Internal Assessment · Biology · 10 min read
What a full IB science IA looks like: an annotated example (Biology / ESS field study)
On this page
The most common thing my IB students say about the IA is not "it's hard". It's "I have no idea what the finished thing is supposed to look like". They've read the criteria, seen a few bullet points, and still can't picture the document.
So here is one, section by section. It's a field investigation of the kind that suits Biology or ESS: soil moisture and plant growth along a stream. Every section shows an excerpt of what the student would write, followed by why it works. The data, the graphs and the statistics are all worked through, so you can see how the numbers turn into a conclusion.
Read this first. The investigation, the data and the wording below are invented for this guide. They show the shape of a strong IA. Don't copy them. Your teacher has to authenticate your work, and an IA built on someone else's design or numbers is academic misconduct, even if you collected similar data yourself. Use this to understand what to aim for, then design your own.
The whole IA on one page
The current science IA (Biology, Chemistry and Physics, first assessed 2025) is a report of up to 3,000 words, marked on four criteria of 6 marks each. The IB states that half of the marks are for the conclusion and the evaluation. Most students spend 80% of their effort on the experiment and write the last two sections in a rush. That's backwards. The science IA criteria guide breaks down what each of the four criteria rewards.
One caution if you take ESS: this kind of field study suits the ESS IA too, but ESS marks it differently. The ESS "individual investigation" has six criteria worth 30 marks in total, including one on a strategy that tackles an environmental issue, so use the structure below for the data, conclusion and evaluation, and check the ESS criteria for the rest.
| Section | What it does | Criterion | Rough word budget |
|---|---|---|---|
| Title + research question | One precise question | Research design | — |
| Background and hypothesis | Why the question matters, what you expect and why | Research design | 400–500 |
| Variables and method | Exactly what you did, so someone could repeat it | Research design | 400–500 |
| Raw and processed data | Tables, with uncertainties | Data analysis | (tables don't count) |
| Graphs and statistics | Patterns, tests, what the numbers say | Data analysis | 300–500 |
| Conclusion | Answer the question, with evidence and science | Conclusion | 400–600 |
| Evaluation | Strengths, weaknesses, their effect, improvements | Evaluation | 600–800 |
| References | Sources you used | — | (not counted) |
The budgets are my suggestion, not an IB rule. Tables, graphs, equations, headers and the bibliography typically don't count towards the limit, but check the rules in your own subject guide.
1. Title and research question
How does distance from a stream bank (0–50 m, in 10 m intervals) affect soil moisture content (%) and the mean height (cm) of stinging nettle (Urtica dioica) along three parallel transects at [named nature reserve]?
Why this works: one independent variable with a range and intervals, two dependent variables with units, a named species (binomial name), and a specific site. Anyone reading it knows exactly what was measured. See how to write a research question for the before-and-after version.
2. Background and hypothesis
Stinging nettle grows in moist, nutrient-rich soil and is common along watercourses. Soil water affects growth directly: water is needed for turgor and transport, and too little water closes stomata and limits photosynthesis. Too much water is also a problem. Waterlogged soil is low in oxygen, which limits aerobic respiration in the roots and therefore active uptake of mineral ions.
Moving away from the stream bank, the soil should become drier. I therefore expect soil moisture to decrease with distance. For nettle height, I don't expect a straight line: plants very close to the bank may be limited by waterlogging, and plants furthest away by drought. My hypothesis is that height will be greatest at intermediate soil moisture.
H₀: there is no correlation between soil moisture and nettle height. H₁: there is a correlation between soil moisture and nettle height.
Why this works: the background isn't a textbook summary. Every sentence leads towards the prediction. The student uses biology (respiration, ion uptake, stomata) to predict the shape of the result, a curve rather than a straight line. That reasoning is exactly what the conclusion will come back to later.
3. Variables
| Variable | What | How measured or controlled |
|---|---|---|
| Independent | Distance from stream bank | Tape measure from the water's edge, 0, 10, 20, 30, 40, 50 m (±0.1 m) |
| Dependent 1 | Soil moisture content | Mass method (below), % of wet mass |
| Dependent 2 | Nettle height | Ruler from soil to the highest leaf, 3 nearest plants per point (±1 cm) |
| Controlled | Species | Only Urtica dioica measured |
| Controlled | Sampling day | All data collected on one day, no rain in the previous 48 h |
| Controlled | Soil depth | Samples taken at 5–10 cm depth with the same corer |
| Monitored (not controlled) | Light, soil type, trampling | Recorded at each point, discussed in the evaluation |
Why this works: controls are specific and say how they're held constant. Variables that couldn't be controlled are named honestly as "monitored". That gives the evaluation something real to discuss later.
4. Method
- Three parallel transects were laid perpendicular to the stream, 15 m apart, using a compass bearing of 110° from three marked points on the bank.
- At each sampling point (0, 10, 20, 30, 40 and 50 m), a soil core was taken at 5–10 cm depth, sealed in a numbered bag (ID e.g. T2-30 = transect 2, 30 m) and kept cool.
- The three nettle plants nearest to the point (within 1 m) were measured from soil level to the highest leaf. If fewer than three plants were within 1 m, the nearest plants within 2 m were measured and this was recorded. (This happened at T1-50.)
- Each point was photographed with a ruler for scale.
- In the school lab, each sample was weighed (wet mass), dried at 105 °C until the mass stayed constant, and weighed again (dry mass).
- Moisture content (%) = (wet mass − dry mass) ÷ wet mass × 100.
Safety and ethics: the site was public land; nothing was uprooted; gloves were worn for handling nettles; I stayed out of the water.
Why this works: it's precise enough to repeat, it includes the sample IDs, and the missing-data rule was written before collection. When it was needed at one point, it's reported, not hidden. Note what isn't here: no padding about "first I went to the site". Method marks come from precision, not length. More on fieldwork technique in the ecology field IA guide.
5. Raw data
Table 1. Soil moisture and nettle heights at each sampling point (illustrative data).
| ID | Distance (m) ±0.1 | Soil moisture (%) | Heights of 3 plants (cm) ±1 | Mean height (cm) |
|---|---|---|---|---|
| T1-0 | 0 | 57.1 | 78, 77, 61 | 72.0 |
| T1-10 | 10 | 44.1 | 86, 96, 91 | 91.0 |
| T1-20 | 20 | 39.2 | 75, 106, 91 | 90.7 |
| T1-30 | 30 | 32.0 | 83, 81, 88 | 84.0 |
| T1-40 | 40 | 27.5 | 72, 72, 80 | 74.7 |
| T1-50 | 50 | 18.4 | 27, 44, 35 | 35.3 |
| T2-0 | 0 | 51.2 | 74, 77, 71 | 74.0 |
| … | … | … | … | … |
| T3-50 | 50 | 22.5 | 53, 67, 63 | 61.0 |
(18 points in total; the full table would be in the IA.)
Why this works: every row has an ID that matches the sample bag and the photo. Uncertainties are in the column headers. Raw values are kept, not just the means. Small details like these are what make a data table credible.
6. Processed data and graphs
Table 2. Mean soil moisture and nettle height at each distance (n = 3 transects; heights from 9 plants).
| Distance (m) | Mean moisture (%) | SD | Mean height (cm) | SD |
|---|---|---|---|---|
| 0 | 55.6 | 3.9 | 70.4 | 8.2 |
| 10 | 43.8 | 3.2 | 91.2 | 6.8 |
| 20 | 37.6 | 2.4 | 90.2 | 10.0 |
| 30 | 30.9 | 1.5 | 82.1 | 6.0 |
| 40 | 26.5 | 0.9 | 71.1 | 7.4 |
| 50 | 20.0 | 2.2 | 44.2 | 14.4 |
Soil moisture fell steadily with distance, from a mean of 55.6% at the bank to 20.0% at 50 m. Spearman's rank correlation between distance and moisture across all 18 points was strongly negative (rs = −0.99, p < 0.001), so H₀ for this relationship is rejected.
Height did not change in a straight line with moisture (linear R² = 0.22). A quadratic model fitted the data well: h = −0.102m² + 8.45m − 82.6 (R² = 0.96), with a maximum at m ≈ 41%. Because Spearman's test only detects relationships that keep going in one direction, it was applied to the points away from the bank (10–50 m, n = 15): rs = 0.94, p < 0.001.
Why this works: means with standard deviations, graphs with labelled axes, units and error bars, and a caption that says what the reader is looking at. Most importantly, the student picked the test to fit the data. They noticed the curve, saw that a single straight-line correlation would be misleading, and explained what they did instead. That judgement is what moves Data analysis into the top band. If you're unsure which test fits your own data, the guide to statistical tests for the IA walks through the choice.
7. Conclusion
Soil moisture decreased strongly with distance from the stream bank (rs = −0.99). Nettle height was greatest at intermediate moisture, around 41% (Figure 2), and lower both at the waterlogged bank (mean 70.4 cm at 55.6% moisture) and at the driest points (44.2 cm at 20.0%). This supports my hypothesis of an optimum rather than a linear relationship.
The decrease on the dry side fits reduced water availability: lower turgor, and stomatal closure limiting photosynthesis. The lower heights at the bank fit waterlogging: with little oxygen in the soil, root respiration and active ion uptake are limited. However, my data shows a correlation, not a mechanism. I did not measure soil oxygen, so the waterlogging explanation is a plausible interpretation, not a demonstrated cause.
The variation in height increased at 50 m (SD 14.4 cm, compared with 6.0–10.0 cm elsewhere), mainly because transect 3 was taller there (61.0 cm against 35–36 cm on the other two). Transect 3 ran closer to a hedge, which may have given more shade and moisture. This makes the conclusion at the dry end less certain.
Why this works: it answers the research question directly, with numbers. It explains the pattern with biology, and connects back to the hypothesis. Then it does what most students skip: it says what the data can't show (correlation, not causation; oxygen not measured) and points at the one weak spot in its own data. That honesty is worth marks.
8. Evaluation
| Weakness | Effect on the results | Realistic improvement |
|---|---|---|
| Only 3 transects, so n = 3 per distance | Means at each distance are uncertain (SD up to 14.4 cm at 50 m) | 5–6 transects, chosen randomly along the bank |
| Soil oxygen not measured | Waterlogging explanation at 0 m is not tested | Measure redox potential or record waterlogging (water in the core) at each point |
| Light not controlled; T3 near a hedge | Shade and moister soil may explain T3's taller plants at 50 m (22.5% moisture vs 18–19% on T1 and T2) | Measure light with a lux meter at each point, or choose transects with similar cover |
| One sampling day | Moisture reflects one moment, after 48 h without rain | Repeat after rain and in a dry spell to see if the optimum holds |
| "3 nearest plants" could favour bigger, more visible plants | Possible bias towards taller plants | Use a quadrat and measure every nettle inside it |
Strengths: a clear gradient across the full range, a written missing-data rule applied consistently, a photo record of every point, and an analysis chosen to fit the shape of the data rather than forced into a linear test.
Why this works: every weakness has three parts: what it was, how it affected these specific results, and a fix that's actually doable. Compare that with "the sample size was small and there may have been human error", which could be written about any IA. Specific beats generic every time.
9. References
Use whichever referencing style your school asks for, consistently. Cite any source you rely on for background, any secondary data, and any AI tool you used, following the IB's rules (see using AI honestly).
How long this takes, realistically
| Week | What happens |
|---|---|
| 1 | Topic, site visit with photos, draft research question, approval from teacher |
| 2 | Pilot: one transect, check the method works, fix problems |
| 3 | Full data collection (1–2 days), lab drying, data into one master spreadsheet |
| 4 | Processed data, graphs, statistics. Write the method now. |
| 5 | Conclusion and evaluation |
| 6 | Background and introduction last, then a full read-through and the word count |
Turn this into your own IA
Use the free IA Planner Notion template. It has the research question checklist, the variables table, the missing-data protocol, a test chooser and the evaluation table from this example. Then read: