Communicating evidence-based physics conclusions

A physics conclusion is not a retelling of a measurement but an argument that links a claim to evidence through a physical law (reasoning). The working pattern is claim → evidence → reasoning → uncertainty. For a pendulum, the claim "the period grows with length" is supported by a table of measurements and the relation $T = 2\pi\sqrt{L/g}$. Data are quoted as numbers with units and uncertainty: $T = 2.00 \pm 0.02\ \mathrm{s}$ at $L = 1.00\ \mathrm{m}$. Observation is kept separate from interpretation: "the measured value agrees with the prediction within uncertainty" is a conclusion, while "the theory is proven" is overclaiming. The strength of a conclusion is set by the quality of the data: the number of repeats, the spread, systematic effects and the conditions of the experiment. A conclusion therefore states its range of validity and its limitations.

A student times a pendulum at two lengths (one trial each, instrument: $\pm 0.01\ \mathrm{s}$):

Arrange the parts of the conclusion using the pattern claim → evidence → reasoning → uncertainty, and write the conclusion.

From a series of measurements a graph of $T^2$ against $L$ is plotted. The points fall on a straight line with slope $k = 4.00 \pm 0.05\ \mathrm{s^2/m}$. Determine $g$ with its uncertainty and state a conclusion that is grounded in the data.

A classmate writes: "We measured the period of a pendulum of length $1.00\ \mathrm{m}$ three times and got $2.00\ \mathrm{s}$, $2.01\ \mathrm{s}$, $1.99\ \mathrm{s}$. This proves that the formula $T = 2\pi\sqrt{L/g}$ is correct." Identify the overclaiming and unsupported parts, then rewrite the conclusion so that it is evidence-based.

A safe self-study plan using a simulation (no real masses are hung).

1. Open a pendulum simulation (for example, the PhET "Pendulum Lab") and set a small angle of no more than $15^\circ$; keep the mass and the material of the bob unchanged.
2. Choose lengths $L = 0.20,\ 0.40,\ 0.60,\ 0.80,\ 1.00\ \mathrm{m}$. For each length take 3 period measurements and record a table: $L$, three values of $T$, the mean $T$, and the spread.
3. Plot $T^2$ against $L$ with vertical spread bars and find the slope $k$ with an uncertainty estimate.
4. Write a four-part conclusion: claim (what you assert), evidence (the slope and $g = 4\pi^2/k$ with its uncertainty), reasoning (the relation $T = 2\pi\sqrt{L/g}$), uncertainty and limitations (number of repeats, range of $L$, small angle). Use verbs such as "agrees with" and "supports", not "proves".

To show that an evidence-based conclusion has a checkable structure: each part of the conclusion must be tied to a specific source (table, graph, law) or to an explicitly named limitation.

A four-block "skeleton" diagram of a conclusion, joined by arrows from left to right: Claim → Evidence → Reasoning → Uncertainty and limitations. Under the Evidence block sits a miniature data table and a graph of $T^2$ against $L$ with points and spread bars. Under the Reasoning block sits the formula $T = 2\pi\sqrt{L/g}$. A feedback arrow runs from Uncertainty and limitations back to Claim, labelled "refine or weaken the claim". To the right of the diagram are two short label examples: "supports within $\pm 0.12\ \mathrm{m/s^2}$" (green) and "proves" (red, crossed out).

You time the same runner over $100\ \text{m}$ three times with a phone stopwatch: $13.2\ \text{s}$, $13.5\ \text{s}$, $12.9\ \text{s}$. The mean is $13.2\ \text{s}$ and the spread is $0.6\ \text{s}$. A useful report states: 'mean speed $= \frac{100\ \text{m}}{13.2\ \text{s}} = 7.6\ \text{m/s}$, but because reaction time varies by about $\pm 0.3\ \text{s}$ the speed lies between $7.4\ \text{m/s}$ and $7.8\ \text{m/s}$'.

This exercises the core of evidence-based communication: the claim ('the runner's speed is ...') is separated from the evidence (the three measured times), and a range is reported instead of a single number. It also forces you to name the source of uncertainty — human reaction time — rather than hide it.

On a dry road a car braking from $50\ \text{km/h}$ ($13.9\ \text{m/s}$) stops in $14\ \text{m}$; on a wet road the same car stops in $22\ \text{m}$. From $s = \frac{v^2}{2a}$ the mean decelerations are $a_{\text{dry}} = \frac{(13.9)^2}{2 \cdot 14} = 6.9\ \text{m/s}^2$ and $a_{\text{wet}} = \frac{(13.9)^2}{2 \cdot 22} = 4.4\ \text{m/s}^2$. Reporting these numbers requires naming the assumptions too: constant deceleration, no reaction time included, and one measurement per road condition.

This exercises stating assumptions and limitations explicitly, distinguishing a measured quantity ($s$) from one inferred from it ($a$), and warning that a single trial per condition cannot establish a general law — exactly the honesty required when communicating a physics conclusion.

Three thermometers at $1.5\ \text{m}$ height on the school field read $21.4\ ^\circ\text{C}$, $21.6\ ^\circ\text{C}$, $21.3\ ^\circ\text{C}$. The mean is $21.4\ ^\circ\text{C}$ and the range is $0.3\ ^\circ\text{C}$. The nearest official station reads $20.9\ ^\circ\text{C}$ over grass. A complete conclusion is: 'air temperature here is $21.4 \pm 0.2\ ^\circ\text{C}$; this is $0.5\ ^\circ\text{C}$ above the station value, which is consistent with our site being over asphalt and $2\ \text{m}$ above a warm surface'.

This exercises comparing your value with an independent source and interpreting the difference as either a real site effect or a measurement discrepancy; it uses the reporting habit of quoting a mean with an uncertainty and offering a physical reason for disagreement instead of ignoring it.