IB Sciences

IB Sciences Data Analysis Guide | Graphs, Uncertainties & Conclusion Writing

From plotting best-fit lines to propagating uncertainties and writing evaluations, this guide unifies the data analysis skills that separate high scorers across all IB science subjects.

What Do IB Science Examiners Actually Look for in Data Analysis?

IB science examinations—whether Biology, Chemistry, or Physics—test data analysis skills with remarkable consistency across papers. The short answer: examiners want evidence that you understand what your data means, not just that you can plot points and draw a line. That understanding is demonstrated through precise graph construction, correct handling of uncertainties, a clear distinction between error types, and a logically structured conclusion.

This guide walks through each of those areas in the order that will serve you best when sitting Paper 2 or Paper 3 response questions, writing your Internal Assessment (IA), or reviewing for final exams.


How Do You Construct a Graph That Scores Full Marks?

A graph is not just a visual summary—it is an argument. Every element you include or omit signals to the examiner whether you understand the data.

The Non-Negotiables

Choosing Between a Straight Line and a Curve

This is a common source of lost marks. Ask yourself: does the underlying physics or chemistry predict a linear relationship? If yes, draw a straight line—even if one or two points deviate. If the theoretical relationship is non-linear (e.g., exponential decay, inverse proportion), draw a smooth curve. Never force a straight line through data that curves; doing so suggests you are fitting the theory rather than reading the data.

Linearisation

When your relationship is non-linear, examiners sometimes award additional credit for linearising the data—for example, plotting ln(rate) against 1/T for an Arrhenius-type relationship. This demonstrates deeper mathematical fluency. Check whether your subject guide or mark scheme expects this for a given question type.


What Is the Difference Between Absolute and Relative Uncertainty?

Understanding uncertainty is arguably the most tested—and most misunderstood—part of IB science data analysis.

TypeDefinitionWhen to Use
Absolute uncertaintyThe ± value in the same units as the measurement (e.g., ±0.05 cm)Reporting a single measurement; drawing error bars; addition and subtraction of quantities
Relative (fractional) uncertaintyAbsolute uncertainty ÷ measured value (expressed as a decimal or percentage)Multiplying or dividing quantities during error propagation
Percentage uncertaintyRelative uncertainty × 100%Comparing the significance of different error sources

How to Propagate Uncertainties

The rule is simple once you remember the logic:

  • Adding or subtracting quantitiesadd the absolute uncertainties
  • Multiplying or dividing quantitiesadd the relative (percentage) uncertainties

For example, if you calculate density from mass and volume, both of which carry uncertainties, you add the percentage uncertainty in mass to the percentage uncertainty in volume to get the percentage uncertainty in density. Then convert back to absolute by multiplying by the calculated value.

Significant Figures in Uncertainties

As a general principle, uncertainties are quoted to one or two significant figures, and the measured value is rounded to match the last decimal place of the uncertainty. For instance: 12.3 ± 0.2 g, not 12.347 ± 0.2 g. Confirm the exact convention expected with your teacher and the current subject guide.


What Is the Difference Between Random and Systematic Errors—and Why Does It Matter?

This distinction carries marks in almost every data analysis question and is frequently examined in the context of evaluation.

FeatureRandom ErrorSystematic Error
Effect on dataScatter above and below the true valueAll values shifted consistently in one direction
Visible on graphLarge scatter around best-fit lineParallel line offset from the origin, or a non-zero intercept when zero is expected
Reduced byRepeating measurements and averagingRecalibrating equipment; redesigning the procedure
ExampleInconsistent reaction time when using a stopwatchA balance that reads 0.3 g when nothing is on it (zero error)

Why Examiners Care

Marks for "evaluation" in extended response questions require you to identify which type of error is present and explain its likely source. Saying "I repeated measurements to reduce error" without specifying random error, and without explaining why repetition does not address systematic error, typically earns only partial credit.

For a deeper look at how these principles apply specifically to lab work in biology, see the IB Biology HL guide, and for chemistry-specific contexts such as titrations and spectroscopy, the IB Chemistry HL guide covers evaluation criteria in detail.


How Should You Structure a Conclusion in a Paper 2/3 Response?

A conclusion is not a description—it is an argument. Examiners following mark schemes reward a specific logic sequence.

The Four-Step Structure

Deviating from this sequence—or treating the conclusion as a simple restatement of the procedure—consistently costs marks. The most frequent gap is step 3: students identify errors but fail to explain whether those errors are large enough to actually matter.

Quantitative vs. Qualitative Conclusions

Where your data allows, put numbers in your conclusion. "The rate increased" is weaker than "the rate approximately doubled with each 10 °C rise in temperature, consistent with a Q₁₀ of roughly 2." Quantitative statements demonstrate that you are reading the graph, not just describing it.


How Do These Skills Connect to the Internal Assessment?

The IA is where data analysis skills are formally and holistically assessed against published criteria. Everything in this guide—graph quality, uncertainty propagation, error classification, conclusion structure—maps directly onto the IA's evaluation and conclusion criteria.

The IB IA writing guide goes into the full criterion-by-criterion breakdown. If you are also approaching the Extended Essay in a science subject, the same data analysis principles apply—see the IB Extended Essay guide for how expectations shift in that context.


A Quick Reference Checklist Before Submitting Any Data Analysis

TaskDone?
Axes labelled with quantity and unit
Scale fills the majority of the grid
Error bars shown where applicable
Best-fit line or curve justified by theory
Absolute uncertainties recorded for raw data
Relative uncertainties calculated for derived values
Random and systematic errors identified separately
Conclusion follows: trend → hypothesis → errors → improvements
Uncertainties used to assess the significance of discrepancies
Specific, costed improvements proposed

Data analysis is a learnable set of skills, not an innate talent. Once the logic of each component is clear, it becomes systematic rather than stressful. If you find you are losing marks in a particular area—uncertainty propagation, conclusion structure, or graph interpretation—working through it with someone who has navigated IB science examinations recently can make the pattern click faster than practising alone. That is exactly the kind of targeted support Quick IB's IB-experienced tutors are there for.

FAQ

What is the relationship between uncertainty and error bars?
Error bars are a graphical representation of uncertainty on each data point. They show the range within which the true value likely lies, and the best-fit line should pass through or near all error bars to indicate a consistent relationship.
How do I find the uncertainty in a graph's gradient?
Draw the steepest and shallowest best-fit lines that still pass through the error bars, calculate both gradients, and take half the difference as the uncertainty in the gradient. Confirm the expected method with your teacher or the current subject guide.
What is the difference between Conclusion and Evaluation?
A Conclusion states what the data shows and compares it with the original hypothesis. An Evaluation critically examines the experiment's limitations, identifies sources of error, and proposes specific improvements to methodology or equipment.
When should I use relative uncertainty instead of absolute uncertainty?
Use relative uncertainty when multiplying or dividing quantities—you add the relative uncertainties of each value. Use absolute uncertainty for addition and subtraction, and when reporting a final result with its associated range.
How can I reduce random error in my experiment?
Increase the number of repeated trials and calculate the mean. A larger spread of values with more repeats allows better identification of outliers, and averaging reduces the effect of random fluctuations on your final result.
#IB Sciences#data analysis#uncertainty#error analysis#graph skills#conclusion writing

← Back to all articles