The Excellence discussion: critical issues
What Excellence asks you to discuss
Excellence is awarded for a discussion that addresses critical issues in your investigation. You need the equivalent of two good discussion points — depth beats quantity, and a list of vague "sources of error" earns nothing.
The standard names five kinds of critical issue. Any two, done properly, will do.
1. Why there is a limit at either end of your range
- Explain what stopped you going lower, and what stopped you going higher, in terms of the reliability of the measurement — not simply "that was all the equipment we had".
- Too small: the change in the dependent variable becomes comparable to the measurement uncertainty, so the readings are dominated by error.
- "Below m the period was under s, so my s reaction-time error was more than 20% of the reading."
- Too large: the physics being assumed may break down, or the measurement becomes impractical.
- "Above m the pendulum could not be released without swinging sideways, so the motion was no longer in one plane."
- Too small: the change in the dependent variable becomes comparable to the measurement uncertainty, so the readings are dominated by error.
2. Justifying why a variable had to be controlled
- Pick a variable that would significantly affect the result, and explain the mechanism by which it would change the dependent variable.
- Weak: "I kept the mass the same to make it a fair test."
- Strong: "I kept the amplitude below 10°, because at larger angles the restoring force is no longer proportional to displacement, so the period increases with amplitude — a changing amplitude would then have changed my measured period even at constant length, hiding the relationship I was looking for."
3. Difficulties in measuring, and how you overcame them
- Name the specific difficulty, say why it made measurement hard, and state what you did about it.
- "Judging the exact moment the bob reached the top of its swing was difficult because it moves slowly there and appears to hang. I overcame this by placing a fiducial marker at the centre of the swing, where the bob moves fastest, and timing from there."
4. The relationship between your findings and physics ideas
- Connect your measured relationship to the physics that explains it.
- "My result matches the inverse-square law, which follows from the light spreading out over the surface of a sphere: the area of a sphere is , so the same power spread over four times the area at twice the distance gives a quarter of the illuminance."
- Where you can, compare a measured constant to an accepted value.
- "My gradient of s2 m−1 gives m s−2, close to the accepted m s−2."
5. Unexpected results, their cause, and their effect on the conclusion
- All three parts are needed: what was unexpected, what could have caused it, and what it does to the validity of your conclusion.
- "The point at m sat well above the line. The clamp stand was noticeably flexing at this length, which would let the pivot move and lengthen the effective pendulum, increasing the period. Since the remaining five points still lay close to a straight line, the relationship is still supported, but the evidence at the top of my range is weaker than at the bottom."
Worked ExampleTurning a weak error list into a discussion point
A student writes: "Sources of error: human reaction time, the ruler was not exact, and the pendulum sometimes swung in a circle." Rewrite the third of these as a proper Excellence discussion point.
Step 1 — Name the issue specifically
The pendulum sometimes swung in a conical path rather than in a single vertical plane, particularly at the longer lengths.
Step 2 — Explain why it affects the measurement
When the bob moves in a circle rather than back and forth, it does not pass the fiducial marker once per half-swing in the same way, so the count of oscillations is unreliable — and the motion is no longer the simple pendulum motion the investigation assumes.
Step 3 — State the effect on the results
This would make the measured period inconsistent at those lengths, adding scatter to the points at the top of the range. It could also make the measured period slightly longer, which would push those points above the line of best fit and reduce the measured gradient.
Step 4 — Say what was done, or should be done, about it
Releasing the bob from rest, without any sideways push, and using a card guide at the release point keeps the motion in one plane. Repeating any run in which a conical swing was noticed, rather than recording it, keeps the affected data out of the analysis.
Step 5 — State the effect on the validity of the conclusion
Because the affected runs were repeated and the final points still lie close to the line, the conclusion remains supported — but if a conical swing had gone unnoticed, the gradient would have been underestimated, and the relationship obtained would have been systematically wrong rather than merely scattered.