1.5 cm/s increase in level is a height-change rate. Converting it into volume flow would require additional information about the container’s cross-sectional area.

A completed planning template and lesson sequence

Plan approximately 35 minutes: five to identify quantities and units, eight to choose and demonstrate scales with Run A, six to compare the poor graph with the repaired graph, ten for independent Run B work and six to review evidence and limitations. These are illustrative planning allowances. The success check is a defensible graph and explanation, not completion at a prescribed speed.

Before plotting, ask learners what a point means. “(4, 8)” means that at 4 s the model level is 8 cm. It does not mean eight squares up until the scale has been declared. Record a prediction about the general pattern, but do not let that prediction replace checking all the coordinates. A visually straight line can still have incorrect labels or units.

Demonstration: choose a useful scale

The completed Run A graph uses 0–12 s horizontally and 0–20 cm vertically. Major horizontal ticks represent 2 s; major vertical ticks represent 5 cm. The minor grid gives 1 s and 1 cm. Equal distances along an axis represent equal changes in its quantity. The scales on the two axes need not match because their units and ranges differ.

All points fit, and the zero origin makes the nonzero starting level easy to see. In some scientific graphs a clearly labelled nonzero vertical start is useful, but it must not disguise the displayed range. Here there is enough space to begin at zero without compressing the pattern. Explain that this is a reasoned choice for this dataset rather than a rule that every graph must use the same bounds.

Model (6 s, 11 cm): move to 6 on the time axis, then to 11 on the height scale. Put a small cross at the intersection. Plot the other six pairs the same way. The supplied model is linear, so draw one straight line through its plotted points within 0–12 s. For noisy measurements, an appropriate fitted line or curve might not pass through every point. Do not teach automatic dot-to-dot joining for all scientific data.

The poor graph and its repair

The deliberately poor version places vertical labels 0, 2, 5, 8, 11, 14 and 20 at equal physical gaps, omits the vertical unit and titles the picture “Faster and faster”. This distorts the numerical scale: equal gaps represent changes of 2, 3, 3, 3, 3 and 6 units. Even if every label looks tidy, the coordinate system is unreliable.

Ask learners to name the error and the consequence before repairing it. A useful response is “the final vertical gap covers six centimetres while most other gaps cover three, so the visual slope is misleading”. Feedback can then point to the repaired scale with constant one-centimetre minor intervals. Restore “Liquid level (cm)” and use a descriptive title such as “Simulated liquid level over time”. A straight rising graph here means a constant height-change rate, not acceleration.

Interpretation with working and units

Between 2 s and 10 s, the level rises from 5 to 17 cm. Average height-change rate = (17 − 5)/(10 − 2) = 12/8 = 1.5 cm/s. Subtract both coordinates. The tempting answer 17/10 = 1.7 uses the final height as if the starting height were zero, and it ignores the interval in the question. On this straight-line model, all nonzero intervals have the same gradient.

At 5 s, interpolate between (4 s, 8 cm) and (6 s, 11 cm). Five is halfway between four and six, so the estimate is 8 + 0.5 × (11 − 8) = 9.5 cm. The model equation confirms the same value. Reading a paper graph to 9.5 cm is reasonable if the grid supports that precision; reporting 9.50000 cm would not make the estimate more reliable.

A bounded conclusion is: “In the simulated Run A interval from 0 to 12 s, liquid level increases linearly from 2 to 20 cm, at 1.5 cm/s.” The graph alone does not identify a pump, prove a physical mechanism or show indefinite increase. Extending the formula to 14 s gives 23 cm algebraically, but the supplied model is defined only to 12 s. That extension is unsupported by this lesson’s evidence.

Evaluation: resolution and repeat readings

For an additional measurement discussion, imagine a display rounded to the nearest 0.5 cm. Rounding alone can differ from the underlying value by up to about 0.25 cm. For two readings, the worst-case rounding contribution to a difference can be about 0.50 cm because one could round upwards while the other rounds downwards. This is a bound under the stated rounding model, not a complete uncertainty analysis.

Across the eight-second interval used above, 0.50/8 = 0.0625 cm/s is the corresponding worst-case contribution to the calculated gradient if time is treated as exact. It does not include timing error, calibration, changing conditions or systematic offset. Choosing a longer interval can reduce the relative effect of the same reading resolution, but that choice must still describe the interval the question asks about.

Three additional simulated readings at 6 s are 10.5, 11.0 and 11.5 cm. This separate example includes reading variation beyond rounding; it is not the exact Run A formula sampled three times. Their mean is 33.0/3 = 11.0 cm and their range is 11.5 − 10.5 = 1.0 cm. The spread illustrates repeatability, not proof of accuracy. Repeating readings does not automatically remove a systematic offset. Ask what further check would help, such as checking the reference mark against a known length standard in the hypothetical measurement setup.

Independent learner tasks and feedback

1. G1, plotting: Plot all seven Run B points on the supplied blank grid. Choose and label the axes, units and regular scales. Answer: the pairs are (0,1), (2,3), (4,5), (6,7), (8,9), (10,11) and (12,13). They lie on h = 1 + t over the given interval. Accept other clear, regular scales that include every point.
2. G2, interpretation: Estimate Run B at 5 s and calculate its average rate from 2 to 10 s. Answer: 6 cm by interpolation; (11 − 3)/(10 − 2) = 1.0 cm/s. A value of 1.1 cm/s reveals the final-value-over-final-time error.
3. G3, evaluation: Critique “Run A rises faster, so it must contain more liquid and will keep rising forever”. Answer: A has the larger height-change rate in the supplied interval, 1.5 versus 1.0 cm/s. Volume needs container geometry and a reference amount. Neither run supplies evidence beyond 12 s.
4. G4, repeats: Find the mean and range of the three six-second readings, then name a remaining limitation. Answer: 11.0 cm and 1.0 cm. A shared calibration offset could remain despite repeated readings.

Use the separate answer guide to award evidence for the learner’s graph, working and explanation. If a learner needs spoken data or a larger grid, keep the quantities and reasoning available. A table can preserve interpretation tasks, but it does not itself demonstrate visual plotting. For that target, agree an appropriate accessible plotting format rather than claiming that a spoken answer proves a graph was drawn.

After the lesson, record a specific next step: for example, “G1 points correct but unit missing; require axis labels before joining” or “G2 used final coordinates without subtraction; practise interval endpoints”. These observations are more useful than a general comment that graphs were difficult.

Accessibility background: Home Office guidance on image alternatives. The simulated data, graphs, tasks and teaching decisions are original.

Pack contents

• Scientific graphs learner worksheet: three pages, in editable Word and PDF.
• Scientific graphs answer guide: two pages, in editable Word and PDF, matched to G1–G4.
• Graph demonstration sheets: a two-page PDF with the completed and deliberately poor Run A graphs.
• Editable CSV datasets and original SVG/PNG graph companions, including the blank grid and Run B answer graph.