Why Sign Changes Matter
A function’s average value over an interval is based on its signed integral. When the graph lies above the horizontal axis, it contributes positively; when it lies below, it contributes negatively. These contributions can cancel, so a graph with substantial positive and negative regions can still have a small signed average—or an average of zero.
This is different from adding the sizes of all the regions between the graph and the axis. The signed average records the net effect of the function, not the total amount of area on either side.
Definition of the Signed Average
Let \(f\) be continuous on an interval \([a,b]\), where \(a<b\). Its signed average value on that interval is its definite integral divided by the interval length.
The interval length \(b-a\) is positive. Thus, the sign of the average is the sign of the net integral: a positive net integral gives a positive average, a negative net integral gives a negative average, and a zero net integral gives an average of zero.
Separating Positive and Negative Contributions
For a continuous function, define its positive part and negative part by \(f_+(x)=\max(f(x),0)\) and \(f_-(x)=\max(-f(x),0)\). Both are nonnegative, and at every \(x\), \(f(x)=f_+(x)-f_-(x)\). The integral of \(f_+\) measures the total positive contribution, while the integral of \(f_-\) measures the magnitude of the negative contribution.
Here \(P\geq 0\) and \(N\geq 0\). The signed integral is \(P-N\), so the average is the net contribution divided by the interval length.
When the graph has finitely many crossings, these quantities can usually be computed by finding the zeros and integrating over the intervals between them. For example, if \(f\) is nonnegative on \([a,c]\) and nonpositive on \([c,b]\), then additivity gives \(\int_a^b f(x)\,dx=\int_a^c f(x)\,dx+\int_c^b f(x)\,dx\). The shared endpoint \(c\) is allowed because the function may equal zero there; the assumptions do not require it to be both strictly positive and strictly negative at that point.
Why the Separation Formula Works
The separation formula follows directly from the pointwise relationship between the positive and negative parts. This justification also explains exactly what is being counted as cancellation.
At each point, the function’s value is its positive part minus the magnitude of its negative part: \(f=f_+-f_-\).
By linearity of the definite integral, \(\int_a^b f=\int_a^b f_+-\int_a^b f_-\).
The average-value definition gives \(f_{\mathrm{avg}}=(P-N)/(b-a)\).
Thus, equal positive and negative contributions cancel exactly in the signed integral. The average then reflects what remains after that cancellation, spread over the interval’s length.
Worked Examples
Worked Example: A Linear Function with Equal Opposing Contributions
Find the signed average of \(f(x)=x-1\) on \([0,2]\), and identify its positive and negative contributions.
First find where the graph crosses the horizontal axis:
For \(0\leq x\leq 1\), \(x-1\leq 0\). For \(1\leq x\leq 2\), \(x-1\geq 0\). An antiderivative is \(F(x)=\frac{x^2}{2}-x\), since \(F'(x)=x-1\).
The negative signed contribution is
Its magnitude is \(N=\frac12\). The positive contribution is
Therefore \(P=\frac12\), and the net integral is \(P-N=\frac12-\frac12=0\). The interval length is \(2-0=2\), so
Check: Direct integration over the whole interval gives \(\int_0^2(x-1)\,dx=F(2)-F(0)=0-0=0\), which agrees with the separated contributions. The average is zero even though the function is not zero throughout the interval.
Worked Example: A Polynomial with Two Positive Regions
Find the signed average of \(f(x)=x^2-1\) on \([-2,2]\), separating its positive and negative contributions.
The zeros satisfy \(x^2-1=0\), so \(x=-1\) and \(x=1\). The function is nonnegative on \([-2,-1]\) and \([1,2]\), and nonpositive on \([-1,1]\). An antiderivative is \(F(x)=\frac{x^3}{3}-x\), because \(F'(x)=x^2-1\).
On the left positive interval,
On the right positive interval,
Thus \(P=\frac43+\frac43=\frac83\). The signed integral over the central, nonpositive interval is
Hence \(N=\frac43\), and the net integral is \(P-N=\frac83-\frac43=\frac43\). The interval length is \(2-(-2)=4\), giving
Check: Integrate directly across the full interval: \(\int_{-2}^2(x^2-1)\,dx=F(2)-F(-2)=\frac23-(-\frac23)=\frac43\). Dividing by \(4\) again gives \(\frac13\). The positive contribution is larger than the negative contribution, so the average is positive.
Worked Example: A Shifted Sine Function
Find the signed average of \(f(x)=\sin x-\frac12\) on \([0,2\pi]\), and compare its positive and negative contributions.
The zeros occur where \(\sin x=\frac12\). On this interval they are \(x=\frac{\pi}{6}\) and \(x=\frac{5\pi}{6}\). The function is positive between these zeros and nonpositive on the outer intervals. An antiderivative is \(F(x)=-\cos x-\frac{x}{2}\), since \(F'(x)=\sin x-\frac12\).
The positive contribution is
To find the total signed integral, integrate over the whole interval:
Since the net integral is \(P-N=-\pi\), the negative magnitude is \(N=P+\pi=\sqrt{3}+\frac{2\pi}{3}\). The interval length is \(2\pi\), so the signed average is
Check: The separated contributions give \(P-N=(\sqrt{3}-\frac{\pi}{3})-(\sqrt{3}+\frac{2\pi}{3})=-\pi\). Dividing by \(2\pi\) gives \(-\frac12\), matching the direct integral. The negative contribution is larger, so the average is negative.
Signed Average Versus Average Absolute Value
The signed average and the average absolute value answer different questions. The signed average uses \(P-N\), so opposing contributions cancel. The average absolute value uses \(P+N\), so it measures the average size of the function without regard to sign.
For \(f(x)=x-1\) on \([0,2]\), the signed average is \(0\), while the average absolute value is \((\frac12+\frac12)/2=\frac12\). Confusing these quantities changes the question being answered: one describes net contribution, the other describes total magnitude.
A Common Mistake
A frequent error is to replace every contribution with its positive area before calculating the signed average. For \(f(x)=x-1\) on \([0,2]\), this incorrect calculation uses the magnitudes of both triangles:
This is not the signed average; it is the average absolute value. The region on \([0,1]\) lies below the axis and contributes \(-\frac12\), not \(+\frac12\), to the signed integral. The correction is
Before dividing by the interval length, check whether each integral is being used with its sign. Geometric area is nonnegative, but a definite integral over a region below the axis is negative.
A Reliable Calculation Strategy
When the graph changes sign, use the following sequence to keep the signs and interval boundaries clear:
These identify where the graph may change sign.
Use the function’s formula or a test value in each interval.
Keep negative contributions negative; if listing their magnitudes separately, subtract those magnitudes from the positive total.
The denominator is \(b-a\), even when the function changes sign several times.
The sum of the signed contributions must equal \(\int_a^b f(x)\,dx\).
Check Your Understanding
Answer these questions using signed contributions, not just the sizes of geometric regions.
- For a continuous function on \([a,b]\), what formula defines its signed average value?
- If the positive and negative contributions have equal magnitude, what is the signed integral and signed average?
- A function is nonnegative on \([a,c]\) and nonpositive on \([c,b]\). How do you express its signed integral over \([a,b]\) using the two subinterval integrals?
- For \(f(x)=x-1\) on \([0,2]\), what are the positive contribution, the magnitude of the negative contribution, and the signed average?
- How does the average absolute value differ from the signed average when a function changes sign?
