1.13 Function Model Selection and Assumption Articulation

Function model selection means picking the right type of function to represent a real-world situation, then stating the assumptions and restrictions that make your model make sense. Linear functions fit constant rates of change, quadratics fit symmetric data with one peak or valley, and higher-degree polynomials fit data with several zeros or turning points.

Why This Matters for the AP Precalculus Exam

This topic is where the work you have done on rates of change, polynomials, and rational functions becomes useful for actual modeling. Free-response questions in AP Precalculus ask you to do more than calculate. You have to explain and justify your choices, and modeling questions are a common place to do that.

When you choose a model, you are using number patterns and context clues to decide which function family fits. When you state assumptions and restrictions, you are showing that you understand a model is a simplified version of reality with limits. Both skills show up when you justify conclusions in words, which is a big part of clear free-response work. Being precise with units and variable meaning is important here too.

more resources to help you study

practice multiple choiceFRQ practice & scoringcheatsheetsscore calculator

Key Takeaways

• Match the function type to the pattern: constant rate of change means linear, constant second differences means quadratic, constant nth differences means degree n.
• Use context clues. Two-dimensional area problems often fit quadratics, and volume or three-dimensional problems often fit cubics.
• Polynomials are a good fit when data shows several real zeros or several maxima and minima.
• A polynomial of degree n or less can pass through any set of n + 1 points with distinct input values.
• Every model carries assumptions about what stays constant and how quantities change together.
• Set domain and range restrictions from the context, like time greater than or equal to zero or rounding to whole-number outputs.

Function Models Based on Differences and Degree

In AP Precalculus, a model is a mathematical representation of a real-world situation. It is a simplified version of the situation that you use to make predictions or understand how quantities relate. Models can be built from many function types, and the right choice depends on the pattern in the data and the context of the problem.

Linear Functions

Linear functions model data sets or contextual scenarios that show roughly constant rates of change. They have the general form $y = mx + b$, where m is the slope and b is the y-intercept. These work well when the output changes by about the same amount each time the input changes by a fixed step.

Example: A farmer wants to predict money made from selling crops based on acres planted, with this data:

• Number of acres planted (x): 0, 10, 20, 30, 40
• Money made from selling crops (y): 0, 800, 1600, 2400, 3200

The output increases by 800 for every 10 acres, so the rate of change is constant and a linear model fits. Using the points (0, 0) and (10, 800):

$m = \frac{y_2 - y_1}{x_2 - x_1} = \frac{800 - 0}{10 - 0} = \frac{800}{10} = 80$

$b = 800 - (80 \cdot 10) = 0$

So the model is $y = 80x$. To predict earnings for 50 acres, substitute x = 50: $y = 80(50) = 4000$ dollars.

Quadratic Functions

Quadratic functions model data sets or scenarios that show roughly linear rates of change, or data that is roughly symmetric with one unique maximum or minimum value. They have the general form $y = ax^2 + bx + c$, where a, b, and c are constants. The graph is a parabola with a single peak or valley.

Quadratics are a good choice when the rate of change itself is increasing or decreasing in a steady way, or when the data rises to a single high point and comes back down (or the reverse).

Some situations where a quadratic model can apply:

• The parabolic path of a projectile such as a thrown ball
• The height of a point on a roller coaster over a section of track
• Crop yield as a function of the amount of fertilizer applied, where too little and too much both lower the yield

These are applications of the concept, not required AP content. The key reason a quadratic fits is the underlying pattern, not the specific story.

Context Clues: Geometry

Geometric situations give strong hints about model type. Contexts involving area or two dimensions can often be modeled by quadratic functions. Contexts involving volume or three dimensions can often be modeled by cubic functions, which have the general form $y = ax^3 + bx^2 + cx + d$, where a, b, c, and d are constants.

If a problem asks about the area of a region as one side length changes, expect a quadratic. If it asks about the volume of a box as one dimension changes, expect a cubic.

Polynomial and Piecewise Functions

Polynomial functions model data sets or scenarios with multiple real zeros or multiple maxima or minima. A polynomial has the form $y = a_nx^n + a_{n-1}x^{n-1} + ... + a_1x + a_0$, where n is the degree and the a values are constants. The degree is the highest power of x.

Two patterns help you choose the degree:

• A polynomial of degree n models data that shows roughly constant nonzero nth differences. So if the first differences are constant, use degree 1 (linear). If the second differences are constant, use degree 2 (quadratic). If the third differences are constant, use degree 3 (cubic), and so on.
• A polynomial of degree n or less can model a graph of n + 1 points with distinct input values. With n + 1 points, you can find a polynomial that passes through all of them.

A piecewise-defined function is defined by different rules over different, nonoverlapping input intervals. It is useful when a situation behaves one way over one stretch of inputs and a different way over another. For example, a piecewise model can describe a process that follows one rule before a certain time and a different rule after.

Assumptions and Restrictions

Choosing a model is only half the job. You also have to describe what the model assumes and where it applies.

1. A model may have assumptions about what stays consistent. These can come from math clues or from the context. For example, a physical model might assume a rate or condition stays the same over the interval you care about.
2. A model may have assumptions about how quantities change together. For example, a model might assume two quantities stay directly proportional as they vary. These assumptions are what let the model make predictions.
3. A model may need domain restrictions based on mathematical clues, contextual clues, or extreme values in the data. The domain is the set of allowed input values. For example, a model of how far a projectile travels might only allow time greater than or equal to zero, since negative time has no meaning here.
4. A model may need range restrictions, such as rounding values, based on mathematical clues, contextual clues, or extreme values in the data. The range is the set of allowed output values. For example, a model that predicts a count of items might round to whole numbers, and a cost model might only allow nonnegative outputs.

As an example of a built-in restriction, the function $y = \frac{1}{x}$ cannot have x equal to 0, because the function is undefined there. Its domain is $(-\infty, 0) \cup (0, \infty)$ and its range is $(-\infty, 0) \cup (0, \infty)$.

Models are not perfect, and they do not capture every detail of a real situation. That is why naming assumptions and setting domain and range limits matters. A well-built model still captures the general trend, which is what makes it useful for predictions.

How to Use This on the AP Precalculus Exam

Choosing a Model

• Check the differences in the data. Constant first differences point to linear, constant second differences point to quadratic, and constant nth differences point to degree n.
• Read the context. Area or two-dimensional setups suggest quadratics, and volume or three-dimensional setups suggest cubics.
• Count features. Several zeros or several turning points suggest a higher-degree polynomial.

Free Response

• State your reasoning, not just your answer. If you pick a quadratic, say why, such as the data being symmetric with a single maximum or having constant second differences.
• Name at least one assumption your model relies on, and give domain or range restrictions that fit the context.
• Keep units and variable meaning clear so your explanation is easy to follow. This is important for clear exam work.

Common Trap

• Do not assume more points always means a higher-degree polynomial is better. The pattern in the data and the context should drive your choice, not just fitting every point exactly.

Common Misconceptions

• "A model has to match the data perfectly." A model is a simplified version of reality. It captures the general trend, and it is normal for it to not hit every point.
• "Constant differences and constant rate of change are the same for all functions." Constant first differences mean linear. Constant second differences mean quadratic. You have to check which level of differences is constant.
• "Domain and range have no limits unless the math forces them." Context often adds restrictions, like time being nonnegative or outputs being rounded to whole numbers, even when the equation alone would allow more values.
• "Any function that passes through the points is a valid model." Passing through points is not enough. The function type should match the pattern and the context, and a high-degree polynomial that fits every point can behave badly between or beyond the data.
• "Stating assumptions is optional." Describing assumptions and restrictions is part of building a model in this course, and free-response questions expect you to explain them.

Related AP Precalculus Guides

• 1.8 Rational Functions and Zeros
• 1.4 Polynomial Functions and Rates of Change
• 1.7 Rational Functions and End Behavior
• 1.11 Equivalent Representations of Polynomial and Rational Expressions
• 1.3 Rates of Change in Linear and Quadratic Functions
• 1.9 Rational Functions and Vertical Asymptotes