2.14 Logarithmic Function Context and Data Modeling

When logarithmic models fit data

A logarithmic function models data that increases or decreases quickly at first, then slows down and grows/declines without bound — no horizontal asymptote, but a vertical asymptote appears at the input value where the argument equals zero. Use logarithmic regression when scatterplots show this rapid-then-slow, ever-increasing (or decreasing) pattern rather than leveling off.

Building and fitting log models

Given data, you can fit a model of form y = a + b·ln(x) using regression on a calculator or by transforming the data (plotting y vs. ln(x) should look linear if a log model fits). Check the fit visually or with residuals: small, patternless residuals suggest a good logarithmic model.

Interpreting the model in context

Once fitted, use the model to predict outputs for given inputs, estimate rates of change, or find inputs producing a target output (solving the log equation). Always interpret parameters in context: the coefficient b describes how fast output changes per multiplicative change in input, since log functions compress large input changes into smaller output changes.

Key terms

When to use a logarithmic model

When data rises/falls quickly at first, then slows, but keeps growing/declining without bound (no horizontal asymptote).

Form of logarithmic regression model

y = a + b·ln(x), where a and b are constants determined by regression.

Vertical asymptote in log models

Occurs at the x-value that makes the argument of the log equal to zero (often x = 0 for basic models).

Checking model fit visually

Plot y vs. ln(x); if the log model fits, this transformed plot should appear linear.

Meaning of parameter b in y = a + b·ln(x)

Represents how much y changes per multiplicative (percentage) change in x, not per unit change.

Try an AP-style question

pH = −log[H⁺]. If [H⁺] = 10⁻³, what is the pH?

1. −3
2. 30
3. 3
4. 0.001