Paper 3: Limitations

🛠️ 9702 Physics Paper 3: Q2 Evaluation Toolkit

🧠 Examiner’s Insight: The Anatomy of Question 2

Question 2 is the “Evaluation” question. You are given a flawed, low-precision experiment, asked to take only two sets of readings, and then calculate a constant $k$.

The final section (usually worth 8 marks: 4 for limitations, 4 for improvements) is where top students separate themselves. The golden rule is specificity. Generic answers like “human error,” “parallax error,” or “use a computer” will score zero. You must name the specific variable (e.g., $h$, $\theta$, $t$) and describe exactly why it was hard to measure and exactly how to fix it.

---

🏆 1. The “Free Mark” (Always use this)

There is one limitation/improvement pair that applies to 100% of Question 2 experiments. Write this down as your first point every single time.

Limitation (Source of Error)	Improvement
Two readings are not enough to draw a valid conclusion / prove the relationship.	Take more readings (for different values of $x$) and plot a graph to find $k$.

⚠️ DO NOT CREDIT (Banned): Never just write “take repeat readings” or “calculate an average”. The mark scheme strictly requires “more readings AND plot a graph”.

---

⏱️ 2. Motion & Timing (Oscillations, Falling, Rolling)

Very common in experiments involving pendulums, springs, or rolling balls.

Limitation (Source of Error)	Improvement
Difficult to judge the start/end of an oscillation (or highest/lowest point).	Use a fiducial marker (e.g., a pin or line) at the centre of the oscillation (equilibrium point).
$t$ is very short, leading to a large percentage uncertainty in $t$.	Time for more oscillations (e.g., 10 or 20) OR use a video camera with a timer/frame-by-frame playback (slow motion).
Ball/cylinder does not travel in a straight line / wanders off track.	Use a guide track, channel, or parallel rails.
Object moves too fast to start/stop the stopwatch accurately.	Use light gates connected to a data logger/electronic timer.
Difficult to release the object without applying a force (giving it initial velocity).	Use a mechanical release mechanism, e.g., an electromagnet (for steel balls) or a card gate.
Unwanted modes of oscillation (e.g., pendulum swings in a circle instead of a flat plane).	Release from a smaller angle OR use a dual-thread suspension (bifilar pendulum).

---

📏 3. Lengths, Heights & Distances

When measuring extensions, heights of drops, or dimensions of objects.

Limitation (Source of Error)	Improvement
Parallax error when measuring height $h$ because the ruler is not flush with the object.	Read at eye-level OR use a set square against the bench to ensure the ruler is perfectly vertical.
Measurement $x$ is very small, leading to a large percentage uncertainty.	Use a micrometer screw gauge, Vernier calipers, or a travelling microscope.
Difficult to measure to the centre of the mass/bob.	Measure to the top and bottom of the bob using calipers and calculate the average.
Ruler moves or is difficult to hold perfectly vertical/steady.	Clamp the ruler to a retort stand using a boss.
Difficult to measure the distance because an object (e.g., a clamp) is in the way.	Use calipers to measure the outside distance and subtract the thickness of the object.

---

📐 4. Angles & Alignment

When using protractors or trying to keep things horizontal/vertical.

Limitation (Source of Error)	Improvement
Difficult to measure angle $\theta$ because the protractor is handheld / shakes.	Clamp the protractor to a stand.
Difficult to judge if the wooden strip/rule is perfectly horizontal.	Use a spirit level OR measure the height from the bench to the strip at both ends to ensure they are equal.
Difficult to judge if the string/spring is perfectly vertical.	Use a plumb line aligned next to the setup.
The zero-line on the protractor does not perfectly align with the pivot point.	Drill a hole at the origin of the protractor and mount it directly on the pivot nail.

---

🧲 5. Mechanics, Materials & Magnetism

Dealing with friction, putty/plasticine, springs, and magnets.

Limitation (Source of Error)	Improvement
Friction at the pulley / pivot / hinge affects the measurement of force $F$.	Lubricate the pulley/pivot (e.g., oil) OR use a high-quality bearing.
Plasticine / putty / rubber band deforms or stretches over time.	Wait for the material to finish stretching before reading OR use a non-deformable material (if allowed by context).
Mass of the adhesive putty (Blu-Tack) is not taken into account.	Weigh the putty on a top-pan balance and add it to the total mass $M$.
Magnets repel/twist making it hard to measure distance $x$.	Place magnets in a plastic guide tube to restrict sideways movement.

---

💧 6. Fluids (Water/Oil) & Optics

When pouring water, looking through lenses, or dropping objects in liquids.

Limitation (Source of Error)	Improvement
Difficult to judge when the water level reaches the line (meniscus is hard to see).	Add dye/food colouring to the water.
Difficult to focus the image perfectly on the screen / image is blurry.	Conduct the experiment in a dark room OR use a translucent screen and view from behind.
Parallax error reading the syringe / measuring cylinder scale.	Read at eye level to the bottom of the meniscus.
Water spills or leaks when transferring/filling.	Use a burette, pipette, or valve/tap for precise fluid transfer.

---

⚡ 7. Electricity

Usually involving resistance wires, potential dividers, or heating.

Limitation (Source of Error)	Improvement
Current fluctuates / meter reading is unstable.	Clean crocodile clips with wire wool (to remove oxide layer) OR solder the connections.
Wire heats up, which changes its resistance.	Open the switch between readings to allow the wire to cool.
Difficult to place crocodile clip exactly on the measurement mark.	Use a wire with a knife-edge contact (jockey) for better precision.
Resistance of the connecting leads/crocodile clips is not accounted for.	Measure the resistance of the leads alone and subtract from the total reading.

---

🚨 Top Student Secrets: How to Never Lose Marks Here

1. Be specific to the letter: Never write “Difficult to measure the length”. Write “Difficult to measure $L$ because the string bends.” Use the variables given in the question.
2. “Human Error” is banned: Never use the phrase “human error.” If it’s a reaction time issue, state: “Human reaction time is a large fraction of the short time $t$, creating a large percentage uncertainty.”
3. Data Loggers require Sensors: “Use a computer” or “Use a data logger” scores zero. You must pair it with a sensor: “Use a light gate connected to a data logger” or “Use a motion sensor connected to a computer.”
4. Avoid “Parallax error” as a blanket term: Don’t just write “Parallax error.” Write: “Parallax error when reading the height $h$ on the ruler.”
5. Video Recording is the ultimate hack: “Use a video camera and playback frame-by-frame with a timer in the shot” is a universally accepted improvement for almost any experiment involving fast motion, dropping, rolling, or quick oscillations.
6. Double-check the “Not enough readings” rule: Don’t write “Take 3 readings and average”. The board actively rejects this for the 2-reading limitation. The magic phrase is “Take more readings and plot a graph.”

Here is an expertly curated, high-yield cheat sheet based on the recurring themes, frustrations, and praises from the Principal Examiner Reports for Cambridge 9702 Physics Paper 3.

It is formatted in Markdown with Obsidian callouts for easy copying into your vault.

---

👑 9702 Physics Paper 3: The Ultimate Examiner Cheat Sheet

Tags: Physics ALevel Paper3 ExamTechnique DataAnalysis

The Examiner's Mindset

Examiners are actively looking for consistency, precision matching the instrument, and evidence that you are actually doing the experiment rather than just writing down expected physics theory. They heavily penalize “forced” data and vague evaluations.

---

📊 Question 1: Table & Graph Mastery

1. The Results Table (Easy Marks, Easily Lost)

• The Full Range: Do not cluster your independent variable. If you have a 1-meter wire, do not take readings from 10cm to 30cm. Spread them out (e.g., 10cm to 90cm). Examiners specifically check for this.
• Column Headings: MUST have a quantity and a unit separated by a solidus / or brackets (). Example: $1/I{\text{ / A}}^{-1}$ or $1/I{\text{ (A}}^{-1})$. Never attach units to dimensionless quantities (e.g., $\sin \theta$, $n$, or ratios like $L_1/L_2$).
• Raw Data Precision (Decimal Places): Raw data must match the resolution of the instrument exactly.
  • Ruler: Nearest mm (e.g., $25.0\text{ cm}$, NOT $25\text{ cm}$).
  • Micrometer: Nearest $0.01\text{ mm}$ (e.g., $0.35\text{ mm}$).
  • Stopwatch: Nearest $0.1\text{ s}$ or $0.01\text{ s}$ (be consistent down the column).
• Calculated Data Precision (Significant Figures): This must be calculated row-by-row. The sf of the calculated value must equal the sf of the least precise raw measurement used in that specific row, or one more.

  • Fatal Mistake: Do not keep the number of decimal places constant down a calculated column if the significant figures of the raw data change.

2. The Graph

• The “Awkward Scale” Trap: Never use multiples of 3, 6, 7, etc. Never make the lowest and highest table values the exact ends of the axes and subdivide. Use 1, 2, or 5. If you use an awkward scale, examiners will dock the axis mark AND usually the gradient/intercept read-off marks because you will inevitably misread them.
• Grid Usage: Points must occupy at least half the graph grid in both the $x$ and $y$ directions.
• Plotting “Blobs”: Use a sharp pencil and small crosses (x). If your point is larger than half a small square (>1mm), it is classified as a “blob” and scores zero for plotting.
• The Line of Best Fit (LOBF):
  • Do not force it through the origin $(0,0)$.
  • Do not simply connect the first and last points.
  • You need a balanced distribution of points above and below the line. If all top points are above the line and all bottom points are below it, your line needs rotation.
  • Use a transparent 30cm ruler. Short rulers create “kinked” or double lines, which score zero.
  • If you have one anomaly, circle it and label it “anomalous”. Examiners will then ignore it when judging your LOBF. You can only do this for ONE point.

3. Gradient and Intercept

• Gradient Triangle: The hypotenuse of your triangle MUST be greater than half the length of your drawn LOBF. Draw the triangle clearly.
• Read-offs: You must read coordinates from your drawn line, NOT from your data table (unless the table point lies exactly on the line).

• The False Origin Trap: Examiners catch thousands of students here. You can only read the $y$-intercept directly from the $y$-axis if the $x$-axis actually starts at $0$. If the $x$-axis has a false origin (e.g., starts at $10$), you MUST calculate the intercept using $y=mx+c$ with a point from your line.

---

🔬 Question 2: Evaluation & Unstructured Experiment

1. Timing Oscillations

• Never time a single oscillation. Time $n$ oscillations (where $n\ge 5$) and divide by $n$.
• Always record the repeated measurements of $nT$ in your working to prove you did it.

2. Calculating Percentage Uncertainty

• The Absolute Uncertainty: This is almost NEVER just the resolution of the instrument for Q2. It must reflect the practical difficulty of the measurement.
  • Example: For a stopwatch ($0.01\text{ s}$ resolution), human reaction time means absolute uncertainty is $0.2\text{ s}-0.5\text{ s}$.
  • Example: Measuring a bouncing spring with a ruler ($1\text{ mm}$ resolution), absolute uncertainty is $2\text{ mm}-5\text{ mm}$.
• The Half-Range Method: The best way to find absolute uncertainty is to take 3 repeats and use $\frac{\text{Max}-\text{Min}}{2}$. Show this working clearly!

3. Justifying Significant Figures for $k$

The Perfect Answer Formula:

“The number of significant figures in $k$ is [X], because the raw measurements of [Name Variable 1] and [Name Variable 2] were given to [Y] and [Z] significant figures, and the sf of $k$ must match the least of these.” Never use the phrase “raw data” or “raw readings”. You must name the specific measured quantities.

4. The $k$-value Conclusion (Does it support the relationship?)

You must show a 3-step logical chain:

1. Calculate % difference: $\frac{\text{Difference between }{k}_1\text{ and }{k}_2}{\text{Average of }{k}_1\text{ and }{k}_2}\times 100$
2. State the criterion: Compare it explicitly to the % given in the question (e.g., 20%).
3. Conclude: “The percentage difference is X%, which is less than the criterion of 20%, therefore the results support the relationship.” (Or vice versa).

---

🛑 Limitations and Improvements (The “Hit List”)

Examiners are ruthless here. Vague statements score zero. You must follow the formulas below.

Identifying Limitations (Quantity + Reason)

Do not just say “Measurement was difficult.” You must state exactly what was measured and why.

• ❌ Bad: “Parallax error.”
• ✅ Good: “Parallax error when measuring $h$ because the ruler could not be placed flush against the water drop.”
• ❌ Bad: “Hard to time.”
• ✅ Good: “Difficult to measure $T$ accurately because the period is very short, leading to a high percentage uncertainty due to human reaction time.”
• ❌ Bad: “Only two readings.”
• ✅ Good: “Two sets of data are insufficient to draw a valid conclusion about the relationship.” (This is almost always a free mark).

Suggesting Improvements (Action + Detail)

Do not suggest things that should be “standard practice” (e.g., “repeat readings and find an average”, “look at eye level”, “turn off the fans”).

• ❌ Bad: “Take more readings.”
• ✅ Good: “Take more readings for different values of [Independent Variable] and plot a graph to test the relationship.”
• ❌ Bad: “Use a video.”
• ✅ Good: “Record the oscillations using a video camera with a timer in view, and play back frame-by-frame to accurately determine $t$.”
• ❌ Bad: “Clamp the ruler.”
• ✅ Good: “Use a retort stand and boss to clamp the ruler vertically, verified using a set square against the bench.”
• ❌ Bad: “Use a better instrument.”
• ✅ Good: “Use vernier calipers instead of a ruler to measure the diameter $d$ to reduce percentage uncertainty.”

9702 AS-Level Physics Practical (Paper 3) – Ultimate Examiner Insights

This comprehensive analysis synthesizes multiple years of Principal Examiner Reports for CIE A-Level Physics (9702/Paper 3). It details the common pitfalls, exact precision rules, and evaluation strategies required to secure maximum marks.

---

1. Data Collection & Table Design

Precision Rules for Raw Readings

A primary area where candidates lose marks is failing to record raw measurements to the correct precision of the instrument.

Instrument	Typical Precision (Nearest…)	Example of Correct Stating	Common Errors to Avoid
Metre Rule / 30 cm Ruler	$1\text{ mm}$ ($0.1\text{ cm}$ or $0.001\text{ m}$)	$15.0\text{ cm}$ or $0.150\text{ m}$	Writing "$15\text{ cm}$" or "$15.00\text{ cm}$"
Vernier Caliper	$0.1\text{ mm}$ ($0.01\text{ cm}$)	$2.34\text{ cm}$	Writing "$2.3\text{ cm}$"
Micrometer Screw Gauge	$0.01\text{ mm}$ ($0.001\text{ cm}$)	$0.33\text{ mm}$ or $0.033\text{ cm}$	Misreading the half-millimetre mark
Protractor	$1^{\circ }$ (Nearest degree)	$45^{\circ }$	Writing "$45.0^{\circ }$" or "$45.2^{\circ }$"
Analogue Thermometer	$1^{\circ }\text{C}$ or $0.5^{\circ }\text{C}$	$23.0^{\circ }\text{C}$ or $23.5^{\circ }\text{C}$	Stating to $0.1^{\circ }\text{C}$ (unless digital)
Stopwatch	$0.1\text{ s}$ or $0.01\text{ s}$	$1.43\text{ s}$	Writing to three decimal places

The Trailing Zero Trap

Do not add trailing zeros merely to make the number of significant figures match down a column of raw values.

• Incorrect: $15.0\text{ cm}$, $9.00\text{ cm}$, $5.00\text{ cm}$ (inconsistent precision).
• Correct: $15.0\text{ cm}$, $9.0\text{ cm}$, $5.0\text{ cm}$ (consistent decimal places to reflect the $1\text{ mm}$ resolution).

---

Significant Figures (SF) of Calculated Columns

The number of significant figures in a calculated quantity must relate to the raw measurements used in that calculation.

• Rule: The calculated value must be written to the same number of significant figures as, or one more than, the raw measurement with the least number of significant figures.
• Row-by-Row Evaluation: Examine each row individually. If raw values change from 2 SF to 3 SF down the column, the calculated column’s SF can change accordingly.

$t=12.3\text{ s}(3\text{ SF})⟹\frac{1}{t}=0.0813{\text{ s}}^{-1}\text{ or }0.08130{\text{ s}}^{-1}$ $t=9.8\text{ s}(2\text{ SF})⟹\frac{1}{t}=0.10{\text{ s}}^{-1}\text{ or }0.102{\text{ s}}^{-1}$

Exception for Pure Integers

If a calculated value depends only on an exact integer (e.g., number of resistors $n=2,7,8$), SF rules do not apply. Choose a sensible decimal precision that allows for accurate plotting on the graph grid (e.g., write $1/7$ as $0.143$, not $0.1$).

---

Table Formatting Checklist

• Column Headings: Must state the quantity and unit, separated clearly by a solidus / or brackets (). E.g., $1/R/{\Omega }^{-1}$ or $1/R({\Omega }^{-1})$.
• No Units in the Body: Never write units (like $\text{cm}$ or $\text{s}$) inside the table cells. Units belong strictly in the column headings.
• Range of Independent Variable: Use the widest possible range of the apparatus. If you are given $1\text{ m}$ of wire, do not limit readings between $30\text{ cm}$ and $60\text{ cm}$. Extend them from $<10\text{ cm}$ to $>90\text{ cm}$.

---

2. Graphing & Data Analysis

Scale Selection

• Sensible Scales Only: Use simple ratios: $1,2,$ or $5$ units per large square (e.g., $0.1,0.2,0.5,$ etc.).
• Avoid Awkward Scales: Never use scales based on multiples of $3,6,7,12,$ or $15$. These lead to plotting errors and will be penalized.
• Grid Coverage: Points must occupy at least half of the grid in both the $x$ and $y$ directions (at least 6 large squares vertically and 4 large squares horizontally).

Plotted Points

• Fine Pencils: Use a sharp pencil. Dots must have a diameter $\le$ half a small square. Large “blobs” ($>1\text{ mm}$ diameter) will not receive credit.
• Symbols: Use a small, neat cross ($\times$ or $+$) or a small dot inside a circle.

Line of Best Fit (LOBF)

• Equipment: Use a transparent, non-folding $30\text{ cm}$ ruler. Short $15\text{ cm}$ rulers lead to kinked or double lines.
• Balance: Ensure a balanced distribution of points on either side of the line along its entire length.
• Anomalous Points: If a point is clearly anomalous, check calculations and replot. If it remains an outlier, circle it clearly to indicate it is being ignored for the LOBF. You may only identify a maximum of one anomalous point.

      y |
        |         x (Anomalous - circled)
        |        /
        |       / x
        |      /
        |   x / 
        |    / x
        |  x/
        +------------------ x

---

Gradient & Intercept Calculation

Gradient ($\Delta y/\Delta x$)

• Triangle Size: The hypotenuse of the gradient triangle must be greater than $50\%$ of the length of the drawn line.
• Read-off Accuracy: Read coordinates to the nearest half a small square.
• Points Selection: Only use points that lie directly on the line of best fit. Do not use data points from the table unless they happen to lie exactly on the line.

$y$-Intercept ($c$)

• False Origins: Check the $x$-axis. If the scale does not start at $x=0$, you have a false origin.
• Direct Reading Error: Do not read the intercept directly from the $y$-axis line if there is a false origin.
• Calculation Method: Instead, substitute a coordinate $(x,y)$ from the line of best fit and your calculated gradient $m$ into $y=mx+c$:

$c=y-mx$

---

3. Dealing with Uncertainties

Estimating Absolute Uncertainty ($\Delta$)

Do not simply use the resolution of the instrument if the measurement itself introduces additional difficulties.

Actual Uncertainty = Instrument Resolution + Experimental/Human Factors

• Timing oscillations: A stopwatch measures to $0.01\text{ s}$, but human reaction time introduces an uncertainty of $0.2\text{ s}$ to $0.5\text{ s}$.
• Irregular dimensions: When measuring the diameter of a hand-rolled clay ball or a foam ring, the shape is not perfectly spherical/circular. A realistic absolute uncertainty is $2\text{ mm}$ to $5\text{ mm}$, even though the ruler resolves to $1\text{ mm}$.
• Repeated readings: If you repeat a measurement, the absolute uncertainty can be estimated as:

$\Delta =\frac{\text{Range}}{2}=\frac{x_{\text{max}}-x_{\text{min}}}{2}$

---

Percentage Uncertainty vs. Percentage Difference

Do Not Confuse These Two Terms

• Percentage Uncertainty: Calculated for a single measured quantity based on its estimated absolute uncertainty: $\%\text{ Uncertainty}=\frac{\Delta x}{x}\times 100\%$
• Percentage Difference: Calculated to compare two experimental values of a constant (e.g., $k_1$ and $k_2$) to see if they support a suggested relationship: $\%\text{ Difference}=\frac{|k_1-k_2|}{k_{\text{average}}}\times 100\%$

---

Evaluating the Relationship (Q2)

To state whether your results support the relationship:

1. Compare your calculated percentage difference between $k_1$ and $k_2$ to the percentage uncertainty limit specified in the question (often $10\%$, $15\%$, or $20\%$, or your calculated percentage uncertainty for a key variable).
2. Write a clear concluding statement using precise phrasing:

“Since the percentage difference between my $k$ values ($7.4\%$) is less than the specified criterion of $15\%$, the experimental results support the suggested relationship.”

---

4. Question 2: Limitations & Improvements

To secure high marks in this section, avoid vague, generic answers. You must identify a specific measurement, explain the exact difficulty, and propose a detailed, practical solution.

Non-Creditworthy (Vague) Phrases vs. Creditworthy (Specific) Phrases

❌ Non-Creditworthy (Too Vague)	Creditworthy (Detailed & Specific)
“Human error” or “reaction time."	"It was difficult to judge the exact start/end of an oscillation because the motion was too fast."
"Parallax error."	"There was a risk of parallax error when measuring the height $H$ because the ruler could not be placed directly next to the tube."
"Two readings are not enough."	"Two sets of data are insufficient to draw a valid conclusion."
"Take more readings."	"Take more readings of [independent variable] and plot a graph of [Y] against [X] to test the relationship."
"Use a video camera."	"Record the motion using a video camera with a timer/stopwatch in the frame, and replay frame-by-frame to determine the period $T$."
"Use a robotic arm/mechanical hand.”	Not accepted. Solutions must be realistic classroom alternatives (e.g., clamping a guide).
”Make the rule vertical."	"Use a plumb-line / spirit level to align the rule vertically and clamp it in place.”