Chapter 01.

06
Propagation of Errors

1. Addition of measured quantities

If you have measured values for the quantities X, Y, and Z, with
uncertainties X, Y, and Z, and your final result, R, is the sum or
difference of these quantities, then the uncertainty R is:

Here the upper equation is an approximation that can also serve as an upper
bound for the error. Please note that the rule is the same for addition and
subtraction of quantities.

Example:
Suppose we have measured the starting position as x1 = 9.3+-0.2 m and
the finishing position as x2 = 14.4+-0.3 m. Then the displacement is:

x = x2-x1 = 14.4 m - 9.3 m = 5.1 m

and the error in the displacement is:

(0.22 + 0.32)1/2 m = 0.36 m

2. Multiplication of measured quantities

In the same way as for sums and differences, we can also state the result for
the case of multiplication and division:

01.06.1
01.06.2 Chapter 01.06

Again the upper line is an approximation and the lower line is the exact
result for independent random uncertainties in the individual variables. And
again please note that for the purpose of error calculation there is no
difference between multiplication and division.

Example:
We have measured a displacement of x = 5.1+-0.4 m during a time of t =
0.4+-0.1 s. What is the average velocity and the error in the average
velocity?

v = x / t = 5.1 m / 0.4 s = 12.75 m/s

and the uncertainty in the velocity is:

v = |v| · [ (x/x)2 + (t/t)2 ]1/2 = 12.75 m/s · [(0.4/5.1)2 + (0.1/0.4)2]1/2 =

3.34 m/s

3. Multiplication with a constant

What if you have measured the uncertainty in an observable X, and you
need to multiply it with a constant that is known exactly? What is the error
then? This is easy: just multiply the error in X with the absolute value of the
constant, and this will give you the error in R:

If you compare this to the above rule for multiplication of two quantities,
you see that this is just the special case of that rule for the uncertainty in
c, c = 0.
Propagation of Errors 01.06.3

Example:
If an object is released from rest and is in free fall, and if you measure the
velocity of this object at some point to be v = - 3.8+-0.3 m/s, how long
has it been in free fall?

Answer: we can calculate the time as (g = 9.81 m/s 2 is assumed to be

known exactly)

t = - v / g = 3.8 m/s / 9.81 m/s2 = 0.387 s

The uncertainty in the fall time is then:

t = |-1/g| · v = 0.102 s2/m · 0.3 m/s = 0.03 s

4. Polynomial functions
What if there is a dependence of the result on the measured quantity X that
is not described by simple multiplications or additions? We will state the
general answer for R as a general function of one or more variables below,
but will first cover the special case that R is a polynomial function of one
variable X.

You see that this rule is quite simple and holds for positive or negative
numbers n, which can even be non-integers.

5. General functions
And finally, we can express the uncertainty in R for general functions of one
or more observables. If R is a function of X and Y, written as R(X,Y), then
the uncertainty in R is obtained by taking the partial derivatives of R with
respect to each variable, multiplication with the uncertainty in that variable,
and addition of these individual terms in quadrature.
01.06.4 Chapter 01.06

If a calculation is made with numbers that are not exact, then the calculation itself will have
an error. How do the errors in each individual number propagate through the calculations.

Example 1
Find the bounds for the propagation error in adding two numbers. For example if one is
calculating X  Y where
X  1.5  0.05 ,
Y  3.4  0.04 .
Solution
By looking at the numbers, the maximum possible value of X and Y are
X  1.55 and Y  3.44
Hence
X  Y  1.55  3.44  4.99
is the maximum value of X  Y .
The minimum possible value of X and Y are
X  1.45 and Y  3.36 .
Hence
X  Y  1.45  3.36
 4.81
is the minimum value of X  Y .
Hence
4.81  X  Y  4.99.

One can find similar intervals of the bound for the other arithmetic operations of
X  Y , X * Y , and X / Y . What if the evaluations we are making are function evaluations
instead? How do we find the value of the propagation error in such cases.
If f is a function of several variables X 1 , X 2 , X 3 ,.......,X n 1 , X n , then the maximum
possible value of the error in f is
f f f f
f  X 1  X 2  ....... X n 1  X n
X 1 X 2 X n 1 X n

Example 2
The strain in an axial member of a square cross-section is given by
F
 2
h E
where
F =axial force in the member, N
h = length or width of the cross-section, m
E =Young’s modulus, Pa
Propagation of Errors 01.06.5

Given
F  72  0.9 N
h  4  0.1 mm
E  70  1.5 GPa
Find the maximum possible error in the measured strain.
Solution
72
 3 2
(4  10 ) (70  10 9 )
 64.286 10 6
 64.286
  
  F  h  E
F h E
 1
 2
F h E
 2F
 3
h h E
 F
 2 2
E h E
1 2F F
  2 F  3 h  2 2 E
h E h E h E
1 2  72
 3 2
 0.9  3 3
 0.0001
(4  10 ) (70  10 )
9
(4  10 ) (70  10 9 )
72
 3 2
 1.5  10 9
(4  10 ) (70  10 9 ) 2
 8.0357 10 7  3.2143 10 6  1.3776 10 6
 5.3955 10 6
 5.3955
Hence
 (64.286  5.3955 )
implying that the axial strain,  is between 58.8905 and 69.6815

Example 3
Subtraction of numbers that are nearly equal can create unwanted inaccuracies. Using the
formula for error propagation, show that this is true.
Solution
Let
z  x y
Then
z z
z  x  y
x y
01.06.6 Chapter 01.06

 (1)x  (1)y
 x  y
So the absolute relative change is
z x  y

z x y
As x and y become close to each other, the denominator becomes small and hence create
large relative errors.
For example if
x  2  0.001
y  2.003  0.001
z 0.001  0.001

z | 2  2.003 |
= 0.6667
= 66.67%

Additional information: