When two quantities vary together, a function describes their relationship. Rates of change explain how the output responds to input changes, while concavity describes how those rates themselves change. These are related but different questions.
Average rate is change in output per input unit
Between inputs a and , with the average rate of change is
On a graph, this is the slope of the secant line connecting the two endpoint points. In context, attach units: if is distance in kilometers and input is time in hours, the average rate is kilometers per hour.
The numerator is a difference, not the output value. A car at position 80 km after two hours has not necessarily averaged ; its starting position might not have been zero.
Example: compare equal input intervals
For the values at equal to 0, 1, 2, and 3 are 0, 1, 4, and 9. Over the three unit intervals, output changes are 1, 3, and 5. Since each interval has length 1, these are also the average rates.
The function is increasing on this interval because outputs rise, and its average rates are increasing. The graph bends upward: it is concave up.
For on the same interval, outputs are 10, 9, 6, and 1. Rates are The function is decreasing, and its rates are becoming more negative, so it is concave down. “Decreasing” alone does not tell you the concavity.
For x squared, equal input steps produce increasingly large gains. The secant slopes are 1, 3, and 5.
For x squared, equal input steps produce increasingly large gains. The secant slopes are 1, 3, and 5.
Describe this figure
For x squared, equal input steps produce increasingly large gains. The secant slopes are 1, 3, and 5.
Marked points: filled marker at (0, 0); filled marker at (1, 1); filled marker at (2, 4); filled marker at (3, 9).
Increasing and concave down can occur together
Consider for positive . As increases, the output increases, but successive equal input increases produce smaller output gains. The curve rises while flattening: increasing and concave down.
Conversely, for positive decreases while flattening toward zero. Its rates remain negative but become less negative, so the curve is decreasing and concave up.
Keep two separate descriptions: the sign of the rate tells whether the function rises or falls; whether the rate increases or decreases tells the direction of concavity.
Example: unequal intervals need division
A tank holds 20 L at minute 0, 32 L at minute 2, and 47 L at minute 5. The first average rate is . The second is .
The second output increase, 15 L, is larger than the first, 12 L, but it occurred over a longer time. Comparing raw changes would incorrectly suggest the faster rate. Divide by the input interval before comparing.
These measurements suggest a slowing increase across the sampled intervals. A few data points alone do not prove that the rate decreases at every moment between them. Distinguish what the data show from what a smooth model assumes.
Linear, quadratic, and exponential patterns
A linear function has constant average rate over every interval. With equally spaced inputs, its first differences are constant.
A quadratic has constant second differences for equally spaced inputs. An exponential model has a constant ratio over equal input increments, so its percentage change is constant rather than its absolute change.
For example, outputs 3, 6, 12, 24 double at each step. Their differences are 3, 6, 12, so they do not fit a constant-rate linear model. Recognizing the pattern helps you choose the right function family.
Try it yourself
A quantity has values 50, 44, 40, and 38 at equally spaced times 0, 1, 2, and 3. Is it increasing or decreasing? Are its interval rates increasing or decreasing?
Answer: It is decreasing at the sampled times. The rates are which are increasing toward zero. A smooth model consistent with that pattern would be decreasing and concave up over the interval, though the table does not establish every intermediate behavior.
What to remember
Divide output change by input change, retain the units, and separate direction from bending. Equal spacing is essential when using differences to identify patterns.