Scientific Investigation: Internal Assessment Support
The IB SEHS internal assessment is an individual scientific investigation that you design, carry out and report on. This page takes you through every stage, from choosing a topic to checking your final report.
24%
Of final grade
~10 hrs
Practical time
3,200
Word limit (max)
1. What the internal assessment is
The internal assessment (IA), called the scientific investigation in the current course, is an open-ended task. You formulate your own research question, collect and analyse data to answer it, and write a report. It is marked against assessment criteria, with the strongest weight on conclusion and evaluation.
Feature
What to know
Weighting
24% of the final SEHS grade, at both SL and HL
Practical time
About 10 hours, including developing the question, the method, data collection and processing
Report length
Maximum 3,200 words. Check your subject brief for what counts towards the limit (for example whether references and appendices are included)
Individual report
Every student submits their own report, even if you worked in a group
Group work
Small groups may collaborate, and similar methods may be shared, provided the independent or dependent variable differs and each student's data is unique to them
Marking emphasis
Half of the marks are for conclusion and evaluation. Build your report around answering the question well and judging your own work
Check your official sourceThe IB subject brief for SEHS is the authority on mark allocations, criteria wording and word-count rules. This page is a practical guide that follows the published changes. Confirm the details with your teacher before you start.
The main stages of the investigation. Work through them in order, but expect to revisit earlier stages when results or evaluation show a problem.
2. Choosing a topic
The best topics test one manageable factor, use equipment you can access in school, and produce enough repeated, reliable data to support a justified conclusion. Use this checklist before you commit:
Safe: you can do it without unreasonable risk, and you have a risk assessment that your teacher has approved.
Measurable: the dependent variable can be measured with a reliable, precise instrument.
Controllable: you can keep the other factors constant.
Feasible: you can get enough participants or trials within about 10 hours.
Ethical: participants give informed consent, and no one is pushed beyond a safe level of exertion or discomfort.
Scientific: you can explain the expected result using SEHS theory from the syllabus.
Examples
Weaker topic
Why it is weaker
Stronger version
Does music improve fitness?
Too broad; "fitness" is not measurable
Does playing fast-tempo music change the number of push-ups completed in 60 seconds?
How does training change the heart?
Requires long training periods and invasive measures
Does a 2-minute warm-up change resting heart rate recovery after a set of step-ups?
Which shoe is best for running?
Uncontrolled variables and no clear measure
Does sole thickness change the contact time of a standing vertical jump?
3. Writing the research question
A research question should be focused, testable, measurable and clearly linked to the variables. A common format is:
TemplateHow does [independent variable] affect [dependent variable] in [participants or context]?
Make sure the question names the variables, states the units where relevant, and is possible to answer within your method. Avoid questions that can only be answered "yes" or "no", unless you then measure how much.
Good: How does a 1-minute versus a 3-minute rest interval affect the time taken for a 30 m sprint in Year 12 students?
Too broad: What affects sprint performance?
4. Background and hypothesis
Show that you understand the science behind your question. In the background section:
Explain the relevant theory from the syllabus, using the correct terms.
Summarise what other research has found, with references.
Build a hypothesis that predicts the direction of the effect and gives a reason, based on the theory.
Example hypothesis: Longer rest between sprints will allow more phosphocreatine resynthesis, so sprint times will be faster after 3 minutes of rest than after 1 minute, because the phosphagen system is more fully restored.
5. Variables and method
Identify your variables before you design the method. The independent variable is the one you change on purpose, the dependent variable is what you measure, and the controlled variables are kept the same.
An example of variables for a sprint investigation. The rest interval is changed, the sprint time is measured, and the other factors are kept constant.
Your method must be detailed enough for another student to repeat it. Include:
the participants (number, age, relevant characteristics) and how they were selected;
the equipment, with its precision (for example timing gates to 0.01 s, not a stopwatch to 0.1 s);
a step-by-step procedure, including warm-up, the number of trials, and the order of conditions;
how each controlled variable was kept constant;
the number of repeats, and how you dealt with fatigue and learning effects (for example randomising the order);
safety measures and the risk assessment reference;
how consent was obtained, and how personal data was stored and anonymised.
6. Recording data
Good data is easy to read, complete and traceable. Follow these rules:
Design your table before you collect data. Use one column per variable, with the quantity and unit in the heading, for example "Sprint time (s)".
Record raw data at the time, not afterwards. Write or type directly into the table. Keep the raw data unedited in an appendix.
Use consistent decimal places that match the precision of your instrument.
Record uncertainties for every measurement, for example ±0.01 s for a timing gate.
Repeat trials. Three or more repeats per condition is usually a good minimum, so that you can calculate a mean and spread.
Note anything unusual, such as a participant who stumbled or a timing error, in a comments column. Do not delete the result.
Keep a log of the date, time, conditions and equipment. A notebook or a shared digital log is fine; be consistent.
Back up your data on a second device or cloud folder as you go.
Example table layout
Condition
Trial
Sprint time (s ± 0.01)
Comments
1 min rest
1
4.82
–
1 min rest
2
4.91
–
3 min rest
1
4.67
Slight wind noted
7. Processing and analysing data
Processing turns raw numbers into something you can interpret. Use the methods that suit your data and your question.
Common calculations
Calculation
What it tells you
How to do it
Mean
The typical value across repeats
Sum of values ÷ number of values
Range and standard deviation
The spread of the data (the sample standard deviation is used for a sample)
Excel/Sheets: MAX − MIN for range; STDEV.S for standard deviation
Percentage change
How much a value changed, relative to the start
(new − original) ÷ original × 100
Absolute and percentage uncertainty
How precise a measurement or result is
Absolute: the ± value. Percentage: absolute ÷ value × 100
Correlation
The strength and direction of a relationship between two continuous variables
Excel/Sheets: CORREL(x, y). Report r with its sign and interpret its size
t-test
Whether two means differ by more than chance would suggest
Excel/Sheets: T.TEST. Choose paired or unpaired to match your design
Choosing a graph
Scatter plot with a line of best fit for two continuous variables (for example rest time against sprint time).
Bar chart with error bars for comparing means across categories (for example mean sprint time for each rest condition).
Line graph for change over time (for example heart rate over a 10-minute recovery).
Box plot to show the spread and median of each group.
Every graph needs a title, labelled axes with units, a sensible scale that uses most of the space, and a legend if there is more than one data set. Error bars should show the uncertainty or standard deviation, and the caption should say which.
Tools that can help
Tool
Use it for
Cost
Microsoft Excel or Google Sheets
Tables, means, standard deviation, t-tests, correlation, charts
Free (Sheets) / school licence
Desmos
Plotting, curve fitting and regression lines
Free
GraphPad QuickCalcs
Quick, checked t-tests and other statistics on the web
Free
Kinovea
Frame-by-frame video analysis, timing and joint angles
Free
Tracker
Video analysis of motion, with position and velocity data
Free
Sensors (for example Vernier or smartphone sensor apps)
Logging acceleration, force, heart rate or timing directly
Varies; check school equipment first
Check your workingRecalculate one value by hand, and check that your spreadsheet formulas refer to the right cells. Small errors in formulas are a common cause of wrong conclusions.
8. Writing the conclusion
The conclusion answers your research question directly, using your processed data. Make it:
Direct: start with the answer, for example "The 3-minute rest condition produced faster sprint times than the 1-minute condition."
Quantified: give the key values with their uncertainty, for example the mean difference and its spread.
Linked to the hypothesis: say whether the data supported your prediction, and why.
Linked to theory: explain the mechanism using SEHS concepts, without overclaiming.
Cautious about causation: a correlation or an effect in a small sample supports a conclusion; it does not prove it.
9. Writing the evaluation
Evaluation is where you judge your own investigation. Half of the marks sit in conclusion and evaluation, so give it real effort. A strong evaluation:
Identifies real limitations of the method, data or analysis, and explains how each one could affect the result.
Distinguishes random and systematic error. Random error varies unpredictably (for example reaction time when starting a stopwatch); systematic error pushes results consistently one way (for example a timing gate that is miscalibrated).
Discusses sample size and representativeness (for example a small group of Year 12 students cannot represent all athletes).
Judges the uncertainty of the result: is the difference between conditions larger than the uncertainty?
Suggests specific improvements, each linked to a limitation and explaining what it would change. "Use timing gates instead of a stopwatch, because hand timing adds human reaction error, often around 0.1–0.2 s, which is larger than some of the differences measured" is specific; "use better equipment" is not.
Weak: The results may not be accurate because of human error. Strong: Hand timing introduces reaction-time error of around 0.1–0.2 s, which is comparable to the 0.15 s difference between conditions. Timing gates would reduce this error and would make the difference between conditions more reliable.
10. Referencing
Support your background, method and evaluation with sources. Use one consistent referencing style throughout (for example Harvard or APA), with:
an in-text citation wherever you use another person's idea, data or image;
a reference list at the end, with full details for each source.
Good sources
Use textbooks, peer-reviewed journal articles, official bodies (for example government health guidance or national sports organisations), and the IB syllabus. Be cautious with blogs and general websites: check the author, date and evidence. Wikipedia can help you find sources, but do not cite it as a primary source.
Tools
Reference managers such as Zotero or Mendeley are free and can keep your references organised. Check whether your school allows a particular tool.
Harvard template (book): Surname, Initial. (Year) Title of book. Edition. Place of publication: Publisher. Harvard template (journal): Surname, Initial. (Year) 'Title of article', Journal Name, volume(issue), pp. x–x.
11. Structure and word count
A clear structure helps the examiner find the marks. A typical report includes these sections:
Title, and research question
Introduction and background (theory, prior research, hypothesis)
Method (variables, participants, equipment, procedure, safety and ethics)
Data (raw data tables, with uncertainties)
Data processing and analysis (calculations, graphs)
Conclusion
Evaluation (limitations and improvements)
References and appendices
Stay within the 3,200-word limit. A report that is too long usually has repetition, unnecessary background or raw data that belongs in an appendix. Check with your teacher what is counted.
12. Academic integrity
The report, data and analysis must be your own work. Do not make up or alter data.
Acknowledge any help, including the use of software, sources and any AI tools, in line with your school's policy and the IB's academic honesty rules.
If you collaborate, follow the group rules in the subject brief. Your data must be unique to you, and your report must be individual.
13. Before you submit
Does my research question name the variables and the context?
Can another student repeat my method from what I have written?
Are my tables complete, with units and uncertainties?
Are my graphs labelled, titled and appropriate for the data?
Does my conclusion answer the research question directly, with values?
Does my evaluation name specific limitations and specific improvements?
Is every source cited in the text and listed in the references?
Is the report within the word limit, and is the file named clearly?
14. Common mistakes
A research question that is too broad, or cannot be answered with the data collected.
Method steps that are vague ("measure the athletes") and cannot be repeated.
Raw data that is missing units or uncertainties.
Graphs without axis labels, units or a title.
A conclusion that claims more than the data shows, or restates results without answering the question.
An evaluation that only says "human error" or "small sample", without explaining the effect or proposing a specific improvement.
Simply Sport Science is an independent educational resource. It is not affiliated with, endorsed or sponsored by the International Baccalaureate Organization, Cambridge University Press & Assessment or Pearson/BTEC. All course and exam-board names are used only to identify the specifications the materials support.