The scientific investigation

📚What is not in the word count

Six things are excluded, and knowing this changes how you write: charts and diagrams, data tables, equations, formulas and calculations, citations and references of any style, the bibliography, and headers.

The practical consequence is that a well-annotated table or a clearly labelled graph costs you nothing, while a paragraph describing what the table shows costs you words. Put information into tables and figures wherever you legitimately can.

State at the start of the report: the title, your IB candidate code, the candidate codes of any group members, and the word count. No cover page and no contents page are required.

📈The four criteria, and where the marks really are

Research design (6 marks)

This assesses how well you communicate the purpose and practice of your methodology. For the top band (5–6) you need three things:

• the research question described within a specific and appropriate context — which means naming the independent and dependent variables, describing the system concisely, and giving background theory of direct relevance. “How does temperature affect rate?” is not a research question; “How does temperature (20–60 °C) affect the initial rate of the reaction between sodium thiosulfate and hydrochloric acid, measured by the time for a cross to disappear?” is;
• methodological considerations explained, not merely stated — why that measuring method, why that range and interval, how many repeats and why, which control variables and how each is controlled, and any safety, ethical or environmental issues;
• a description detailed enough that the investigation could be reproduced — specific apparatus, specific concentrations, precise steps — while avoiding repetitive detail.

The verb hierarchy across the bands is stated → described → explained. Almost every mid-band script fails on that word alone: it says what was controlled without saying why it mattered.

Data analysis (6 marks)

Three strands, and the top band needs all three:

• recording and processing communicated both clearly and precisely — clear means the method of processing can be followed; precise means conventions are right: units on every column heading, consistent decimal places, appropriate significant figures, properly annotated graphs;
• evidence of an appropriate consideration of uncertainties — record raw uncertainties from the instruments, propagate them correctly (add absolute uncertainties for addition and subtraction, add percentage uncertainties for multiplication and division, multiply by the power for exponents at HL), and show uncertainty bars on graphs;
• processing carried out appropriately and accurately, with no significant omissions.

Show one worked example of each calculation you performed and then tabulate the rest. That satisfies “clear” without consuming the word count, since calculations are excluded.

Conclusion (6 marks)

The top band requires a conclusion that is justified, relevant to the research question, fully consistent with the analysis — which means interpreting your processed data including its uncertainties — and justified through relevant comparison to the accepted scientific context.

That last phrase is where marks are routinely dropped. You must compare your result with a published value, an accepted theory or a literature figure, and cite the source in enough detail to be traceable. A conclusion that only describes your own graph, however carefully, cannot reach the top band.

Evaluation (6 marks)

The top band asks you to explain the relative impact of specific methodological weaknesses or limitations, and to explain realistic improvements that address them.

Two distinctions decide this criterion:

• Generic against specific. “Human error” and “the equipment was not accurate” are generic and score in the lowest band. “The 30-second sampling interval was too long relative to the 90-second reaction, so the initial rate was underestimated by roughly 10%” is specific.
• Weakness against limitation. A weakness concerns control of variables, precision of measurement or variation in the data. A limitation concerns how far the conclusion can be extended — the range of data collected, the confines of the system, the assumptions made.

Relative impact means ranking: say which weakness mattered most and why. If your percentage error is 15% and your total percentage uncertainty is 3%, then random error cannot explain the discrepancy and something systematic is at work — saying so, and naming the likely culprit, is exactly the reasoning the top band rewards.

🔧Choosing a research question

The research question need not go beyond the syllabus. What it must do is admit a clear independent variable, a measurable dependent variable, and enough data to analyse.

Approaches the guide allows, alone or combined: hands-on laboratory work, fieldwork, use of a spreadsheet for analysis and modelling, extraction and analysis of data from a database, and use of a simulation. A database investigation is a legitimate route, not a fallback — but the methodology must then focus on how the data were filtered and sampled, exactly as a practical methodology focuses on how measurements were made.

Chemistry topics that work well because they generate genuinely quantitative data with controllable variables:

⚖️Academic honesty

The report must be your own work. Your teacher may comment on one draft — giving oral or written advice on how it could be improved, but not editing it — and the next version you submit is final.

Reference everything: published values, methods adapted from a source, background theory, images. Omitted or improper referencing is treated as academic malpractice. Use one standard style consistently, and include the date of access for electronic sources.

Collaborative work is optional and limited to groups of no more than three. If you work in a group, each student must have their own research question, investigated by manipulating a different independent variable, or the same one with a different dependent variable, or different data selected from a common data set. No two students may present the same set of raw data, and all authoring, including the methodology, must be done individually. A group report is not permitted.

The same piece of work cannot be submitted for both the IA and the Extended Essay.

✅Before you submit — a checklist

• Title, candidate code, group members’ codes and word count at the start.
• Research question stated precisely, naming the independent and dependent variables and their ranges.
• Every control variable listed with the method of control and the reason it matters.
• Safety, ethical and environmental considerations addressed — briefly, but present.
• Raw data tables with units and uncertainties in the column headings, and consistent decimal places.
• Qualitative observations recorded, not just numbers.
• One worked example of every type of calculation, including the uncertainty propagation.
• Graphs with labelled and scaled axes, a line or curve of best fit, and uncertainty bars.
• Outliers identified and their inclusion or exclusion justified.
• Conclusion compared with a cited literature value, with the percentage error calculated and compared against the total percentage uncertainty.
• Evaluation naming specific weaknesses, ranked by impact, each with a realistic improvement that addresses it.
• Full bibliography, one consistent style, access dates on electronic sources.