IA Data & Results: 10 Best Practices (IB Students)

The night before an IA deadline, data has a strange talent: it starts to look like a story you want to be true.

A slope that felt convincing at 6 p.m. suddenly looks shaky at midnight. A table that seemed “fine” now has missing units. And you catch yourself thinking, Maybe I can just smooth that out…

This is the moment where strong students separate themselves. Not by having perfect results, but by writing up data and results in an IA with calm precision: clear tables, honest processing, and analysis that respects uncertainty.

This guide walks you through 10 best practices to write up data and results in your IB IA, with the primary focus on making your IA readable, defensible, and easy to reward.

IA data and results checklist (fast scan)

Before you polish anything, run this checklist on your IA:

• Every table has a specific title and consistent formatting.

• Units and uncertainties live in headers, not scattered through cells.

• Graph types match the data (relationship vs comparison).

• Figures and tables are numbered and referenced in the text.

• Raw data, processed data, and qualitative observations are separated clearly.

• You explain how you processed data, not just the final numbers.

• You show variability (SD, error bars, ranges, confidence where appropriate).

• You state patterns and address anomalies.

• You compare to literature or accepted models where relevant.

• You evaluate limitations and propose improvements that actually change outcomes.

If you want criterion-level direction for your draft, start with RevisionDojo’s IA guides and browse more support in the IA blog tag archive.

Best practice 1: Build tables that tell the truth quickly

In an IA, tables are not decoration. They are your evidence ledger.

• Use descriptive titles that include variables and context (not “Table 1”).

• Put units and uncertainties in column headings.

• Keep decimal places consistent with the measurement tool.

A simple trick: if someone looked only at your table, could they still explain what was measured and in what units? If not, the IA loses clarity.

Best practice 2: Choose graph types that match the question

A surprising number of IA marks leak away from “almost right” visuals.

• Use scatter plots for relationships between continuous variables.

• Use bar/column charts for comparing categories or mean values.

• Use line graphs for trends over time when points are sequential.

Your IA graph should make a pattern obvious in three seconds. If the reader needs a minute, simplify.

Best practice 3: Label and reference visuals like a professional report

Your examiner should never have to hunt.

• Number visuals sequentially: Table 1, Table 2, Figure 1, Figure 2.

• Give captions that state the takeaway, not just what it is.

• Refer to each one in the text: “As shown in Figure 2…”

This is also where students quietly gain credibility: consistent structure signals control.

Best practice 4: Separate raw data, processed data, and observations

A high-scoring IA is easy to audit.

• Raw data: direct readings, trials, measurements.

• Processed data: means, rates, gradients, percent change, derived constants.

• Qualitative observations: color change, participant comments, unexpected environmental factors.

When these are mixed together, your IA becomes harder to trust, even if the numbers are correct.

Best practice 5: Show your processing steps (not just outcomes)

Examiners reward transparency.

In your IA, explain:

• the formula used,

• an example calculation,

• how you handled repeats,

• any transformations (log, reciprocal, normalization).

If you’re unsure whether your analysis is “real analysis” or just description, use What real analysis looks like in an IB IA as a calibration point.

Best practice 6: Use uncertainty and variability to strengthen your claim

Uncertainty is not a confession. It is scientific honesty.

• Add measurement uncertainties (instrument limits) where relevant.

• Include standard deviation or range for repeated trials.

• Use error bars when they help interpret reliability.

A good IA doesn’t pretend to be flawless. It explains how confident you can be and why.

Best practice 7: Write results as decisions, not a diary

Many students narrate results like a lab notebook: “The value increased, then decreased.”

Instead, make each paragraph answer one question:

• What pattern is most important?

• What does it suggest about the relationship?

• How does it connect to the research question?

If your IA includes an anomaly, name it directly and interpret it. For a strong framework, see IB IA unexpected results: strategies to handle them like a pro.

Best practice 8: Compare your IA findings to literature (carefully)

Comparison is where your IA moves from “my experiment” to “my understanding.”

• Compare trends to accepted theory, models, or published values.

• Discuss agreement and disagreement.

• Avoid forcing a match. Explain why differences are plausible.

This is also a great moment to use RevisionDojo’s Study Notes to refresh the relevant theory, then test your understanding with the AI Chat when you get stuck on interpretation.

Best practice 9: Evaluate limitations with specific, mark-winning improvements

Generic limitations (“human error,” “not enough time”) don’t help.

Better IA evaluation sounds like:

• Which variable wasn’t controlled?

• How exactly could it shift results?

• What change would reduce that effect next time?

Specific improvements are also easier to defend if your write-up is aligned with IB expectations from the start. Use IB Internal Assessment: a complete guide as a big-picture reference.

Best practice 10: Proofread like you’re the examiner

The last step of an IA is not “make it sound nicer.” It’s “remove friction.”

• Check every unit, axis label, and caption.

• Verify figure numbers match references in text.

• Scan for inconsistent decimal places.

• Make sure every table/graph earns its place by answering the research question.

If you want an external-style sanity check, RevisionDojo’s Grading tools are built for this: they break feedback down by criteria and show you what to fix next. (For example, Psychology students can use the IB Psychology IA grader.)

Closing: Make your IA easy to reward

A strong IA data and results section feels calm. Not because the results are perfect, but because the choices are disciplined: clear visuals, honest uncertainty, and analysis that stays tied to the research question.

If you want to tighten your IA faster, build a simple feedback loop with RevisionDojo: use Study Notes for theory, Flashcards for definitions and methods, AI Chat to unblock analysis, Grading tools for criterion-based improvements, and the Coursework Library to see what top-band work looks like. Then protect your exam performance with Questionbank drills and timed practice using Mock Exams and Predicted Papers.

For your next step, refine your foundation with How to write a strong IA research question, then come back and run these best practices on your IA data and results one final time.