Comparing Linear and Exponential Models

When analyzing data or real-world scenarios, two of the most common patterns you will encounter are linear and exponential models. Knowing how to tell them apart is key to making accurate predictions.

Linear Models: Constant Differences

A linear model describes something that changes at a constant rate by adding or subtracting the same amount each time.

• Key feature: Constant differences between successive $y$-values (when $x$-values increase evenly).
• Equation: $y = mx + b$ (where $m$ is the constant rate of change).
• Example: A salary that increases by a flat \2000$every year. Every single year, the difference in salary from the previous year is exactly$$2000$.

Exponential Models: Constant Ratios

An exponential model describes something that changes by a constant multiplier or percentage.

• Key feature: Constant ratios between successive $y$-values (when $x$-values increase evenly).
• Equation: $y = a \cdot b^x$ (where $b$ is the constant ratio or growth factor).
• Example: A salary that increases by $3\%$ every year. To find the next year's salary, you multiply the current salary by $1.03$. The actual dollar amount of the raise grows as the salary grows.

Identifying the Model from Data

To determine which model fits a set of data, test the differences and the ratios between successive points.

Example: Does the data set $(0,2), (1,6), (2,18), (3,54)$ fit a linear or exponential model?

1. Check for Linear (Differences): Subtract successive $y$-values: $6 - 2 = 4$ $18 - 6 = 12$ $54 - 18 = 36$ The differences ($4, 12, 36$) are not constant. It is not linear.

2. Check for Exponential (Ratios): Divide successive $y$-values: $6 \div 2 = 3$ $18 \div 6 = 3$ $54 \div 18 = 3$ The ratio is constantly $3$. Therefore, this data fits an exponential model.

The Ultimate Winner: Exponential Growth

A fundamental rule in mathematics is that exponential growth will always eventually outpace linear growth.

Even if a linear model starts with a massive constant rate (like adding \1,000,000$a day) and an exponential model starts very small (like doubling a single penny every day), the multiplying effect of exponential growth will eventually catch up and permanently exceed the linear model. In our salary example, the$3%$raise might start smaller than the$$2000$raise, but as the base salary grows, that$3%$will eventually become much larger than a flat$$2000$ bump!