Error Estimation Miscellaneous p1/5

Error Estimation ()
(A) Definition
Error is a measure of the accuracy of the result.
It indicates how the result closes to the true value.

(B) Significant Figures ( )

1. Expressing significant figures
It should be expressed in two ways only.
(1) To quote an error associated with the measured value, i.e. (measured value error) unit.
(2) To express in scientific notation.

Eg1 An object of mass is estimated to lie between 9.235 g and 9.245 g. Write the result in appropriate form.

Remarks
The number of significant figures must be consistent ( ) with the accuracy of the measurement.(later)
Roundoff ( ) the final result to the number of significant figures. (<5 round down, >5 round up)

2. Addition and Subtraction

Roundoff the first column from the left and drop all the digits to its right.

Eg2 The length of 5 rods are 1.36 cm, 16.72 cm, 5 cm, 0.89 cm and 9.3 cm. What is the total length of the rods when
placed in a straight end to end?

Eg3 (a) A student has two $100 note in his pocket. After he has spent $3, he left $?

(b) No. of audience in a concert is estimated to be 200. If 3 men left, the estimated no. of audience becomes

3. Multiplication and Division

The final result has the same number of significant figures as the lowest number of significant figures among the
quantities.

Eg4 A toy car of mass 1.204 kg moves on a horizontal ground with speed 3.2 m s 1. The kinetic energy of the car is
Error Estimation Miscellaneous p2/5

Remark
It is better to carry extra two significant figures along the intermediate steps and the final answer is then rounded
off appropriately.
It is NOT good to copy all the digits displayed by the calculator.

(C) Sources of Errors

1. Instrumental limitations ( )
All measuring instruments have their limitations. It may be due to inaccurate scale or the fineness of the scale.
These errors CANNOT be reduced by taking repeated measurements.
Example: Meter rule having mm scale has a limitation of 0.5 mm.

2. Systemic errors ( )
These errors cause all measurement to be shifted systematically in one direction either larger or smaller than it
should be.
These errors CANNOT be reduced by taking repeated measurements.
Examples
Parallax ( ) in reading scale when viewing the scale always from one side.
A zero error ( ) on any scale.
A calibration error ( ).
A background count ( ) in a radioactivity experiment.
A biased stray magnetic field, electric field ( / ).
An error in meter rules due to thermal expansion.

3. Random errors ( )
They result from unknown and unpredicted variations in experiments. They may due to
(1) Random variations in the quantity being measured.
(2) Unintended slight changes of the conditions of the experiment.
(3) Random variations in the set up of the apparatus.

The effect of the random errors can be reduced by improving experimental techniques and repeating the
measurement a number of times i.e. becoming statistically insignificant.
Examples
Parallax in reading scale when viewing the scale in different directions.
Unpredicted fluctuation ( ) in air temperature or line voltage.
Unbiased estimates ( ) of measurement readings by the observer.
Nonuniformity of diameter of a wire.
Error Estimation Miscellaneous p3/5

Random errors

Systematic errors

Mean value of the True

above measurements value

4. Plain mistakes
These are careless mistakes such as misreading of scale, faulty arithmetic and faulty transcription ( ).

(D) Treatment of errors ( )

1. Instrumental limitations
The scale error is usually taken as half of the smallest division on the scale.

Eg5 The appropriate way in expressing the reading of the thermometer ( scale in oC ) shown is

0 10 20 30

Eg6 Timing Mr. Yip in running 100 m by a digital stop watch gives a reading of 10.12 s. If the reaction time of the
stop watch controller is 0.1 s, the appropriate way of expressing the time will be

2. Systematic error
There is no general rule for the estimation of these errors.
Examples
Check the zero error before experiment, and then eliminate it in the final record of the readings.
Set zero for electrical meters including C.R.O., multimeter, etc.
Check the average background count, and then eliminate it to become corrected count.

3. Random errors
Sample standard deviation ( ) of the data gives the measure of the random errors. (Not important
for AL)
Error Estimation Miscellaneous p4/5

(E) Estimation of errors

1. Three types of expressing errors
Suppose a reading x is obtained of a quantity whose true value is X.

(a) Absolute error ( ) X X x

X X
(b) Fractional error ( )
X x
X X
(c) Percen tage error ( ) 100% 100%
X x

2. Error in single quantity

Ex7 The error in single measurement of wavelength of sodium light is 1 nm. Taking more measurement will reduce
the random errors. For instance, the results are 587, 589, 588, 591, 587, 588, 590, 592, 590 and 589. Then the
mean gives 589.1 = 589 nm. The sample standard deviation is 2 nm. So the result is expressed as 589 2 nm.

3. Combing errors
(a) Sum and Difference
Z = A + B or Z = A B where A and B are independent. Then
Z A B

The errors are always added.

Eg8 If B = (15 2) and A = (76 3) then Z = A B =?

Eg9 The appropriate way of expressing the length of the wire shown is (cf. Eg5)

0 1 2 3 4

cm

In judging the image position, the minimum and maximum image distance measured are 30.0 and 32.0 cm
respectively. Since the error in judging actual position is (32.030.0)/2 = 1 cm, which is larger than the error in
using meter rule (0.1 cm), the result should be expressed as (31 1) cm.

(b) Product and Quotient

Z = A B or A B where A and B are independent. Then
Z A B

Z A B
Error Estimation Miscellaneous p5/5

(c) Power
Z = k An where k and n are nonzero constants with error free.
Z A
n
Z A

Proof of (b) and (c) above

Eg9 Find the maximum possible error associated in the measurement of the kinetic energy (E) of an object travelling
at velocity v if the mass m = (3.5 0.1) kg and v = (20 1) m s1.

(d) Other combinations

If special function such as sine, log are involved, it will be easier to find the maximum and minimum
possible values in order to find the errors.

Eg10In an experiment, x = (3.30 0.05) m and = (40o 1o) find the value of y = x cos.
y = x cos = 3.30 cos40o = 2.53 m
ymin = xmin cosmax = 3.25 cos41o = 2.45 m
ymax = xmax cosmin = 3.35 cos39o = 2.60 m

ymax y = 3.31 3.25 = 0.07 m

y ymin = 3.25 3.19 = 0.08 m

Maximum error = 0.08 m

y = (2.53 0.08) m