Mathematical modelling is one of the most important skills in IB Maths AI HL.

It isn’t just about finding an answer using a formula.

A modelling question usually gives you a real situation and asks you to turn that situation into mathematics.

You might need to work with population growth, finance, medicine, environmental data, motion, statistics, or another real-world context.

The challenge is deciding which mathematical model is appropriate, how to use technology effectively, and whether your final result makes sense in the original context.

This is why modelling questions can feel different from ordinary calculation questions.

You aren’t always given a clear instruction such as “differentiate this function” or “calculate the standard deviation.”

Instead, you need to decide what mathematics to use.

This guide explains how to approach IB Maths AI HL modelling questions systematically.

What Is Mathematical Modelling?

A mathematical model is a mathematical representation of a real situation.

The process can be simplified as:$$\text{Real situation}$$$$\downarrow$$$$\text{Identify important variables}$$$$\downarrow$$$$\text{Make assumptions}$$$$\downarrow$$$$\text{Choose a mathematical model}$$$$\downarrow$$$$\text{Use mathematics and technology}$$$$\downarrow$$$$\text{Interpret the result}$$

The final answer isn’t useful unless it makes sense in the original situation.

For example, if a model predicts that a population will reach 500,000 people, you need to understand what that number represents, what time period it applies to, and whether the model is reasonable for that situation.

1. Start by Understanding the Context

Before writing an equation, read the entire question.

Don’t immediately search your calculator for a regression model.

First identify:

• What is being measured?

• What are the variables?

• What are the units?

• What information is given?

• What are you being asked to predict or calculate?

• What time period does the model cover?

For example, suppose a question describes the number of users of an app over several years.

You might have:$$t=\text{time in years}$$

and:$$P=\text{number of users}$$

Your first task is understanding the relationship between $t$ and $P$.

Only then should you think about choosing a model.

2. Identify the Variables

A modelling problem usually contains at least two important variables.

One is often the independent variable.

The other is the dependent variable.

For example:$$t=\text{time}$$

and:$$P=\text{population}$$

Here, population depends on time.

You could represent this as:$$P=f(t)$$

This simple step helps organise the rest of the problem.

Always check the units

Suppose time is measured in months but your model uses years.

You need to know this before interpreting your answer.

A numerical answer without correct units can be misleading.

3. Look for the Pattern in the Data

If a modelling question gives you a table of data, don’t immediately choose a regression model.

Look at the data first.

Ask:

Is the relationship roughly linear?

Does it appear to curve upward or downward?

Does it increase rapidly and then level off?

Does it show repeated cycles?

Does it appear to grow exponentially?

The shape of the data gives you important information about the type of model you might need.

A scatter plot can be particularly useful here.

4. Understand Common Model Types

Different patterns suggest different models.

Linear Model

A linear model has the form:$$y=ax+b$$

The rate of change is approximately constant.

Quadratic Model

A quadratic model has the form:$$y=ax^2+bx+c$$

It can describe relationships with a curved shape.

Exponential Model

An exponential model can be represented by:$$y=ab^x$$

This is useful when a quantity changes by a roughly constant percentage or proportional rate.

Logarithmic Model

A logarithmic model has a form such as:$$y=a+b\ln x$$

It can be useful when growth increases rapidly at first and then slows.

Other Models

Depending on the problem, you may also encounter models involving periodic behaviour, logistic growth, or other functions.

The important skill is not memorising a list.

It’s recognising which model fits the situation.

5. Use the Graph Before Choosing the Model

A graph can tell you a lot before you calculate anything.

Suppose you plot your data and obtain a roughly straight pattern.

A linear model might be appropriate.

If the graph curves upward, an exponential or quadratic model might fit better.

If the graph rises quickly and then approaches a maximum value, a model with a limiting behaviour may be more appropriate.

But visual appearance isn’t enough

Two models can both look reasonable.

You need to compare how well they represent the data and whether they make sense in context.

This is where technology becomes useful.

6. Use Technology to Explore the Data

Technology is a major part of IB Maths AI.

Your calculator or other permitted technology can help you:

• Create scatter plots.

• Perform regressions.

• Calculate correlation statistics where appropriate.

• Find model parameters.

• Graph functions.

• Find intersections.

• Make predictions.

• Check calculations.

The important point is that technology should help you investigate and justify your mathematical decisions.

Don’t simply enter the data, accept the first regression result, and stop.

7. Don’t Choose a Model Only Because It Has a High $R^2$

The coefficient of determination can help you compare models, but it isn’t the only consideration.

Suppose:$$R^2=0.98$$

for one model and:$$R^2=0.96$$

for another.

That doesn’t automatically mean the first model is the better choice for every purpose.

You should also consider:

• The shape of the data.

• Residual behaviour.

• The context.

• The range of the data.

• Whether the model makes sense outside the observed values.

A mathematically good fit isn’t automatically a good real world model.

8. Understand Residuals

A residual measures the difference between an observed value and the value predicted by the model.

Conceptually:$$\text{Residual}=\text{Observed value}-\text{Predicted value}$$

Residual analysis can help you determine whether a model is appropriate.

If residuals show a clear pattern, your model may not be capturing an important feature of the data.

If residuals appear randomly scattered around zero, the model may be more appropriate.

Why this matters

A regression line can look good on the original graph while residuals reveal a systematic problem.

That’s why checking the model shouldn’t stop at finding an equation.

9. Interpolation and Extrapolation

This distinction is essential in modelling.

Interpolation

You predict a value within the range of your observed data.

For example, if your data covers:$$2015\leq t\leq2025$$

predicting something for 2020 is interpolation.

Extrapolation

You predict outside the observed range.

Predicting something for 2040 would be extrapolation.

Extrapolation is generally more uncertain because you’re assuming the existing pattern continues beyond the data you actually observed.

Exam habit

Whenever you’re asked to make a prediction, ask:

Am I predicting inside or outside the data range?

Then comment on the reliability of the prediction.

10. Make Reasonable Assumptions

Real situations are complicated.

A mathematical model simplifies reality.

For example, a population model might assume that certain conditions remain stable.

A financial model might assume a particular growth rate.

A model of temperature might ignore variables that are difficult to measure.

When a question asks you to discuss assumptions or limitations, don’t write something vague such as:

“The model isn’t perfect.”

Explain why.

For example:

The model assumes the growth pattern observed in the available data continues into the prediction period.

That’s a meaningful modelling limitation.

11. Understand What the Parameters Mean

Finding the equation isn’t the end of the problem.

You should understand what its parameters represent.

Suppose your model is:$$P(t)=500(1.08)^t$$

The value 500 represents the initial value when:$$t=0$$

The factor 1.08 represents an 8% growth factor per time period.

This interpretation can be more important than simply calculating a value.

Ask yourself:

What does each number in my model mean in the real situation?

That question should become automatic.

12. Check Whether Your Model Makes Sense

Suppose you create a model for the population of a town.

Your equation predicts:$$P(50)=2,500,000,000$$

Mathematically, your calculator may have no problem producing that result.

But is it realistic?

Probably not.

This is why modelling questions require interpretation.

A calculator can’t tell you whether an answer makes sense in the real world.

You need to evaluate it.

13. Watch for Domain Restrictions

A mathematical function might be defined for many values of $x$, but your real world model may not be.

Suppose your data represents the population of a town from 2010 to 2030.

Your model may technically produce values for:$$x<2010$$

but those predictions may have no practical meaning.

Similarly, a model for the number of customers cannot reasonably produce negative customers.

Your model has a domain of practical relevance determined by the context.

Always consider it.

14. Long Exam Questions: Break Them Into Stages

AI HL modelling questions can contain a lot of information.

Don’t try to solve the whole question at once.

Break it down.

Stage 1

Identify the variables.

Stage 2

Understand the data.

Stage 3

Create or select an appropriate model.

Stage 4

Use technology where appropriate.

Stage 5

Interpret the model.

Stage 6

Answer the specific question.

Stage 7

Check whether your answer is reasonable.

This prevents you from becoming overwhelmed by a long context.

15. Don’t Ignore Units

Units can help you understand what your model means.

Suppose:$$d=120$$

What does 120 mean?

120 kilometres?

120 metres?

120 people?

120 days?

The number alone isn’t enough.

Your final answer should use the appropriate unit whenever the question provides one.

Units can also help you catch mistakes.

If you’re calculating a rate and your answer has the wrong unit, revisit your reasoning.

16. Model Interpretation Is Part of the Mathematics

Suppose your model is:$$C(t)=250+40t$$

and $C$ represents cost in dollars while $t$ represents months.

The number 40 isn’t just an algebraic coefficient.

It represents a rate of change of:$$40\text{ dollars per month}$$

Likewise, 250 represents the modelled cost when:$$t=0$$

Being able to explain this is an important modelling skill.

17. Don’t Confuse Correlation With Causation

Suppose your data shows that two variables are strongly correlated.

That doesn’t automatically mean one causes the other.

For example, two quantities may both increase over time without one directly causing the other.

When interpreting statistical models, be careful with your language.

Prefer:

“The variables show a strong positive association.”

rather than automatically saying:

“One variable causes the other.”

This distinction becomes especially important when modelling real world data.

18. Use Your Mistakes to Improve Modelling

Modelling mistakes are often different from ordinary calculation mistakes.

You might get the correct calculator result but still lose marks because you:

• Chose an inappropriate model.

• Didn’t define your variables.

• Misinterpreted a parameter.

• Used extrapolation without discussing reliability.

• Ignored units.

• Failed to justify your model.

• Gave an unrealistic prediction.

• Didn’t explain the answer in context.

This means your mistake review should classify the error.

If you want to identify broader mathematical weaknesses before planning your next practice session, use the Math Skill Scanner.

19. Practice Modelling in Three Stages

Stage 1: Understand the Model

Start with questions where the model is already given.

For example:$$P(t)=1200(1.04)^t$$

Practise calculating and interpreting values.

Stage 2: Choose the Model

Move to questions where you’re given data and must decide which model is appropriate.

Stage 3: Build and Evaluate

Finally, work on longer problems where you need to:

• Identify variables.

• Select a model.

• Use technology.

• Interpret parameters.

• Make predictions.

• Discuss limitations.

This progression builds the actual modelling skill.

20. Use Feedback on the Whole Solution

A final answer doesn’t tell you everything about your modelling ability.

Suppose your predicted population is correct.

You may still have made a weak modelling decision.

For example, perhaps you selected the model without justification.

Or maybe you used extrapolation without discussing its limitations.

When you practise IB Maths questions, review the full chain of reasoning:

Context → Variables → Model → Calculation → Interpretation → Evaluation

This helps you improve more than simply checking whether the final number matches the solution.

21. A 90 Minute IB Maths AI HL Modelling Revision Session

15 minutes: Concept Review

Review:

• Model types

• Variables

• Regression

• Residuals

• Interpolation

• Extrapolation

20 minutes: Data Exploration

Take a small dataset and create a scatter plot.

Identify its general shape.

25 minutes: Modelling Practice

Choose an appropriate model and use technology to find its parameters.

20 minutes: Exam Question

Complete one longer modelling problem.

Focus on explaining your reasoning.

10 minutes: Evaluation

Write down:

What did the model assume?

Where is it reliable?

Where might it fail?

This final step is essential.

22. Common IB Maths AI HL Modelling Mistakes

Mistake 1: Choosing the First Model That Looks Reasonable

Compare alternatives and consider the context.

Mistake 2: Relying Only on $R^2$

Model fit isn’t the only consideration.

Mistake 3: Ignoring the Data Range

Extrapolation can be unreliable.

Mistake 4: Giving a Prediction Without Interpretation

Explain what the number means.

Mistake 5: Ignoring Units

Always include appropriate units.

Mistake 6: Treating the Model as Reality

A model is a simplification.

Mistake 7: Using Technology Without Understanding It

You should know what your calculator output represents.

Mistake 8: Failing to Evaluate the Model

Consider assumptions and limitations.

23. Build a Modelling Checklist

Before submitting a modelling question, ask:

Understanding

□ Did I identify the variables?

□ Did I identify the units?

□ Did I understand what I’m being asked?

Model Selection

□ Does my model match the shape of the data?

□ Did I consider whether another model could be more appropriate?

□ Does the model make sense in context?

Technology

□ Did I enter the data correctly?

□ Did I use the correct regression or calculation?

□ Did I check the calculator output?

Interpretation

□ Did I explain what the parameters mean?

□ Did I include units?

□ Did I interpret the final result?

Evaluation

□ Am I interpolating or extrapolating?

□ Is the prediction reasonable?

□ What assumptions does the model make?

□ What limitations should I mention?

Final Thoughts

The biggest mistake students make with IB Maths AI HL modelling is treating it like a calculator exercise.

It isn’t.

Technology can calculate a regression equation quickly, but you still need to decide whether that model is appropriate.

You need to understand the variables, recognise patterns in the data, select a suitable model, interpret its parameters, make predictions carefully, and evaluate the result in context.

When you practise modelling, don’t only ask:

“Did I get the right answer?”

Also ask:

“Did I choose the right model?”

“Can I explain what the model means?”

“Is my prediction reasonable?”

“What assumptions am I making?”

“Am I interpolating or extrapolating?”

These questions turn ordinary practice into useful exam preparation.

If you want to identify your mathematical weaknesses, start with the Math Skill Scanner.

When you’re ready to practise, use IB Maths practice questions and review your complete solution rather than only checking the final answer.

For a more structured approach to your IB Maths preparation, you can also explore Mathzem membership options.

FAQs About IB Maths AI HL Modelling

1. What is mathematical modelling in IB Maths AI HL?

Mathematical modelling involves representing a real world situation using mathematics. You may need to identify variables, choose a model, use technology, make predictions, and evaluate how well the model represents reality.

2. Which models should I know for IB Maths AI HL?

You should understand the main model types relevant to your course, including linear, quadratic, exponential, logarithmic, and other appropriate functions. More importantly, you should understand when each type of model is suitable.

3. How do I choose the best model for a dataset?

Start by examining the data and its graph. Consider the shape, how well different models fit, residual behaviour, the context, and whether the model makes sense for the intended prediction.

4. What is the difference between interpolation and extrapolation?

Interpolation predicts a value within the range of observed data. Extrapolation predicts outside that range and is generally more uncertain because it assumes the observed pattern continues beyond the available data.

5. How can I improve at IB Maths AI HL modelling questions?

Practise the complete modelling process rather than only calculating answers. Work on identifying variables, selecting models, using technology, interpreting results, checking predictions, and evaluating assumptions and limitations.