IB Internal Assessment Examples: 11 Worked IAs by Subject, With Marks and Criteria

Search for IB internal assessment examples and you get libraries. Thousands of student PDFs, sorted by subject, most of them behind a sign-up box. What you rarely get is the thing that actually helps: one example per subject, opened up, with the marks attached and a plain explanation of why it scored what it scored.

That is what this page is.

Quick answer: An IB internal assessment example is a complete piece of IA coursework you can read end to end — a science investigation, a maths exploration, a history investigation, an economics commentary, an oral recording. The useful ones come with a mark for each criterion. Every IB Diploma subject has an IA, and it is worth 20% to 30% of your final grade in that subject before you sit a single exam. Below you will find 11 worked examples, one per subject, each with its research question, its structure, its mark out of the real total, and the single change that would push it higher.

Two honest notes before you scroll. The only examples with authoritative IB marks live in the IB's Teacher Support Material, which your teacher can open, not on a public website, and we show you how to get those. And the examples below were written and marked by Educifly's IB subject specialists, who have coached IAs across maths, the sciences and humanities since 2018. They are modelled on the patterns we see in real student work. They are not IB-published exemplars, and we say so rather than pretending otherwise. For the wider picture, start with our guide to what an internal assessment is.

What counts as a real IB internal assessment example

A useful IA example does three things. Most free ones do one.

It shows the aim, not just the topic. "Enzymes" is a topic. "How does temperature affect the rate at which catalase breaks down hydrogen peroxide between 10°C and 60°C?" is an aim. Examiners mark the aim before anything else.

It shows the working, not the summary. You should be able to follow the data table, the calculation, the diagram.

It comes with marks per criterion. Without that you cannot tell a 21 out of 24 from a 14 out of 24. They look similar on the page. They are two grade bands apart.

Most public "IA example" pages fail the third test. They host the file and stop. That is why students read twenty examples and still do not know what a 7 looks like.

Where to find genuinely marked IB IA examples

There are four places people look. Only two give you marks you can trust.

1. The IB's Teacher Support Material. The IB publishes sample IAs with assessed marks and examiner commentary for every subject. Your teacher can open these on the programme resource centre. This is the only source where the marks are authoritative. Ask. Most teachers are happy to share and simply never get asked.

2. Your school's own archive. Many schools keep past IAs from earlier cohorts, with permission. These carry your own teacher's marks, which matters, because your teacher marks first and the IB moderates a sample afterwards.

3. Public sample libraries. Clastify, Nail IB, RevisionDojo and similar sites host large collections of student IAs. They are useful for seeing length and layout. Two cautions. The grades shown are usually self-reported, not confirmed by the IB. And reusing another student's work is malpractice.

4. Not the IB store. You can buy complete subject guides there. The assessed exemplars sit inside school-access material and are not sold separately.

Read examples for shape. Never read one on your own research question before you have written your own plan. It is very hard to unsee somebody else's structure.

IA length, marks and weighting: every major subject in one table

Before the examples, here is what you are actually working towards. All figures come from the current IB subject guides. The IB's own guide to assessment puts internal assessment at 20% to 30% of a subject in most courses, and as much as 50% in a few.

One pattern worth noticing: at SL the IA often counts for more, because SL students sit fewer exam papers. In Economics, Business management, English A and Psychology, an SL student has a quarter to a third of the grade decided before study leave starts.

IB internal assessment examples by subject

Each example below gives the research question, what the student actually did, the mark we would award, and the one edit that moves it up. Marks are ours, not the IB's.

Biology: catalase and temperature

The IA: the scientific investigation, 3,000 words, marked out of 24 across research design, data analysis, conclusion and evaluation — 6 marks each. Biology, chemistry and physics all share this model.

Research question: How does temperature (10, 20, 30, 40, 50 and 60°C) affect the initial rate of oxygen production when catalase from potato tissue breaks down 3% hydrogen peroxide?

What the student did: measured oxygen volume every 10 seconds for 2 minutes, five repeats per temperature, with pH, enzyme mass and substrate volume all controlled and stated. Calculated initial rate from the gradient of the first 30 seconds. Ran a t-test to compare the 40°C and 50°C means.

Our mark: 19 out of 24. Research design 6, data analysis 5, conclusion 5, evaluation 3.

Why it lost marks. The evaluation said "human error may have affected the results" and "a larger sample would help". Both are the exact sentences moderators see hundreds of times. The conclusion also stopped at "the optimum is around 40°C" without linking that back to protein structure.

The one change that moves it up. Name the limitation, estimate its size, and give the fix. Not "the water bath fluctuated" but "the water bath drifted by ±2°C, which at 50°C sits on the steep part of the denaturation curve and could account for the 8% spread in that group; a thermostatically controlled bath would cut that to ±0.2°C." That single rewrite is worth two marks. Our full IB Biology IA guide works through all four criteria with 50 more research questions.

Chemistry: how fast does aspirin break down?

The IA: scientific investigation, 3,000 words, 24 marks, same four criteria as Biology.

Research question: How does the pH of the solution (pH 2, 4, 6, 8 and 10) affect the rate constant for the hydrolysis of aspirin at 40°C, tracked by colorimetry as the iron(III)–salicylate complex?

What the student did: built a calibration curve first, then tracked absorbance every 2 minutes for 40 minutes at each pH. Plotted ln(absorbance) against time to get a first-order rate constant from each gradient. Propagated uncertainty from the pipette, the colorimeter and the thermostat through to the final rate constants.

Our mark: 21 out of 24. Research design 6, data analysis 6, conclusion 5, evaluation 4.

Why it lost marks. The chemistry is excellent and the uncertainty work is better than most. But the conclusion compared the rate constants to each other and never to a published value, so the reader cannot tell whether the numbers are sensible or a factor of ten out.

The one change that moves it up. Find one literature value, compare, and say what the gap means. "My rate constant at pH 10 is 30% below the published figure, which is consistent with my solution sitting below the stated 40°C for the first 5 minutes." That is a conclusion. Listing your own numbers again is a summary.

Physics: measuring g with a pendulum, properly

The IA: scientific investigation, 3,000 words, 24 marks.

Research question: How does the length of a simple pendulum (0.20 m to 1.00 m) affect its period, and what value of g does the gradient give?

What the student did: timed 20 oscillations, five repeats, at six lengths. Plotted T² against L, found the gradient, and derived g = 9.6 ± 0.5 m s⁻². Drew maximum and minimum gradient lines to get the uncertainty.

Our mark: 18 out of 24. Research design 5, data analysis 6, conclusion 5, evaluation 2.

Why it lost marks. This is the classic physics trap. The experiment is a textbook one, so research design caps out unless the student adds something of their own. And the evaluation treated the 9.6 result as "close enough" instead of doing the one piece of analysis that earns the marks.

The one change that moves it up. Use the uncertainty as a tool, not a footnote. A result of 9.6 ± 0.5 m s⁻² includes 9.81, so the honest sentence is "my value agrees with the accepted value within experimental uncertainty" — not "my value was slightly low due to air resistance". Then add one original variation: repeat at an amplitude of 30° and show where the small-angle approximation starts to fail. That lifts research design too.

Mathematics: the exploration (AA and AI)

The IA: the mathematical exploration, 12–20 pages, 20 marks across presentation (4), mathematical communication (4), personal engagement (3), reflection (3) and use of mathematics (6). Worth 20% in AA and AI, at both levels.

Research question (AI SL): Which of three mobile phone plans is cheapest for my family, given twelve months of our real usage data?

What the student did: pulled 12 months of call, text and data figures from three family bills. Built a piecewise cost function for each plan, graphed all three, found the break-even points algebraically, then ran a short simulation on next year's usage with a 5% growth assumption.

Our mark: 16 out of 20. Presentation 4, communication 3, personal engagement 3, reflection 2, use of mathematics 4.

Why it lost marks. Real data, real question, genuine ownership — the top half of the rubric is safe. Use of mathematics is the problem. Piecewise linear functions and a break-even point are not far enough beyond the start of the AI SL course.

The one change that moves it up. Add one layer of modelling the course expects. Fit a distribution to the monthly data usage, then calculate the probability that each plan is cheapest next year rather than asserting one growth rate. Criterion E is the only criterion where SL and HL are marked differently, and it carries 6 of the 20 marks. Our post on IB Math IA examples breaks down four full explorations, two AA and two AI, with marks for every criterion.

History: a 2,200-word investigation

The IA: historical investigation in three parts — identification and evaluation of sources (6 marks), investigation (15 marks), reflection (4 marks). 2,200 words, 25 marks, 25% at SL and 20% at HL.

Research question: To what extent was the 1953 coup in Iran driven by British oil interests rather than American fear of Soviet influence?

What the student did: evaluated a declassified CIA internal history and a 1954 British Foreign Office memorandum in section 1, then built a two-sided argument in section 2 using five further sources, and reflected in section 3 on how the decades-long classification of the CIA document shaped what historians could write before it surfaced in 2000.

Our mark: 20 out of 25. Sources 5, investigation 12, reflection 3.

Why it lost marks. Section 2 leaned on one school of interpretation and mentioned the counter-argument in a single sentence near the end. The reflection talked about historians in general rather than about this investigation.

The one change that moves it up. Give the counter-argument its own paragraph and then say which case is stronger and why. Examiners reward a judgement that is argued, not a balance that refuses to land. Our IB History IA guide covers source evaluation and the reflection in detail.

Economics: three 800-word commentaries

The IA: a portfolio of three commentaries on news articles, 800 words each, from three different syllabus sections and three different sources and countries. Marked out of 45 — 14 marks per commentary plus 3 for the portfolio as a whole. Worth 30% at SL, 20% at HL.

Commentary 2 (macroeconomics): an article on Japan's central bank raising interest rates for the first time in 17 years.

What the student did: drew an accurate AD/AS diagram showing the leftward shift in aggregate demand, explained the transmission mechanism from policy rate to investment, and evaluated the effect on savers, borrowers and exporters separately.

Our mark: 11 out of 14 for this commentary.

Why it lost marks. The diagram was correct and well labelled, but the text described it instead of arguing with it. The evaluation listed winners and losers without weighing them or putting a timescale on anything.

The one change that moves it up. Use the diagram as an instrument. Say what shifted, why, where the new equilibrium sits, and what follows. Then evaluate with magnitude and time: "the effect on exporters depends on how much the yen appreciates, and the pass-through to export volumes usually takes 6 to 12 months, so the article's claim of immediate damage is overstated." That is economic judgement, and it is the difference between a 3 and a 4 on the analysis criterion.

Business management: the research project

The IA: a business research project of up to 1,800 words on a real issue at a real organisation, built around one of four key concepts — change, creativity, ethics or sustainability. Marked out of 25. Same task and same criteria at SL and HL. Worth 30% at SL, 20% at HL.

Research question: Should Café Verde open a second branch in Bandra, or invest the same capital in delivery capacity at its existing site?

What the student did: interviewed the owner, surveyed 84 customers, and used a break-even analysis and an investment appraisal alongside a decision tree. Concluded in favour of delivery capacity.

Our mark: 19 out of 25.

Why it lost marks. The survey of 84 customers produced a lot of description. Several tools were included to show range rather than because the decision needed them.

The one change that moves it up. Cut one tool and make the survey earn its place. Ask two questions tied to the actual decision — how far customers travel, and what they would pay for delivery — then feed those answers straight into the break-even calculation. A tool that changes the answer is analysis. A tool that sits there is decoration.

Psychology: the experimental study

The IA: a report on a simple experiment that replicates a published study, 1,800–2,200 words, marked out of 22 across introduction (6), exploration (4), analysis (6) and evaluation (6). Worth 25% at SL, 20% at HL. Students plan and collect data in groups but write the report alone.

Research question: Does the wording of a question about a filmed car crash change the speed participants estimate, replicating Loftus and Palmer (1974)?

What the student did: independent measures design, 30 participants, two conditions ("smashed" and "contacted"), calculated means and standard deviations, ran a Mann-Whitney U test, and reported the result against a 0.05 significance level.

Our mark: 17 out of 22. Introduction 5, exploration 4, analysis 5, evaluation 3.

Why it lost marks. The discussion reported whether the hypothesis was supported and stopped. It never asked what a smaller effect than the original study might mean.

The one change that moves it up. Compare your result with the original study and explain the difference. "My mean difference was 4.2 mph against Loftus and Palmer's 9.0 mph, which may reflect a sample who watch far more video footage than 1974 participants did." That connects data to theory, which is exactly what the evaluation criterion asks for.

Geography: the fieldwork report

The IA: a written report of up to 2,500 words based on primary data you collect yourself, marked out of 25 across six criteria — fieldwork question and geographic context, methods, quality of information, written analysis, conclusion and evaluation. Worth 25% at SL, 20% at HL.

Fieldwork question: How does surface temperature change along a 3 km transect from the centre of Pune to the Aundh suburb, and does tree cover explain the pattern?

What the student did: took temperature readings at 15 points at the same hour on three days, estimated canopy cover from photographs at each point, and plotted temperature against canopy percentage with a correlation coefficient.

Our mark: 20 out of 25.

Why it lost marks. The analysis reported the correlation but did not deal with the three points that sat well off the line, which is where the geography actually was.

The one change that moves it up. Explain the outliers. Two of them sat next to a lake and one beside a metal-roofed market. Naming that turns a scatter plot into an argument about urban heat islands, which is what the criterion is asking for.

English A: the individual oral

The IA: a 15-minute oral — 10 minutes of prepared analysis, then 5 minutes of teacher questions — examining how a global issue is presented in one literary work and one non-literary body of work. Marked out of 40. Worth 30% at SL, 20% at HL.

Global issue: how algorithmic recommendation shapes what teenagers believe about their own bodies.

What the student did: paired an extract from a novel with three posts from a fitness influencer's feed, analysing narrative voice in the first and image cropping, caption register and comment framing in the second.

Our mark: 32 out of 40.

Why it lost marks. The non-literary analysis was sharp. The literary extract got less time and slipped into plot summary in the middle two minutes.

The one change that moves it up. Time the two halves in rehearsal and give the literary work its full share. Then land one sentence that holds both texts together. Examiners are listening for a global issue explored across two works, not two separate mini-analyses.

Language B: the individual oral

The IA: a spoken assessment worth 25% at both SL and HL, marked out of 30. At SL you describe and discuss a photograph linked to one of the five course themes; at HL you work from an extract of a literary work. You get 15 minutes of supervised preparation first.

Stimulus (Spanish B SL): a photograph of a family recycling at a community collection point, linked to the theme Sharing the planet.

What the student did: described the image for roughly 90 seconds, connected it to recycling schemes in Bogotá, then handled questions on household waste and local government.

Our mark: 22 out of 30.

Why it lost marks. Accurate Spanish, but the language stayed in the present tense and the safest vocabulary. The description also ran long, leaving the discussion rushed.

The one change that moves it up. Cut the description to about 60 seconds and spend the extra time on opinion, where you can reach for the subjunctive, conditional and past tenses. The mark scheme rewards range, and range is easiest to show when you are arguing rather than listing.

The same investigation, marked two ways

Topic choice matters less than students think. Execution matters more. Here is one biology investigation written two ways.

Same potato. Nine marks apart. In a subject where the IA is 20% of the grade, that is close to two grade boundaries.

What an IA mark is worth in grade terms

There is no fixed table that turns an IA mark into a subject grade. The IB sets grade boundaries fresh every session, judged against the grade descriptors and checked against statistics, so 19 out of 24 is not a permanent 6. Anyone who tells you otherwise is guessing.

What you can work out is how much of your final grade the IA has already decided. Two steps. Turn your mark into a percentage of the component, then multiply by the component's weighting.

Those are the marks from the worked examples above, run through the arithmetic.

Two things follow. First, an SL English student walking into exams with 24 of 100 already banked is in a different position from one carrying 15, and neither of them can change it in May. Second, the same gain is worth more in a heavier component — five extra marks on a 30% IA moves your subject total further than five marks on a 20% one. If you tutor one IA harder than the others, pick the one with the biggest weighting.

One part of the process students rarely hear about: your teacher's marks are not final. The IB re-marks a sample from every school and uses statistical comparison to decide whether the whole cohort's marks shift up or down. If a school's marks are cut by an average of 15% or more of the component maximum, the school can ask for re-moderation. So a generous teacher does not rescue a weak IA, and a harsh one does not sink a strong cohort.

What separates a top IA from a middling one, whatever the subject

Across subjects, the same four things decide it.

A question narrow enough to answer. If you cannot fully answer your research question inside the word limit, it is too wide. Add a range, a place, a time or a method until you can.

Analysis that interprets. "The value was 0.42" is a report. "The value of 0.42 is below the 0.5 threshold, which suggests the relationship is weaker than the model predicts" is analysis. For every number in your draft, ask: so what, for my question?

Engagement you can point to. Examiners do not award marks for saying a topic fascinates you. They award them for choices only you would have made — your data, your site, your method.

Reflection where it happens. Reflection saved for the conclusion reads like a checklist. Reflection placed at the moment a result surprised you reads like thinking. Move it.

The mistakes examiners see most

How to use an example without crossing the line

The line is closer than students assume.

Reading examples is allowed and encouraged. Copying structure sentence by sentence is not. Submitting anything generated for you is malpractice, and both you and your teacher sign a declaration that the work is your own. Schools run submissions through similarity software, and IB investigations can end with no grade awarded in that subject.

A safe way to use examples: read two or three in your subject before you choose a question, note only the section headings and roughly how long each section runs, then close the files and write your own plan. Come back to them after your first draft to check formatting, not content.

Your teacher can read and comment on one draft. Hand in a complete draft, not a sketch, because you only get one.

When to start: a realistic IA timeline

Most students start their IAs about two months later than they should. A workable shape for a two-year diploma:

1. End of year 1 — choose the subject area and read two examples. Nothing more.

2. Summer between years — settle the research question and check it with your teacher.

3. September of year 2 — collect data or read sources. This always takes twice as long as planned.

4. October to November — write the full first draft and submit it for feedback.

5. December to January — act on the feedback and finish. Most school deadlines land here, before the mocks.

Sciences, geography and psychology have a constraint the others do not: if your data collection fails, you need time to run it again. Build in four spare weeks.

How Educifly helps with IAs

Educifly is a boutique 1-on-1 tutoring practice. Students are hand-matched to a specialist who teaches one or two subjects deeply, and they keep the same tutor every week.

IA, EE and TOK coaching is a paid add-on rather than part of regular subject tutoring, so you can book it as a short block around your deadline instead of a full term. In an IA coaching session we work on the research question first, then the analysis and evaluation sections, because that is where nearly all the recoverable marks sit. We give feedback and teach the skill. We do not write coursework, and no legitimate tutor will.

To see how that works, book a free trial class and bring whatever draft you have.

Frequently asked questions

What is an IB internal assessment example?

An IB internal assessment example is a complete piece of student IA coursework you can read from start to finish — a science investigation, a maths exploration, a history investigation, an economics commentary or a recorded oral. The useful ones include the mark awarded for each assessment criterion, so you can see which choices earned which marks rather than guessing.

Where can I find real IB IA examples?

The only examples with authoritative IB marks are in the Teacher Support Material your teacher can open on the programme resource centre. Your school may also keep an archive of past student work. Public sites such as Clastify, Nail IB and RevisionDojo host large libraries, which are useful for seeing layout and length, but their grades are usually self-reported and not confirmed by the IB.

How much is the IB internal assessment worth?

It depends on the subject and the level. The IA is 20% in the sciences and mathematics at both levels, 25% at SL History, Geography and Psychology, and 30% at SL in Economics, Business management and English A. At HL it is usually 20%. Language B is the exception at 25% for both SL and HL.

Are IB IA examples the same for SL and HL?

The task is usually identical. What changes is the expectation. In mathematics, criterion E is the only criterion worded differently for SL and HL, and work that earns 6 marks at SL can earn 4 at HL. In the sciences, the criteria are the same at both levels but examiners expect more analytical depth from HL students.

Can I use an IB IA example as a template?

You can use it to understand structure and length. You cannot reuse its wording, its research question or its analysis. Both you and your teacher sign a declaration that the work is your own, schools check submissions with similarity software, and proven malpractice can mean no grade in that subject.

How long should an IB IA be?

Word limits vary by subject: 3,000 words for the sciences, 2,500 for Geography, 2,200 for History, 1,800–2,200 for Psychology, 1,800 for Business management, and 800 words per commentary in Economics. The mathematics exploration is guided at 12–20 pages rather than a word count. Orals are timed instead — 15 minutes for English A.

What is the most common reason IAs lose marks?

Describing results instead of interpreting them. Students report their data accurately and never explain what it means for their research question. The second most common is generic evaluation — naming a limitation without saying what effect it had or how to fix it. Both are fixable in an afternoon on a finished draft.

Who marks the IB internal assessment?

Your own teacher marks it first, using the IB criteria. The IB then asks the school to send a sample of marked IAs to an external moderator, who checks whether the school applied the criteria the way the IB intends. If the moderator disagrees, every mark in that school's cohort can be adjusted up or down, not just the sampled ones.

Can my teacher help me with my IA?

Yes, within limits. For most subjects your teacher can discuss your research question, answer questions as you work, and give written feedback on one complete draft. After that draft, they can point out where the work needs improving but cannot edit it for you. Hand in a finished draft rather than a partial one, because you only get the one round.

Do universities look at IB internal assessments?

No. Universities see your subject grades out of 7 and your total out of 45, not the IA marks that fed into them. That said, the IA is part of the grade they do see, and in an SL subject it can be 30% of it, so it shapes the number on your transcript even though it never appears there.

What if my IA data collection fails?

Tell your teacher straight away. If there is time, redesign and collect again — this is why science and geography students should leave four spare weeks. If there is no time, write up honestly what happened. An investigation with disappointing data and a sharp evaluation still scores; an investigation with invented data is malpractice.

Can I change my IA topic after starting?

Usually yes, if you move early. Before your first draft, a change costs you time and nothing else. After the draft has been marked and returned, most schools will resist, because you cannot get a second round of feedback. If the question is genuinely unanswerable, raise it with your teacher immediately rather than pushing on and hoping.