How to Write a Practical Report at O-Level: Structure, Templates, and Worked Example
TL;DR
O-Level practical reports in Singapore follow a specific structure that is different from university lab reports. The exam paper itself guides the structure (you fill in tables, draw graphs, and answer questions), but you still need to know the conventions for each section. This guide covers every section from raw data to evaluation, with templates and a worked example.
O-Level practical reports vs university lab reports
If you search "how to write a lab report" online, you will find university-level guides from Monash, Adelaide, and Toronto. These are not what you need. O-Level practical reports differ in several important ways:
| Feature | O-Level practical (Paper 3/5) | University lab report |
| Format | Fill in a structured answer booklet provided by SEAB | Write a full report from scratch |
| Introduction/hypothesis | Not required (the question provides the context) | Required (you write your own) |
| Method | Follow the instructions in the question paper | Design and write your own method |
| Data recording | Fill in tables with pre-printed headers or draw your own | Create tables from scratch |
| Graph | Draw on the grid provided in the booklet | Create graphs using software |
| Conclusion | Answer specific questions about the data | Write a full discussion section |
| Evaluation | Identify errors and suggest improvements (ACE) | Write a full error analysis |
| References | Not required | Required |
The key difference: at O-Level, the exam paper provides the structure. Your job is to fill it in correctly, following SEAB conventions.
The six sections of a practical report
Every O-Level practical report - whether in Physics, Chemistry, or Biology - follows the same six-section structure, which maps directly to the SEAB assessment strands.
Section 1: Raw data table (MMO marks)
This is where you record what the instruments show. The conventions:
- Column headers must include quantity and unit, separated by a forward slash: "Length / cm", "Time / s"
- Units go in the header only - do not repeat units in every cell
- All values in a column must have the same number of decimal places, matching the precision of the instrument
- Record exactly what the instrument shows - do not round or truncate
Example of a correct raw data table:
| Trial | Length / cm | Time for 20 oscillations / s |
| 1 | 30.0 | 22.06 |
| 2 | 30.0 | 21.94 |
| 3 | 40.0 | 25.38 |
| 4 | 40.0 | 25.42 |
Notice: Length is to 1 decimal place (ruler precision: 0.1 cm). Time is to 2 decimal places (stopwatch precision: 0.01 s). These are consistent within each column.
Section 2: Processed data table (PDO marks)
This is where you calculate derived quantities from your raw data. The conventions:
- Add new columns for processed quantities - do not overwrite raw data
- Show the formula you used (e.g., "Period T = time for 20 oscillations / 20")
- Match decimal places to the precision justified by your raw data - do not give more significant figures than your measurements support
- Include units in the processed column header
Example:
| Length / cm | Mean time for 20 oscillations / s | Period T / s | T² / s² |
| 30.0 | 22.00 | 1.10 | 1.21 |
| 40.0 | 25.40 | 1.27 | 1.61 |
Section 3: Graph (PDO marks)
Draw the graph on the grid provided in the answer booklet. Follow the conventions in our Graph Skills Guide:
- Label both axes with quantity and unit
- Choose a scale that fills more than half the grid
- Plot with small crosses
- Draw a best-fit line (not dot-to-dot)
- If calculating gradient, use a large triangle on the best-fit line
Section 4: Calculations (PDO/ACE marks)
Show all working clearly:
- Write the formula first
- Substitute values with units
- Calculate and state the answer with units
- Round to appropriate significant figures (2 - 3 s.f., consistent with your data)
Example:
Gradient = (T₂² − T₁²) / (L₂ − L₁)
= (1.61 − 1.21) / (40.0 − 30.0)
= 0.40 / 10.0
= 0.040 s² cm⁻¹
Section 5: Conclusion (ACE marks)
A conclusion must reference your data - it is not just "the hypothesis is correct." The template:
"The graph of [y-quantity] against [x-quantity] shows a [linear / proportional / non-linear] relationship. The gradient of the best-fit line is [value with unit], which [confirms / is consistent with / does not support] the expected relationship because [reason linking back to the physics/chemistry/biology]."
Example:
"The graph of T² against L shows a linear relationship passing close to the origin. The gradient is 0.040 s² cm⁻¹. This is consistent with the relationship T² = (4π²/g)L, confirming that the period squared is proportional to the length of the pendulum."
Section 6: Evaluation (ACE marks)
The evaluation has two parts: sources of error and improvements. Each improvement must match its corresponding error.
Template:
| Source of error | Effect on results | Improvement |
| Reaction time when starting/stopping the stopwatch manually | Period measurements may be too high or too low by up to 0.3 s | Use a light gate and datalogger to measure the period electronically |
| The pendulum did not swing in a single plane | Effective length varied during oscillations, making period inconsistent | Use a fiducial marker and ensure the bob swings in a single plane |
| Heat loss from the beaker surface (for thermal experiments) | Final temperature is lower than expected, reducing calculated energy | Insulate the beaker with cotton wool or use a polystyrene cup |
Phrases that earn marks:
- "Parallax error when reading the scale"
- "Repeat the experiment and calculate the mean to improve reliability"
- "Use a light gate and datalogger to reduce the effect of reaction time"
- "Allow the system to reach thermal equilibrium before taking the temperature reading"
Phrases that earn zero marks:
- "Human error" - always name the specific source
- "The experiment was not accurate" - too vague
- "Be more careful next time" - not a measurable improvement
For a complete list of examiner-approved phrases, see the Paper 3 Marking Guide.
How MMO, PDO, and ACE map onto report sections
| Report section | SEAB strand | What the examiner checks |
| Setting up and taking readings | MMO | Did you handle apparatus correctly and read instruments to the right precision? |
| Raw data table | MMO + PDO | Are units in headers, decimal places consistent, repeat readings taken? |
| Processed data table | PDO | Are calculations correct, headers labelled, significant figures appropriate? |
| Graph | PDO | Scale, labels, plotting accuracy, best-fit line, gradient calculation |
| Conclusion | ACE | Does the conclusion reference specific data values and link to the relationship? |
| Evaluation | ACE | Are sources of error specific and are improvements matched to each error? |
Worked example: simple pendulum experiment
Aim: Investigate how the period of a simple pendulum depends on its length.
Raw data:
| Length L / cm | Time for 20 oscillations (Trial 1) / s | Time for 20 oscillations (Trial 2) / s | Mean time for 20 oscillations / s |
| 20.0 | 18.12 | 18.06 | 18.09 |
| 30.0 | 22.06 | 21.94 | 22.00 |
| 40.0 | 25.38 | 25.42 | 25.40 |
| 50.0 | 28.50 | 28.44 | 28.47 |
| 60.0 | 31.12 | 31.18 | 31.15 |
Processed data:
| Length L / cm | Period T / s | T² / s² |
| 20.0 | 0.90 | 0.82 |
| 30.0 | 1.10 | 1.21 |
| 40.0 | 1.27 | 1.61 |
| 50.0 | 1.42 | 2.02 |
| 60.0 | 1.56 | 2.43 |
Graph: T² (y-axis) vs L (x-axis). Best-fit straight line through the data.
Gradient calculation: Using points (20.0, 0.82) and (60.0, 2.43) on the best-fit line:
Gradient = (2.43 − 0.82) / (60.0 − 20.0) = 1.61 / 40.0 = 0.040 s² cm⁻¹
Conclusion: The graph of T² against L is a straight line through the origin, confirming that T² is proportional to L. The gradient (0.040 s² cm⁻¹) is consistent with the theoretical value of 4π²/g.
Evaluation: Reaction time affects stopwatch readings - improvement: time 20 oscillations (not 1) to reduce percentage error. Pendulum may not swing in a single plane - improvement: use a fiducial marker and a narrow gap to constrain the swing.
Five mistakes that cost the most marks
- Missing units in table headers. Every column header needs "Quantity / unit". This is worth at least 1 PDO mark.
- Inconsistent decimal places. If one reading is 22.06 s, all readings in that column must be to 2 decimal places. Mixing 22.1 and 22.06 costs a mark.
- Conclusion without data. "The hypothesis is correct" earns nothing. You must quote specific values (gradient, intercept, calculated quantity).
- "Human error" in evaluation. This phrase earns zero marks every time. Name the specific physical cause.
- Gradient from data points. Always calculate the gradient using two points on the best-fit line, not from your plotted data points.

