IB Internal Assessment: A Student Guide to Scoring Well

It is a Sunday night in Sushant Lok. Aarav has his first Physics IA draft due on Wednesday. His document has one line in it: "How does temperature affect resistance?" He has been sitting with it for two hours. His mother thinks he is working. He is mostly scrolling.

Some version of this evening happens in most IB homes in Gurugram, around October and again in February. The internal assessment is rarely hard because the physics is hard. It is hard because nobody explains how to start, and the deadline feels far away right up until it isn't.

Here is what the IA is worth, how to pick a question you can finish, and what costs marks.

What is an IB internal assessment?

An internal assessment is coursework you produce for each Diploma Programme subject. Your own teacher marks it against published criteria, and the IB moderates a sample of the school's marks. It usually carries 20 to 30 per cent of your subject grade, and you write it across the two years, not in an exam hall.

The shape changes by subject. In the sciences it is an investigation you design and run. In Maths it is an exploration. In the language groups it is a recorded oral. In Economics, short commentaries on real news articles.

The good part is that you control the conditions: no clock, no surprise question, and you can redraft. The cost is months of work alongside the Extended Essay. If your course load is still unsettled, read our guide to choosing HL and SL subjects first.

How much is the IA worth in each subject group?

Most IAs carry 20 to 30 per cent of the subject grade. Standard Level usually carries the higher share, because Higher Level students sit an extra exam paper that dilutes it. So the same piece of work can be worth 30 per cent at SL and 20 per cent at HL.

Treat that as a map, not a rulebook. The IB updates subject guides on a rolling basis, and limits move with them. Ask your teacher for the current guide, or check the official IB page for your syllabus year.

Choosing a research question that is narrow enough

A workable IA question names one thing you will change, one thing you will measure, and the exact setting. "How does temperature affect resistance?" is a topic. "How does temperature between 20 and 80 degrees Celsius affect the resistance of a 30 cm nichrome wire?" is a question you can finish.

Wide questions feel safe because they leave room. That room is what stops you analysing anything properly.

Wide questions and the fixed versions

• History. "What caused Partition?" becomes "To what extent did the Cabinet Mission Plan of 1946 make Partition unavoidable?"
• Biology. "How does light affect plants?" becomes "How does light intensity at five fixed distances affect oxygen release in Elodea?"

Test it: write your answer as one sentence with a number or a clear judgement in it. If you cannot imagine that sentence, the question is too wide.

There is a trade-off. Go too narrow and you have nothing to discuss, which starves the evaluation criterion. Aim for a predictable trend plus one thing that might misbehave.

Planning your IA against the marking criteria

Your teacher marks the IA with a published grid of criteria, each worth a set number of marks. Print that grid before you write a word. Build your section headings from the criteria names. Give each section the same share of your word count as its share of the marks.

The usual failure: eight pages of method and background, then an evaluation finished in four rushed lines the night before. If evaluation carries 6 of 24 marks, it deserves a quarter of your effort.

1. Print the criteria grid and keep it on your desk, not buried in a laptop folder.
2. Read the top band descriptor for each criterion and highlight the verbs it uses.
3. Turn each criterion into a heading in a blank document, with a target word count beside it.
4. Under each heading, bullet what must appear before it can reach the top band.
5. Only then start writing. You are filling a frame, not hunting for one.

Building a realistic drafting timeline

Work backwards from your school's internal deadline, not the IB's. Give yourself about ten weeks. The biggest single gain comes from finishing a complete, ugly first draft by week seven. Your supervisor's feedback then arrives while you still have time to act on it.

1. Week 1. Shortlist three topics, with two lines on why each interests you.
2. Week 2. Pick one, draft the question in full sentences, and read the criteria grid properly.
3. Week 3. Run a pilot or trial read. Most students change their method after it, which is the point.
4. Weeks 4 and 5. Collect data or gather sources, logging everything the same day.
5. Week 6. Process and graph your results. Do not write prose yet.
6. Week 7. Write the full first draft. Let it be bad. It only has to be complete.
7. Week 8. Hand it to your supervisor and move to other subjects while you wait.
8. Weeks 9 and 10. Rewrite, fix citations, check the word count, proofread.

The students who score well are not the cleverest ones. They are the ones who finished a bad first draft early enough to fix it.

All this sits on top of normal classes, so protect it inside a weekly plan. Our post on building a daily study routine that holds covers how to carve out that block. If the pile-up is costing sleep, the notes on managing exam stress help.

How to get useful feedback from your IA supervisor

In most schools you get a few short check-ins and written feedback on exactly one complete draft. That single round is precious. Arrive with specific questions and a real draft and you get specific answers. Arrive with "is this okay?" and you get "looks fine", which helps nobody.

Questions that get you something

• "Which criterion is my draft weakest on, and which band would you put it in right now?"
• "My evaluation says the thermometer was a limitation. Is that specific enough, or does it need a number?"
• "Is this question narrow enough for the top band, or should I fix one more variable?"
• "I am 400 words over the limit. Which section would you cut?"

Know the rule your supervisor works under. They may comment on one draft, but cannot edit it or mark corrections on the final version. That is an academic honesty requirement, and it applies to outside tutors too.

Academic honesty, citations and AI tools

Every idea, image, dataset, translation or phrase that is not yours needs a citation, in one consistent style. The IB treats unacknowledged copying, unacknowledged AI-generated text and shared work as malpractice. A confirmed case can cost you the subject grade or the whole Diploma, so do not cut this corner.

• Keep a running reference list from day one, in the same document; adding sources later is how students lose them.
• Save a screenshot or PDF of any web source you cite, with the date.
• If you used an AI tool to brainstorm or check grammar, declare it where your school asks. Never paste generated text into your draft.
• If your class shared raw data from a group experiment, state which numbers are yours.
• Apply the simplest test: you must be able to explain every line out loud, without notes.

Why do IAs lose marks?

Most lost marks come from structure, not from weak subject knowledge. Moderators see the same problems repeatedly: a question that is too broad, an evaluation full of vague limitations, and data reported without uncertainties. A discussion that never returns to the original question costs marks too.

• Vague limitations. "Human error" scores nothing. Name the error, say which way it pushed your result, and suggest a fix you could really do.
• Data without uncertainties. In the sciences, a table with no uncertainties and a graph with no error bars cap your analysis mark, however neat the work looks.
• Background that never connects. Two pages of textbook theory that never explain your prediction are wasted word count.
• Going over the limit. Assessors stop reading at the limit, and your conclusion is what falls off the edge.
• Engagement on paper only. "I have always loved chemistry" is not engagement. A reaction you noticed in your kitchen is.
• Late panic. A draft written in one weekend reads like one, and formatting does not hide it.

Reading your own work as a marker would, criteria in hand, fixes half this list. Our guide on stopping avoidable mistakes and the piece on revising so it sticks both apply.

What to do this week

Pick the IA with the nearest deadline and give it ninety minutes this week. Print the criteria, write your question as one full sentence, and email your supervisor to book a feedback slot. Three small actions in one sitting are usually enough to break the stuck feeling Aarav had.

1. List all six IAs with their school deadlines on one page, in date order.
2. Download the current subject guide for whichever IA starts first, and print the criteria pages.
3. Write three candidate questions and ask your teacher on Monday which one is narrow enough.
4. Block two fixed hours a week for IA work and treat them like a class.
5. Set your own draft deadline two weeks before the school's, and tell someone at home.

If the subject itself is the problem rather than the process, a specialist helps more than another late night. Dheeru Home Tuitions teaches IB students across Golf Course Road, the DLF Phases and Nirvana Country; see the subjects we cover and the learning models.