Paper 3 Advanced Practical Skills

Assemble securely with scales visible, electrical polarity/ranges correct and a switch available. Check zero, alignment and that the intended variable can change without disturbing controls.

Take a pilot reading at low risk. Confirm the response is measurable and within instrument range; adjust the planned interval/range before collecting the full series.

Collect an appropriate spread of paired values—typically at least six where the instructions allow—covering the widest safe useful range. Let readings stabilise and repeat where scatter is plausible.

Repeat readings where random variation is expected, then calculate a mean. For dimensions, repeat at different positions/orientations to sample non-uniformity.

Reduce scatter at source: view scales normally, wait for steady readings, release moving systems consistently, avoid parallax and keep control variables fixed.

Judge quality by scatter about the trend, agreement of repeats and—only when provided—supervisor or accepted values. Investigate an outlier by repeating the same condition before excluding it.

Record raw readings immediately and to consistent precision for the same instrument: metre-rule values share decimal places; digital displays retain all shown digits.

Calculated values normally use the same number of significant figures as, or one more than, the least precise measured input. Preserve guard digits during working, then round the recorded result.

Show one representative calculation outside the table with substitution and units. For an uncertainty column, state absolute uncertainty with the same unit and compatible decimal place as the value.

Choose simple linear scales (1, 2 or 5 units per major interval) that use at least about half the available grid in both directions. Avoid awkward factors such as 3 per large square.

Plot small crosses or fine points accurately—typically within half a small square. Include uncertainty bars only when required and check every coordinate against the table.

Draw one thin best-fit straight line with balanced scatter above/below, or a smooth best-fit curve. Do not join point-to-point and do not force through the origin unless evidence/theory supports it.

For a tangent, choose the stated curve point, place a ruler along the local direction, draw a long line touching there, then use widely separated points on the tangent for its gradient.

Use a justified reading uncertainty: commonly half the smallest division for a clear analogue scale, one least-significant displayed digit for digital readings, or a larger uncertainty when alignment/judgment limits dominate.

When uncertainty bars/lines are required, draw steepest and shallowest acceptable trends, calculate mmax and mmin, then estimate Δm≈(mmax−mmin)/2.

Compare an experimental and accepted value using percentage difference or overlapping uncertainty ranges before claiming agreement. A close-looking value without uncertainty is weak evidence.

Name a specific procedural limitation and its measured effect, e.g. 'end point is judged by eye, so the measured length varies between repeats'—not just 'human error'.

Pair each limitation with a feasible change that directly reduces it: video/frame timing for reaction time, fiducial marker/mirror scale for parallax, insulation/lid for heat loss, larger measured interval for fractional timing uncertainty.

An extension changes range or an independent factor to test whether the model remains valid; it is not a repair for a limitation unless it reduces the named uncertainty.

Check zero, read analogue scales at eye level, use the full safe range, close calipers/micrometers squarely and use the micrometer ratchet without overtightening. Repeat dimensions at several positions.

Select correct meter range/polarity; connect ammeter in series and voltmeter in parallel. Switch off before altering connections and limit current/heating in components.

Clamp stands with a stable base, keep falling masses clear of feet, prevent rolling objects leaving benches and keep eyes away from stretched springs or snapping wires.

Use heatproof mats/tongs for hot equipment, allow cooling before handling and keep liquids away from electrical supplies. State the hazard and the control together.