Reading an Exponential Model
An exponential model can summarize a quantity that changes by the same percentage over equal time intervals. Its formula is compact, but each symbol has a specific meaning. Correctly interpreting those meanings is essential before using the model to describe a situation or make a calculation.
We will focus on the form \(A(t)=A_0b^t\). Here, \(A(t)\) is the amount at time \(t\), \(A_0\) is the initial amount, and \(b\) is the multiplier for one unit of time. The interpretation depends on the time unit used for \(t\): if \(t\) is measured in months, one unit of \(t\) means one month.
The amount \(A(t)\) and initial amount \(A_0\) have the same measurement units. The base \(b\) is a dimensionless multiplier: it is a ratio, not an amount measured in grams, dollars, or people. The exponent \(t\) is the numerical count of the chosen time units, so the time unit must be stated when interpreting the model.
The Initial Amount
Set \(t=0\) to find the model's value at its starting time. Because \(b^0=1\) for \(b>0\), the formula gives \(A(0)=A_0\). Thus, \(A_0\) is not the amount after one period; it is the amount at the time designated as zero.
The choice of starting time matters. If \(t\) counts months since January 1, then \(t=0\) refers to January 1, and \(A_0\) is the amount on that date. If a different date is chosen as time zero, the initial amount in the model must refer to that new starting point.
The Base and the Percentage Change
The base \(b\) tells how much the amount is multiplied by when \(t\) increases by one. The amount at time \(t+1\) is \(A(t+1)=A_0b^{t+1}\). Since \(b^{t+1}=b^t b\), it follows that \(A(t+1)=bA(t)\). The same multiplier therefore relates the amounts at every pair of consecutive times one unit apart.
To express the change as a percentage, compare the change in amount with the amount before the change. Provided \(A(t)\ne 0\), the relative change over one time unit is \((A(t+1)-A(t))/A(t)\). Substituting \(A(t+1)=bA(t)\) gives \(b-1\). Multiplying this relative change by \(100\%\) expresses it as a percentage.
The base itself is not the percentage change. For instance, a base of \(1.08\) means the new amount is \(108\%\) of the previous amount, which is an increase of \(8\%\). A base of \(0.92\) means the new amount is \(92\%\) of the previous amount, which is a decrease of \(8\%\).
Worked Examples
Worked Example: Interpreting a Growth Model
A population is modeled by \(P(t)=1200(1.05)^t\), where \(t\) is the number of years since a study began and \(P(t)\) is measured in individuals. Interpret the parameters and find the predicted population after three years.
The initial amount is the coefficient: \(P(0)=1200(1.05)^0=1200\) individuals. The base is \(1.05\), so each increase of one in \(t\)—one year—multiplies the population by \(1.05\). The percentage change per year is \(100(1.05-1)\%=100(0.05)\%=5\%\). Since the base exceeds \(1\), this is growth.
For three years, substitute \(t=3\):
The model predicts \(1389.15\) individuals, or about \(1389\) individuals if a whole-person estimate is needed. As a check on the interpretation, the model's one-year value is \(P(1)=1200(1.05)=1260\), and \(1260/1200=1.05\), the stated multiplier. The unrounded prediction after three years follows directly from the formula.
Worked Example: Interpreting a Decay Model
The amount of a substance is modeled by \(M(t)=800(0.92)^t\), where \(t\) is time in hours and \(M(t)\) is measured in milligrams. Identify the initial amount and hourly percentage change, then find the amount after two hours.
At \(t=0\), \(M(0)=800(0.92)^0=800\) milligrams. The multiplier \(0.92\) applies for each one-hour increase in \(t\). Its percentage change is \(100(0.92-1)\%=100(-0.08)\%=-8\%\), so the amount decreases by \(8\%\) per hour.
Substituting \(t=2\) gives:
Check the consecutive-hour interpretation: \(M(1)=800(0.92)=736\) milligrams, and \(M(2)=736(0.92)=677.12\) milligrams. The second amount is exactly \(92\%\) of the first, confirming the multiplier and the \(8\%\) decrease.
Worked Example: Keeping Track of Time and Amount Units
A mass is modeled by \(S(t)=48(1.25)^t\), where \(t\) is the number of weeks and \(S(t)\) is measured in kilograms. State what each parameter means and evaluate the model after two weeks.
The initial mass is \(S(0)=48(1.25)^0=48\) kilograms. The multiplier \(1.25\) applies per week, not per day or per year. The percentage change per week is \(100(1.25-1)\%=25\%\), an increase because \(1.25>1\). The base is a dimensionless factor, while both \(S(t)\) and \(48\) are in kilograms.
After two weeks:
For a check, \(S(1)=48(1.25)=60\) kilograms, and \(S(2)=60(1.25)=75\) kilograms. Each week the mass is multiplied by \(1.25\). This verifies both the calculation and the meaning of the time unit.
Why the Interpretation Works
The parameter meanings follow from the structure of the formula, rather than from a label attached to a particular example. For any \(t\) for which the model is being used, and for a positive amount model, the following identities give a direct justification.
Since \(b>0\), \(b^0=1\), so \(A(0)=A_0b^0=A_0\). This identifies the initial amount.
Using exponent addition, \(A(t+1)=A_0b^{t+1}=A_0b^tb=bA(t)\). Thus \(b\) is the multiplier over one time unit.
The relative change is \((A(t+1)-A(t))/A(t)=(bA(t)-A(t))/A(t)=b-1\). Multiplying by \(100\%\) gives the percentage change.
The same comparison also explains why the percentage is constant from one period to the next: the ratio \(A(t+1)/A(t)\) is always \(b\), so the relative change is always \(b-1\). This reasoning assumes the amount is nonzero when the relative change is formed; \(A_0>0\) and \(b>0\) ensure positive, nonzero model values.
Common Mistake: Calling the Base the Percent Change
A frequent error is to read \(A(t)=500(1.08)^t\) as “an increase of \(108\%\) per time unit.” That confuses the new amount as a percentage of the old amount with the percentage increase.
For example, if the current amount is \(500\), multiplying by \(1.08\) gives \(500(1.08)=540\). The increase is \(540-500=40\), and the relative increase is \(40/500=0.08=8\%\), not \(108\%\). The base says that the new amount is \(108\%\) of the old amount; subtracting \(100\%\) gives the \(8\%\) increase.
Units and Limits of Interpretation
State the time unit and amount unit whenever you describe a model. In \(A(t)=A_0b^t\), the numerical value \(t=3\) means three units of the time scale stated in the model. If \(t\) is in months, it means three months. The model's base then gives the multiplier per month. The coefficient and the predicted amount share the same unit, such as dollars, grams, or individuals.
Do not change the time unit without also reconsidering the numerical value of the base. A multiplier described per month is not automatically the multiplier per year. The formula only tells you the multiplier for an increase of one in its specified \(t\)-scale. Also, if a model is used for fractional values of \(t\), evaluate the formula at those values; whether such predictions are meaningful in a particular setting depends on the situation being modeled.
Interpreting parameters does not establish that a model is a good description of real data. It tells you what the formula asserts: a starting amount and a constant multiplier per specified time unit. Whether those assumptions are appropriate is a separate modeling question.
Summary
- \(A_0\) is the amount at \(t=0\), with the same measurement units as \(A(t)\).
- \(b\) is the dimensionless multiplier for one unit of the specified time scale.
- The percentage change per time unit is \(100(b-1)\%\).
- A base above \(1\) gives growth; a base between \(0\) and \(1\) gives decay; a base of \(1\) gives no change.
- The meaning of \(t\), including its time unit and starting point, must be stated to interpret the model.
Check Your Understanding
For each question, pay attention to the difference between a multiplier and a percentage change.
- In \(Q(t)=250(1.04)^t\), what is the initial amount, and what is the percentage change per time unit?
- In \(R(t)=90(0.75)^t\), does the model describe growth or decay? What is the percentage change per time unit?
- If \(A(t)=A_0b^t\), show how evaluating the model at \(t=0\) identifies \(A_0\).
- A model uses \(t\) measured in days. What time interval does a one-unit increase in \(t\) represent, and what does its base describe?
- Explain why a base of \(1.12\) represents a \(12\%\) increase, not a \(112\%\) increase.
