Starting Soon

Uncertainty in measurement

Md. Shaidur Rahman

– Assistant Director (Met), Head of Length &
Dimension Measurement Lab. NML-BSTI
– [email protected]

2
Acceptance of the Measurement Results

3
 Outline

• Introduction
• Basic Statistics
• What is Measurement Uncertainty ?
• How is Measurement Uncertainty estimated?
• How can Measurement Uncertainty be reported?
• Conclusions and Remarks

4
 When should you evaluate
uncertainties of measurement results ?

• When a procedure is introduced inside your

laboratory
• When a critical factor changes in the procedure
(instrument, operator, …)
• During / together with procedure validation
 An individual evaluation process is NOT needed for
every individual result produced !

5
 Why do we need uncertainty?
• It is required by ISO 17025 - Accreditation
• The uncertainty of the result demonstrates the
metrological QUALITY of the measurements
• It improves the knowledge about the measurement
procedure
• In laboratory  document/prove in transparent way
the measurement procedure
• For end-user  give the result with proper confidence
• To allow comparison of results

6
 Why do we need uncertainty? (2)
• A well documented uncertainty statement
underpins/Support/ justify your results and
provides transparency!
• Identify major uncertainty contributors
- find out ways to improve the procedure
• Demonstrate compliance with limits (legal or
contractual) and the establishment of acceptance
criteria
 Your best defence in discussions!
 Repeating the measurement 2, 10 or 100 times
does not give you all information to have reliable
results!

7
 What is Measurement?

Measurement is
‘Set of operations having the object of
determining a value of a quantity.’ (
VIM 2.1 )
Note: The operations may be
performed automatically.

8
 Measurement

The objective of a measurement is

to determine the value of the
measurend, that is, the value of the
particular quantity to be measured

9
 Measurand

• Particular quantity subject to measurement

( VIM 2.6 )
• Example : vapour pressure of a given sample of
water at 20 OC
• Note : The specification of a measurand may
require statements about quantities such as time,
temperature and pressure.

10
 Principles of Measurement
A measurement begins with
• an appropriate specification of the measurand
• the method of measurement and
• the measurement procedure

Method
DUT of Result
Comparison

Standard

11
 ISO/IEC 17025

• 5.4.7.2
– The laboratory shall use methods which
meet the needs of the client

12
 Measurements
 Every measurement is subject to some
uncertainty.
 A measurement result is incomplete without a
statement of the uncertainty.
 When you know the uncertainty in a
measurement, then you can judge its fitness for
purpose.
 Understanding measurement uncertainty is the
first step to reducing it

13
 Measurement Error
Measurement Error

Real Number System

Measured Value True Value

Measured values are inexact observations of a

true value.
The difference between a measured value and
a true value is known as the measurement
error or observation error.

14
 Random Errors

Result of a measurement minus

the mean result of a large number
of repeated measurement of the
same measurand.( VIM 3.13 )

15
 Random Errors
• Random errors result from the fluctuations in
observations

• Random errors may be positive or negative

• Measurements with random errors have no obvious bias

– The average bias approaches 0 as more
measurements are taken
– Requires that experiments be repeated a sufficient
number of time to establish the precision of
measurement

16
 Random Errors
• Some sources of random errors include the
following:
– Exerting a different pressure when measuring
with a micrometer
– Reading a tape measure from different angles
– Connecting an instrument in a slightly
different fashion each time

17
 Systematic Errors

same measurand minus a true value of

the measurand. ( VIM 3.14 )

18
 Systematic Errors
• A systematic error is a consistent deviation in a
measurement
• A systematic error is also called a bias or an offset

• Systematic errors have the same sign and magnitude

when repeated measurements are made under the
same conditions

• Statistical analysis is generally not useful, but rather

corrections must be made based on experimental
conditions.

19
Measurement Error

20
Precision, Accuracy, and Bias
• Precision:
– Precision is a measure of the repeatability of a
measurement
– Small random error
• Accuracy:
– Accuracy is a measure of the closeness of all
measurements to the true value
– Small random error and small systematic error
– Small total error
• Bias:
– Bias is a measure of the closeness of the average
measurement to the true value
– Small systematic error

21
 Precision and Accuracy

Measurements that are close to the “correct”

value are accurate.

Measurements which are close to each other are

precise.

Measurements can be
- accurate and precise
- precise but inaccurate
- neither accurate nor precise

22
 Precision, Accuracy, and Bias

Low Precision Low Precision

Large Bias Small Bias

High Precision High Precision

Large Bias Small Bias

ACCURATE

23
 Uncertainty

„a parameter associated with the result

of a measurement, that characterises
the dispersion of the values that could
reasonably be attributed to the measured ‟

Result = Value ± uncertainty

(22.7 ± 0.5) mg/kg

The value is between 22.2 mg/kg and 23.2 mg/kg

24
 Uncertainty
• Estimated quality of a result is expressed as
the uncertainty

•The uncertainty is an essential part of a

measurement result:
Measurement result = Estimate ± uncertainty

•The uncertainty gives the limits of the range in

which the “true” value of the measurand is
estimated to be at a given probability..

25
Basic Statistics
 Systematic Errors
• Three types of systematic errors:
1.Natural error resulting from
environmental effects
2.Instrument error resulting from improper
calibration
3.Personal error resulting from habits of
the observer

27
 Histograms
• When making many measurements,
there is often variation between
readings. Histogram plots give a visual
interpretation of all measurements at
once.
• The x-axis displays a given
measurement and the height of each
bar gives the number of measurements
within the given region.
• Histograms indicate the variability of
the data and are useful for determining
if a measurement falls outside of
“specification”.

28
For a large number of experiment replicates the results
approach an ideal smooth curve called the GAUSSIAN or
NORMAL DISTRIBUTION CURVE

Characterised by:

The mean value – x

gives the center of the
distribution

The standard
deviation – s
measures the width
of the distribution

29
 Average
• The most basic statistical tool to analyze a series of
measurements is the average or mean value :
“Sum of”
Individual measurement

x
 x i

n Number of measurements

The average of the three values 10  15  12.5

10, 15 and 12.5 is given by: x  12.5
3

30
 Deviation

Deviation = individual value – avg value

d i  xi  x
Need to calculate an average or “standard” deviation

To eliminate the possibility of a zero deviation, we square di

31
 Standard Deviation
• The average amount that each measurement deviates
from the average is called standard deviation (s) and is
calculated for a small number of measurements as:

s
( x  x )2
i
Sum of deviation squared

n 1 xi = each measurement
x = average
n = number of measurements

Note this is called root mean square: square root of the mean of the squares

32
Standard Deviation

33
 Standard Deviation
For example, calculate the standard deviation of the
following measurements: 10, 15 and 12.5 (avg = 12.5)

s
 (x  x)
i
2


10  12.52  15  12.52  12.5  12.52
n 1 n 1

(2.5) 2  (2.5) 2 12.5

   2.5
2 2
The values deviate on average plus or minus 2.5 :12.5 ± 2.5
10.0 12.5 15.0

34
Standard Deviation of the Mean

The uncertainty in the best measurement is given

by the standard deviation of the mean (SDOM)

s

n

38
 Gaussian Distribution
• Given a set of repeated measurements which
have random error.
• For the set of measurements there is a mean
value.
• If the deviation from the mean for all the
measurements follows a Gaussian probability
distribution, they will form a “bell-curve”
centered on the mean value.
• Sets of data which follow this distribution are
said to have a normal (statistical) distribution of
random data.
39
 POPULATION DATA
For an infinite set of data,
n→∞ x → µ and s→σ
population mean population std. dev.

The experiment that produces a small

standard deviation is more precise .
Remember, greater precision does not
imply greater accuracy.

Experimental results are commonly

expressed in the form:
_
mean  standard deviation xs

40
 A. Normal Error Curve
m
• 68.3% of measurements
Relative frequency, dN / N

will fall within ±  of the

-1 +1 mean.

• 95.5% of measurements
will fall within ± 2 of
the mean.
-2 +2
• 99.7% of measurements
-3 +3 will fall within ± 3 of
the mean.
30 40 x
50i 60 70

41
 Gaussian Distribution

– For an interval of -1 < x < +1, one can be confident that any
random measurement of x will lie within one standard deviation of
the mean value 68.3% of the time.
– An interval of -2 < x < +2, one can be confident that 95.4% of
the time, x will lie within two times the standard deviation of the
mean value.

42
 95% Confidence Interval

– The 2 interval, or the 95% confidence interval is

the engineering standard.
– It is the implied uncertainty of a value even when
not written explicitly.
Resistor value is given as 100  10  means that the true
value of a marked 100 ohm resistor will be between:
90 and 110 ohms, 95% of the time
95 and 105 ohms, 68% of the time

43
EXAMPLE
Replicate results were obtained for the measurement
of a resistor. Calculate the mean and the standard
deviation of this set of data.

Replicate ohms
1 752
2 756
3 752
4 751
5 760

44
_
x  xi
Replicate ohms
n 1 752
752  756  752  751  760
  754.2 2 756
5
3 752
 x  x
2
4 751
s i

n 1 5 760


752  754.22  756  754.22  752  754.22  751 754.22  760  754.22
5  1


 2.22  1.82   2.22   3.22  5.82
4
 3.77 NB DON’T round a
std dev. calc until
s 3.77
   1.69 the very end.
n 5
45
 x  754.2
s  3.77   1.69
Also:
s 3.77
RSD  
x 754
s 3.77
%RSD   100   100
x 754

Variance  s 2  3.77 
2

46
 Student's t-Distribution
• The t distributions were
discovered by William S.
Gosset in 1908. Gosset
was a statistician employed
by the Guinness brewing
company which had
stipulated that he not
publish under his own
name. He therefore wrote
under the pen name
``Student.''

47
52
53
 Propagation of Error
• Typical experimental methods of analysis involve several
steps, each of which has an uncertainty associated that
contributes to the net uncertainty of the analysis.

A. Addition and subtraction

x=p+q-r Where p, q, & r each have uncertainties
sp , sq , & sr

sx = sp2 + sq2 + sr2

56
 Propagation of Error

A. Addition and subtraction

x=p+q-r Where p, q, & r each have uncertainties
sp , sq , & sr

sx = sp2 + sq2 + sr2

58
 Propagation of Error
• Addition and subtraction
Example, suppose you needed to perform the following
calculation:
1.76  0.03  1.89  0.02  0.59  0.02
• The following formula has been derived to calculate the final
uncertainty:

sx  s  s  s 2
p
2
q
2
r

sx  0.032  0.022  0.022

 0.04

59
 Propagation of Error
• The error in the final answer is calculated as:

1.76  0.03  1.89  0.02  0.59  0.02

 3.06  0.04

60
Guide to the Expression of
Uncertainty
in Measurement (1993)
 Guide to the Expression of Uncertainty
in Measurement (1993)

BIPM - International Bureau of

International Organisations Weights and Measures
http//: www.bipm.org
IEC - International Electrotechnical
Commision
http//: www.iec.ch
IFCC - International Federation of
Clinical Chemistry
http//: www.ifcc.org
IUPAP - International Union of
Pure and Applied Physics
http//: www.iupap.org
IUPAC - International Union of
Pure and Applied Chemistry
http//: www.iupac.org
ISO - International Organisation for
Standardisation
http//: www.iso.ch
OIML - International Organisation
for legal metrology
http//: www.oiml.org

67
 Introduction to GUM

• When reporting the result of a measurement of a

physical quantity, it is obligatory that some
quantitative indication of the quality of the result
be given so that those who use it can assess its
reliability.
• Without such an indication, measurement results
can not be compared, either among themselves
or with reference values given in the
specification or standard.
– GUM 0.1

68
 Uses of MU
• QC & QA in production
• Law enforcement and regulations
• Basic and applied research
• Calibration to achieve traceability to national
standards
• Developing, maintaining, and comparing
international and national reference standards and
reference materials

–GUM 1.1

69
 Terms specific to the GUM
• standard uncertainty,
– the uncertainty of the result of a measurement expressed as
a standard deviation
GUM 2.3.1
• Type A evaluation (of uncertainty)
– method of evaluation of uncertainty by the statistical
analysis of a series of observations

GUM 2.3.2
• Type B evaluation (of uncertainty)
– method of evaluation of uncertainty by means other than
the statistical analysis of series of observations

GUM 3.2.3

70
 Terms specific to the GUM

• combined standard uncertainty

the standard deviation of the result of a
measurement when the result is obtained from the
values of a number of other quantities. It is obtained
by combining the individual standard uncertainties
(and covariances as appropriate), using the law of
propagation of uncertainties, commonly called the
"root-sum-of-squares" or "RSS method.

GUM 2.3.4

71
 Terms specific to the GUM
• expanded uncertainty
– quantity defining an interval about the result of a
measurement that may be expected to
encompass a large fraction of the distribution of
values that could reasonably be attributed to the
measurand.
GUM 3.2.5

• coverage factor, k
– numerical factor used as a multiplier of combined
standard uncertainty in order to obtain expanded
uncertainty
GUM 3.2.6

72
 Basic concepts
• Result of a measurement
–is only an estimate of a true value and only
complete when accompanied by a statement
of uncertainty. GUM 3.1.2
–is determined on the basis of series of
observations obtained under repeatability
conditions GUM 3.1.4

• Variations in repeated observations are

assumed to arise because influence
quantities Gum 3.1.5

73
 Influence quantity

• Quantity that is not the measurand but that

affects the result of measurement.
Example - temperature of a micrometer used to
measure length.

( VIM
2.7 ) 74
 Basic concepts
• The error in a measurement
– Measured value – True value.
– This is not known because:

• The true value for the measurand

– This is not known
GUM 2.2.4

– The result is only an estimate of a true value and only

complete when accompanied by a statement of
uncertainty. GUM 3.2.1

75
 Basic concepts

• Random error
– Presumably arises from unpredictable
temporal and spatial variations
– gives rise to variations in repeated
observations
– Cannot be eliminated, only reduced.

GUM 3.2.2

76
 Basic concepts
• Systematic error
– If a systematic error arises from a recognized effect of
an influence quantity
• the effect can be quantified
• can not be eliminated, only reduced.
• if significant in size relative to required accuracy, a
correction or correction factor can be applied to
compensate GUM 3.2.3
• then it is assumed that systematic error is zero.
• It is assumed that the result of a measurement has
been corrected for all recognised significant
systematic effects GUM 3.2.4
77
 What is measurement uncertainty?
“A parameter associated with the result of a measurement,
that characterises the dispersion of the values that could
reasonably be attributed to the measurand”. (VIM definition)

Xu
“The uncertainty represents the range of values that the
analyst believes could reasonably be attributed to the
measurand.”
“The uncertainty is a range in which we expect and claim
our result to lay in with a certain confidence.”

78
What is NOT a measurement uncertainty!

• It is not an error!
• It is not traceability!
• It is not just a statistical factor (parameter):

79
 Uncertainty
• Estimated quality of a result is expressed as
the uncertainty
•The uncertainty is an essential part of a
measurement result:
Measurement result = Estimate ± uncertainty
•The uncertainty gives the limits of the range in
which the “true” value of the measurand is
estimated to be at a given probability..

80
 Uncertainty
• The result of a measurement after
correction for recognized systematic
effects is still only an estimate of the
value of the measurand because of the
uncertainty arising;
– from random effects and
– from imperfect correction of the result for
systematic effects
GUM 3.3.1

81
82
83
 Error versus uncertainty

• It is important not to confuse the terms error and

uncertainty
• Error is the difference between the measured value and
the “true value” of the thing being measured
• Uncertainty is a quantification of the doubt about the
measurement result
• In principle errors can be known and corrected
• But any error whose value we do not know is a source of
uncertainty.

84
Process of Uncertainty Estimation

• Specify Measurand

• Identify all Uncertainty Sources

• Quantify Uncertainty Components

• Calculate Combined Uncertainty

85
1. Specify the Measurand
 The measurand?

Measurand = particular quantity

subject to measurement
[VIM 2.6 / GUM B.2.9]

Example: the conventional mass

of a 1kg weight.
GUM 1.2

87
2. Identify all Uncertainty
Sources
 Identify all Uncertainty Sources

• Listthe possible sources of uncertainty. This will

include sources that contribute to the uncertainty on
the parameters in the relationship specified in Step 1,
but may include other sources and must include
sources arising from chemical assumptions.

•Tool for forming a structured list is the Cause and

Effect diagram.

89
 ISO/IEC 17025
• 5.4.7.2
– attempt to identify all the components of
uncertainty
• 5.4.7.3
– All uncertainty components which are of
importance shall be taken into account

90
 Possible Sources of MU
– incomplete definition of the measurand
– imperfect realization of the definition
– imperfect mathematical model
– sampling method
– uncertainties in values of measurement standards
and reference materials
– uncertainties in constant or other parameters
obtained from other sources
– environmental factors
– random variation in repeated observations
– instrument resolution
– etc

91
 Sources of uncertainty

Sources that contribute to uncertainty may

include
• condition of the sample
• calibrators,
• reference standards,
• Method used,
• equipment used,
• operator,
• environmental conditions

92
 Causes for uncertainty
Measurement Measuring
standard methods
Calibration certificate
Secular change

Measuring Measurement
instrument results Measurer

Manufacturer’s Peculiarities in
specification readout
Resolution
Measurement Number of
environment measurements
Dispersions in repetition

93
 3. Quantify
Uncertainty Components
 Quantify
Uncertainty Components
• Measure or estimate the size of the uncertainty
component associated with each potential source of
uncertainty identified.
• It is often possible to estimate or determine a
single contribution to uncertainty associated with a
number of separate sources.
• It is also important to consider whether available
data accounts sufficiently for all sources of
uncertainty.
• If necessary plan additional experiments and
studies carefully to ensure that all sources of
uncertainty are adequately accounted for.

95
 Classification of uncertainty
components
• Type A components: those that are evaluated
by statistical analysis of a series of
observations
• Type B components: those that are evaluated
by other means
– Both based on probability distributions
• standard uncertainty of each input estimate is obtained
from a distribution of possible values of input quantity:
both based on the state of our knowledge
• Type A founded on frequency distributions
• Type B founded on a priori distributions

96
 Measurement Model

For each input quantity

1. Obtain knowledge of that quantity
2. Assign a probability distribution to each quantity
consistent with that knowledge
Often a Gaussian (normal) or a rectangular distribution

99
 Type A Evaluation
• For component of uncertainty arising from
random effect Applied when multiple
independent observations are made under
the same conditions
• Data can be from repeated measurements,
control charts, curve fit by least-squares
method etc
• Obtained from a probability density function
derived from an observed frequency
distribution (usually Gaussian)

100
 Type A Evaluation
Arithmetic mean
• Best estimate of the expected value of a
input quantity -

1 n
q   qk
n k 1

101
 Type A Evaluation
Experimental standard deviation

Distribution of the quantity

102
 Type A Evaluation
Experimental standard deviation of the mean
• spread of the distribution of the means -

__
s(qk )
s( q ) 
n

103
 Type A Evaluation
• Type A standard uncertainty

u ( xi )  s(q)
• degrees of freedom

 i  n 1
104
Example
A digital multimeter is used to measure a high
value resistor and the following readings are
recorded.

The standard uncertainty, u, is therefore 0.008 83

kΩ.
105
 Type B Evaluation of Standard
Uncertainty
• Previously measured data.
• Experience with or general knowledge of
the behavior and properties of relevant
materials and instruments.
• Manufacturers specifications.
• Data provided in calibration and other
certificates.
• Uncertainties assigned to reference data
taken from handbooks.
GUM 4.3.1

108
 Type B Evaluation
•Assumed Normal distribution
If the value is obtained from a calibration certificate,
standard uncertainty and degrees of freedom need to be
retrieved from cal certificate
– Standard uncertainty
u(xi) = U/k
(U -reported expanded uncertainty in a certificate,
k - coverage factor)
k = 1.96 95% C.L
k=2 95.45% (approx 95%) C.L. (2 )
k=3 99.7% C.L (3 )
– degrees of freedom
check from t-distribution table

109
 Type B Evaluation
Normal distribution
Example
A calibration certificate reports the measured value of
a nominal 1kg OIML weight class F2 at approximately
95% confidence level as:

0.999999kg  10mg

u B  u  xi  
quoted uncer 10mg
  5mg
k 2

110
 Normal Distribution

Frequency

±1s

±2s

-4 -2 0 2 4

0.999989kg 0.999999kg 1.000009kg

The measured value of a nominal 1kg weight

111
 Rectangular distribution
“It is likely that the value is somewhere in
that range”
Rectangular distribution is usually described in terms of: the
average value and the range (±a)Certificates or other
specification give limits where the value could be, without
specifying a level of confidence (or degree of freedom).

The value is
between the limits 2a(=  a)
a  a
The expectation
y xa
1/2a

X
112
 Rectangular
distribution
Range = 2a ,
Semi-range = Range /2 = a
a a

P=1/
2a
A B

113
 Rectangular
B
distribution
B
   x Pdx  P  x dx
2 2 2

A A
B
P 
1
2a
 1  2  1  3
;      x dx    X / 3 A
2 B
 
 2a  A  2a 
B  a ; A  a
 1
 2   
   a 3


 a 3
 a 2


 2a   3 3  3
a
 
3
114
Example of Rectangular distribution

Example
• From the previous example, if the
Maximum Permissible Error (MPE)
according to OIML class F2 is used;
then

a 16mg
uB    9. 23mg
3 3

115
 Example of Rectangular
distribution
Handbook
• A Handbook gives the value of coefficient of
linear thermal expansion of pure copper at 20 C

 20 Cu   16.52106 / C
and the error in this value should not exceed,
 0.40 106 / C assuming rectangular distribution the
standard uncertainty is:
semi  range; a  0.40 106
0.40 106
u xi    0.23106 / C
3

116
Rectangular distribution

a a

16. 12  10 6 / C 16. 52  10 6 / C 16. 92  10 6 / C

117
 Example of Rectangular distribution

• From Manufacturer’s Specification

Temperature indicator range 0-100 OC ,accuracy ± 1 OC
of full scale

semi - range ( a ) = 1 OC

a 1C
u xi     0.6C
3 3

118
 Example of Rectangular distribution
Resolution of a digital indication
•If the resolution of 6
the digital device is 1 2 3 4 5
dx, the value of X 4
can lie with equal Range   x
probability anywhere in x
the interval X - dx /2 Semi - range 
2
to X + dx /2 and
thus described by a u ( xi ) 
x
rectangular probability 2 3
distribution with the
 0.29 x
width dx
119
 Example of Rectangular distribution
•Example
•A digital balance
having capacity of 210g
and the least significant
digit 10 mg. The
standard uncertainty
contributed by this
u xi  
0,01
balance is: g
2 3
 2.9 mg

120
 Triangular distribution
Distribution used when it is suggested that values near the
center of range are more likely than near to the extremes
2a (=±a)

y xa 2a (=a)

1/a
Assumed standard deviation:

s  a 1 / 6 X

122
 Example of Triangular distribution

Values close to x are more likely than near the

boundaries than for rectangular distribution.
Example (volumetric glassware)
The manufacture quotes a volume for the flask of
(100 ± 0.1) ml at T = 20° C.
Nominal value most probable!
Assuming triangular distribution the standard uncertainty is:

s  u ( x )  a 1 / 6  0 . 1 / 6  0 . 0 4 m l

In case of doubt, use the rectangular distribution

123
 Type B Evaluation
Rectangular Triangular
a a a a

a_ m a+ a_ m a+

a  a
Best estimate m =
2
Standard uncertainty of the best estimate, m
a a
3 6
Degrees of freedom = 
124
 Which is better A or B?
It should be recognized that a Type B
evaluation of a standard uncertainty can
be as reliable as a Type A evaluation,
especially in a measurement situation
where a Type A evaluation is based on a
comparatively small number of
statistically independent observation.
GUM 4.3.2

125
 4. Calculate
combined uncertainty

126
 Uncorrelated input quantities
For uncorrelated input quantities r (xi , xj)
=0 N 1 N
Then 2  c c ux ux r x , x   0
i 1 j i 1
i j i j i j

N 2

uc2  y    ciuxi 
i 1

For ci =1
combined standard uncertaint y, u c  y 
 
u c y  u x1  u x 2  u x 3  ...u x n
2 2 2 2

133
Determine the expanded uncertainty
U = k uc(y)
• k - coverage factor
– chosen from the desired level of confidence
– requires detailed knowledge of the probability distribution
– under certain conditions, normal distribution can be
assumed due to Central Limit Theorem
– for normal distribution: chosen from t -distribution table
according to effective degrees of freedom
• Level of confidence
Most cal labs adopt 95.45% which gives k  2
for effective degrees of freedom  30

137
 Coverage factor
• k=1.00 68.27% confidence
• k=1.64 90%
• k=1.96 95%
• k=2.00 95.45%
• k=2.58 99%
• k=3.00 99.73%
• One can assume that taking k=2 produces an
interval having a confidence of 95% and taking
k=3 produces an interval having a confidence
interval of 99%. GUM 6.3.3

138
 Effective number of degrees of
freedom
• Welch-Satterthwaite

4
uc ( y )
 eff
 N 4
ui ( y )

i 1  i

140
Effective number of degrees of freedom
Example
• A steel rod is measured 4 times. The calculated .

u A  3.5 mm and uB  2.3 mm;

uc  y   4.2 mm
• The effective degree of freedom:
uc  y  4.2
4 4
veff  4   6.22
u A u B

4
3.5  04
4

vA vB 4  1
• Therefore;  eff  6 @95% confidence level and from
“t student” table, we get k = 2.52
141
 Effective number of degrees of
freedom
Therefore,the expanded uncertainty U is:

U  k .uc  y 
 2.52  4.2
 11 mm.

142
Reporting Result
 Reporting results
• should include
– result of measurement
– expanded uncertainty with coverage factor and
level of confidence specified
– description of measurement method and
reference standard used
– uncertainty budget
• example of uncertainty statement
e.g.The expanded uncertainty of measurement is ±
____ , estimated at a level of confidence of
approximately 95% with a coverage factor k =
____.

144
 Reporting results
The numerical value of the uncertainty of measurement
should be given to at most two significant figures. The
numerical value of the measurement result should in the
final statement normally be rounded to the least
significant figure in the value of the expanded
uncertainty assigned to the measurement result. For the
process of rounding, the usual rules for rounding of
numbers have to be used (for further details on rounding
see ISO 31-0:1992, Annex B). However, if the rounding
brings the numerical value of the uncertainty of
measurement down by more than 5%, the rounded up
value should be used.
EA-4/02

145
 Reporting results
6.7 Uncertainties are usually expressed in bilateral terms (±) either in
units of the measurand or as relative values, for example as a
percentage (%), parts per million (ppm), 1 in 10X, etc. ……………

6.8 The number of figures in a reported uncertainty should always

reflect practical measurement capability. In view of the process for
estimating uncertainties it is seldom justified to report more than
two significant figures. It is therefore recommended that the
expanded uncertainty be rounded to two significant figures, using
the normal rules of rounding. The numerical value of the
measurement result should in the final statement normally be
rounded to the least significant figure in the value of the expanded
uncertainty assigned to the measurement result.

6.9 Rounding should always be carried out at the end of the process in
order to avoid the effects of cumulative rounding errors.
UKAS M3003

146
 Reporting result
1. Best estimate ± Uncertainty

Best estimate = average value = 2.4

probable range : 2.3 to 2.5

How to express ?

Measured value = 2.40.1

(measured value of x) = xbest ±x

147
 Reporting result
2. Significant Figures
(measured g) = 9.82 ± 0.02385 m/s2
4 figures

(measured g) = 9.82 ± 0.02 m/s2

Rule for stating Uncertainties

Experimental uncertainties should almost
always be rounded to one significant figure.

Exception
If the leading digit is 1 (one), two figures is better.

Not x=0.1 but x=0.14

148
 Reporting result
2. Significant Figures
(measured speed) = 6051.78 ± 30 m/s

>2 or <8 No meaning

(measured speed) = 6050 ± 30 m/s

Rule for stating Answers

The last significant figure in any stated answer
should usually be of the same order of magnitude
(in the same decimal position) as the uncertainty.

149
 Reporting result
2. Significant Figures
Examples : 92.81
uncertainty

0.3 : 92.8  0.3

3 : 93  3

30 : 90  30

150
 Reporting result
2. Significant Figures
Any numbers to be used in subsequent calculations should
normally retain at least one significant figure more than is
finally justified.

Unit should be placed after the uncertainty. 6050 ± 30 m/s

6050 m/s ± 30 m/s

In the scientific notation, the answer and the uncertainty

should be in the same form.
(1.61 ± 0.05) x 10-19 coulombs
1.61 x 10-19 ± 5 x 10-21 coulombs

151
 Reporting Conventions
• 1000 (30) mL
– Defines the result and the (combined) standard
uncertainty
• 1000 +/- 60 mL
– Defines the result and the expanded uncertainty
(k=2)
• 1000 +/- 60 mL at 95% confidence level.
– Defines the expanded uncertainty at the specified
confidence interval

152
Calibration of a Mercury Thermometer
 Calibration of a mercury thermometer
measurement set-up
• task of measurement
(schematic)
determination of the error of indication
of a mercury thermometer at 20°C.

• principle of measurement
temperature measurement in a medium
of known temperature.
tINDX tINDS
• method of measurement
comparison of two temperatures.
• procedure of measurement
comparison of the indication of the
thermometer to be calibrated (calibration
object) with the indication of a calibrated
thermometer (working standard) in a
stirred water bath.

154
 Calibration of a mercury thermometer

cause effect propagation

tX - error of indication of the thermometer to be calibrated at

the temperature of 20 °C,
tINDX - temperature indicated by the calibration object,
 tINDX - possible, but unknown deviation resulting from the finite
resolution of the indication of the calibration object,
tX - temperature in the stirred bath at the calibration object,

155
 Calibration of a mercury thermometer
tS - temperature in the stirred bath at the working standard,
 tBATH - possible, but unknown difference between the
temperature of the bath at the calibration object and the
temperature of the bath at the working standard,
tINDS - temperature indicated by the working standard,
 tINDS - possible, but unknown deviation resulting from the
finite resolution of the indication of the working standard,
tINDS - error of indication of the working standard at time of its
last calibration,
 tSD - possible, but unknown deviation of the error of
indication of the working standard from the value of its last
calibration due to drifts.

156
 Problem 6.2: Calibration of a mercury
thermometer (part 4)
Assessment of the knowledge about the input quantities

• tINDX : 9 observations of the temperature indicated by the calibration object

•  tINDX : scale division of the calibration object 0,1 K;
the resulting maximum deviation due to the finite resolution
is 0,05 K;
• tINDS : 10 observations of the temperature indicated by the working standard
•  tINDS : scale division of the working standard 0,05 K;
the resulting maximum deviation due to the finite resolution
is 0,025 K;
•  tBATH: the modulus of the temperature gradient in the stirred bath
is known from detailed investigations to be less than 0,5 Km-1;
the two thermometers are immersed at a distance of 60 mm;
the resulting maximum deviation is 30 mK;

158
 Calibration of a mercury thermometer

Assessment of the knowledge about the input quantities (continued)

• tINDS : the temperature assigned to the indication of 20,0°C

of the working standard is according to the calibration
certificate 19,96°C ± 30 mK (k = 2);
the resulting error of indication is 0,04°C ± 30 mK (k = 2);
•  tSD : it is known from the history of the working standard that
the value of the error of indication changes randomly
between two calibrations but not more than 40 mKa-1;
Since the working standard has been calibrated half
a year before the deviation due to drift is within ± 20 mK.

159
 Calibration of a mercury thermometer
Readings of the indicated temperatures
(because there is no pre-processing on the readings they equal the observations)
tINDX: tINDS:
number value
number value
1 20,05°C
1 19,8 °C
2 19,95°C
2 19,9 °C
3 20,05°C
3 19,9 °C
4 20,10°C
4 20,0 °C
5 20,00°C
5 19,8 °C
6 20,10°C
6 19,8 °C
7 19,9 °C
7 20,05°C
8 19,8 °C 8 19,95°C
9 19,9 °C 9 20,15°C
10 20,10°C

160
 Calibration of a mercury thermometer
quantity value standard distri- deg. of sensitivity- uncertainty
uncertainty bution freedom coefficient contribution
Xi xi u( xi ) i ci ui ( y )
tINDX
 tINDS
 tBATH
tINDS
 tINDS
tINDS
 tSD
tINDX
Result of measurement tX = ???
Expanded uncertainty of measurement (P = 0,95; k = ???) U = ???

161
 Calibration of a mercury thermometer
quantity value standard distri- deg. of sensitivity- uncertainty
uncertainty bution freedom coefficient contribution
Xi xi u( xi ) i ci ui ( y )
tINDX 19,867°C 23,6 mK normal 8 1,0 24 mK
 tINDX 0,0 K 28,9 mK rectang  1,0 29 mK
 tBATH 0,0 K 17,3 mK rectang  1.0 17 mK
tINDS 20,050°C 21,1 mK normal 9 1,0 21 mK
 tINDS 0,0 K 14,4 mK rectang  1,0 14 mK
tINDS 0,040 K 15,0 mK normal 50 1,0 15 mK
 tSD 0,0 K 11,5 mK rectang  1,0 12 mK
tINDX -0,143 K 52 mK 120

Result of measurement tX = -0,14 K

Expanded uncertainty of measurement (P = 0,95; k = 2) U = 0,10 K

162
Conclusions and Remarks
 Basic concepts
“…The evaluation of uncertainty is neither
a routine task nor a purely mathematical
one; it depends on detailed knowledge
of the nature of the measurand and of
measurement…”

GUM 3.4.8

164
Uncertainty Estimation in Practice

Martti Heinonen 30.5.2006

165
166
 REMEMBER!!!

„Live with Uncertainty‟

not
„Live in Uncertainty‟

Thank You!
167