3.5 Model Selection and Representation

From Pattern to Model

The functions chapter comes together in modeling. AEPA Mathematics expects candidates to interpret different representations of functions and analyze real-world problems involving linear, polynomial, exponential, logarithmic, rational, radical, absolute value, and piecewise functions. A model-selection item may not say which family to use. The evidence is in the pattern, graph shape, units, or context.

A model is a mathematical representation of a situation. It can be an equation, graph, table, verbal rule, diagram, or recursive statement. Good modeling is not curve matching. It requires naming variables, deciding which variable is independent, choosing a reasonable family, interpreting parameters, and checking whether answers make sense in the original situation.

Function-Family Cues

These cues are starting points, not automatic answers. A table with only two points can fit infinitely many functions, so use context. A plan with a signup fee and a per-month charge is linear even if only two data points are shown. A bouncing-ball height over time is often quadratic for one flight, but repeated bounces may require a piecewise or exponential decay description.

A Modeling Routine

Use a repeatable routine on exam items:

1. Define variables and units.
2. Identify the type of change or graph feature.
3. Choose a function family and write the model.
4. Interpret coefficients, intercepts, factors, vertices, or asymptotes.
5. State contextual domain and range.
6. Check the answer by estimation or substitution.

Worked example: Plan A charges $35 plus $0.12 per minute. Plan B charges $20 plus $0.18 per minute. Let m be minutes and C be cost in dollars. The models are A(m) = 35 + 0.12m and B(m) = 20 + 0.18m. The fixed charges are the y-intercepts; the per-minute charges are slopes. To find the break-even point, solve 35 + 0.12m = 20 + 0.18m. Then 15 = 0.06m, so m = 250. Below 250 minutes, Plan B costs less because its lower fixed charge matters more. Above 250 minutes, Plan A costs less because its lower slope matters more. The contextual domain is m >= 0, and fractional minutes may or may not make sense depending on billing rules.

Representation Conversions

A graph shows shape, intercepts, extrema, and asymptotes quickly. A table shows differences and ratios. An equation gives exact calculation and parameters. A verbal description supplies units and constraints. AEPA-style answer choices often preserve one representation but distort another. For example, a graph may have the right intercept but the wrong slope, or an equation may match data points but use x and y in reverse.

For a line, slope in a graph should match the unit rate in the story. For an exponential model, the multiplier should match the percent change. For a quadratic, the vertex should match the maximum or minimum described. For a rational model, excluded input values may indicate physical impossibilities, such as dividing by a rate difference of zero.

Misconceptions and Model Limits

Teacher candidates need to recognize why a student's model is inappropriate. If data increase by 4, 4, 4, and a student chooses exponential because the graph rises, the misconception is confusing increasing with multiplicative growth. If a student forces a linear trend through curved data because two points determine a line, the misconception is ignoring the remaining data and context. If a student extrapolates a population model 500 years into the future, the issue may be model validity beyond the observed domain rather than algebra.

Also watch for discrete versus continuous variables. A function can use real-number inputs, but the context may restrict inputs to whole numbers. Number of students, row number, and item count are discrete. Time, length, and temperature are often continuous. A correct-looking equation can produce outputs that should be rounded, rejected, or interpreted cautiously.

Exam-Ready Model Checking

Before choosing an answer, ask three questions. Does the model match the change pattern? Do the parameters match the context? Are the domain and units reasonable? This catches most distractors. The calculator can compare numerical outputs, but the exam often rewards the explanation behind the match. A defensible model says what each number means, where the rule applies, and what behavior it predicts.

Comparing Candidate Models

When answer choices give several possible models, test structure before testing many numbers. A model can pass through one or two data points and still be wrong for the relationship. Check whether differences, ratios, intercepts, and units match the story. If a context says each additional hour adds the same fee, the slope should be constant. If a context says the amount retains the same percent each year, the multiplier should be constant.

Model quality also includes limits. A formula for ticket revenue may have a mathematical maximum at a noninteger price, but the context may require cents or whole-dollar pricing. A model for medicine concentration may approach 0 without becoming negative. Naming these limits turns a computation into the kind of mathematical reasoning the exam is designed to sample.