Download as pdf or txt
Download as pdf or txt
You are on page 1of 43

Measurement Techniques

Lecture Notes
Significant Digits

Rounding Off

Importance of Error Band

Random and Systematic Errors

Quoting Results and Errors

Error Analysis

Uncertainty Analysis

Statistical Analysis
•Statistical Analysis of Experimental Data
•Probability Distributions
•Histograms

Gauss Distribution – Bell Curve


Significant Digits

 Engineering measurements are generally accurate to at most only a few digits. Three
(sometimes four) digits of accuracy are considered “standard” for engineering analysis.

 The number of significant digits is defined as the number of relevant or useful digits in a
measurement.

 The best way to illustrate is to write the number in standard exponential (scientific) notation
instead of common real number (engineering) notation, and then count the number of
digits.

 Here are some examples:

Common Underlined Exponential # significant


Comments
notation notation notation digits
134.2 134.2 1.342 × 102 4 just count the number of digits

0.0056 0.0056 5.6 × 10-3 2 the leading zeroes are not significant
the leading zeroes are not significant, but any zeroes
0.00506 0.00506 5.06 × 10-3 3
between two numbers are significant
the leading zeroes are not significant, but the trailing
0.00560 0.00560 5.60 × 10-3 3
zeroes are significant
Significant Digits

Common Underlined Exponential # significant


Comments
notation notation notation digits
integer values have an infinite number of
400 400 4 × 102 infinite
significant digits
a decimal point (or underline) indicates that all
400. 400 4.00 × 102 3 digits to the left of the decimal point are significant,
and that this is not an integer value
the zero to the right of the decimal point is
400.0 400.0 4.000 × 102 4
significant
a decimal point (or underline) indicates that all
40,300. 40,300 4.0300 × 104 5 digits to the left of the decimal point are significant,
and that this is not an integer value
integer values have an infinite number of
40,300 40,300 403 × 102 infinite significant digits; do not use a decimal point when
writing an integer in exponential notation
400 (to 2 words in parenthesis are necessary to indicate a
significant 400 4.0 × 102 2 smaller number of significant digits in common
digits) notation whenever trailing zeroes are present
Significant Digits

 Things become somewhat difficult and ambiguous when dealing with large numbers. For
example, suppose someone reports the population of a large city as 3,485,000, and says
nothing about significant digits.

 Is it rounded to the nearest thousand (3.485 × 106; 4 significant digits)?


 Is it rounded to the nearest hundred (3.4850 × 106; 5 significant digits)?
 It is impossible to know. We suspect that the population is not exactly 3,485,000 (7
significant digits), although that is a remote possibility.

 One way around this ambiguity is to underline the least significant digit.
 In our population example, if the population were rounded to the nearest thousand, we
would write 3,485,000 since the first zero is not significant.
 If the population were rounded to the nearest 100, we would write 3,485,000 since the first
zero is significant, and so on.
Significant digits
 When performing multiplication or division calculations, the answer has the same number of
significant digits as the component with the least number of significant digits.

 Example:
 Given: A force of 4.210 N is measured, and it is applied to a mass of 2.23 kg so as to
accelerate this mass.
 To do: Calculate the acceleration.
 Solution: Use Newton's second law, i.e. F = ma, and solve for the acceleration:

4.210 kg∙m/s
1.88789 m/s The answer we report is:
2.23 kg
 Note some important points in this simple example:
1.89 m/s

 The final answer is reported to only three significant digits, since m is precise to only 3
significant digits, and 3 < 4, where F is precise to 4 significant digits. The answer is not
1.88789238, even though that is what the calculator shows! All the digits following the first
three are meaningless.
 However, it is good to write down the answer to several additional significant digits, as
above, if this value is to be used in subsequent calculations. Failure to do so can result in
round-off error.
Significant digits
Scientific method is something really sweet
Collecting lots of data can make your life complete
Significant Digits – song – YouTube Analyzing data means you get to make a graph
Before you reach conclusions you might need to do some math!

http://youtu.be/kB2szfcwu1A You take the measured numbers, and substitute them in


Into a math equation, calculations can begin
You check your calculator and you write the answer down
But heres the great big question, where the heck to round?

When adding or subtracting numbers right there in your face


Look for the one thats least precise and round it to that place
If you have some measurements to multiply or divide
Round to the fewest sig figs counting from the leftward side

If you add two numbers up, dont sit there looking tense
Lets say ones to the hundredths place, the other to the tenths
3.25 plus 3.3 gives us 6.55
Now its time to round it off, and still get out alive!

3.25s measurement precise to hundredths place


But 3.3 is to the tenths, thats less by one space
Round your answer to the place that goes the least far out
To the tenths, thats 6.6, thats what its all about!

When adding or subtracting numbers right there in your face


Look for the one thats least precise and round it to that place
If you have some measurements to multiply or divide
Round to the fewest sig figs counting from the leftward side

Now you have two numbers that you need to multiply


3.5 X 9.95, watch your fingers fly
Across your calculator, here comes the answer, live!
What flashes quick upon the screen? 38.825

3.5 has two sig figs, 9.95 has three


So you must round to two sig figs, because its less, you see
So you count in two sig figs from the leftwards side
Your answer rounds to 39, this roundings quite a ride!

When adding or subtracting numbers right there in your face


Look for the one thats least precise and round it to that place
If you have some measurements to multiply or divide
Round to the fewest sig figs counting from the leftward side
Rounding Off

 There are standard rules for rounding off values to a desired number of significant digits.
First, the number is truncated to its desired length. Then, the excess (leftover) digits are
examined as if they were a decimal fraction:

 If the decimal fraction is less than 0.5, truncate the excess digits.
 If the decimal fraction is greater than 0.5, round up the least significant digit in the
number by one.
 If the decimal fraction is exactly 0.5, the convention is to round up if the least significant
digit is odd, and to truncate (round down) if the least significant digit is even. [Note: The
digit zero is even.]
 Here is a little trick to remember this, courtesy of Steve Galamba: Either way, you get an
even digit!

 Examples:
 Round 548,392 to three significant digits. Answer: 548,000 [round down].
 Round 548,592 to three significant digits. Answer: 549,000 [round up].
 Round 548,500 to three significant digits. Answer: 548,000 [round down since 8 is even].
 Round 547,500 to three significant digits. Answer: 548,000 [round up since 7 is odd].
Rounding Off

 Things become a little tricky when adding or subtracting numbers. For example, suppose the population
of a large city is 3,485,000, rounded to the nearest thousand (3.485 × 106 or 3,485,000; 4 significant
digits). If 12 people move into the city, what is the new population?

 We are tempted to say 3,485,000 + 12 = 3,485,012, but this implies 7 significant digits of precision.
Actually, since we cannot have fractions of people, it implies infinite precision, i.e., an integer
number. The correct answer is 3,485,000, since the 12 extra people do not change the population
to the nearest thousand – we round off.

 Suppose 1,862 people move to the city. The new population is 3,485,000 + 1,862 = 3,486,862
rounded off to the nearest thousand to maintain the original number of significant digits, i.e.,
3,487,000.

 When performing addition or subtraction, the number of significant digits is determined by the leftmost
decimal column that contains a least significant digit. The best way to add or subtract numbers is to align
the decimal point, and highlight the leftmost significant digit. For example, here is how to add 13.68 +
0.08672:
13.68
+ 0.08672
13.76672

 We highlight the fourth column from the left, since it contains the leftmost least significant digit. After
rounding up, our final answer is therefore 13.77, precise to four significant digits.
Rounding Off

 When adding or subtracting numbers, it is possible for the result to have a greater number of
significant digits than any of the component numbers. For example, 5.86 + 7.21 = 13.07 [result
precise to four significant digits]. This is useful when calculating mean (average) values of a
measurement sample.

 Example: Consider the following twelve numbers:

 7.53, 8.76, 7.42, 8.15, 7.79, 7.88, 7.91, 8.24, 8.13, 7.74, 7.80, and 8.06.

 The average is calculated by adding up all the numbers and dividing by 12. The sum is 95.41
(four significant digits). Calculation of the average yields 95.41/12 = 7.950833333.

 However, we must round to a maximum of four significant digits because of the division. Here,
although each data point is precise to only 3 significant digits, our final answer is 7.951
(rounded to four significant digits). However, as mentioned previously, if we need to use this
average in further calculations, we should carry along a few more digits to avoid potential
round-off errors.

 Note: Some argue that an average cannot have more significant digits than its components. They
would say that the correct final answer to the above example is 7.95 (rounded to three significant
digits). There is some support for this alternative answer here since there is so much scatter in
the original data.
Importance of Error Band

Why are errors important?

 For an infant, two measurements of body temperature before and after a drug is
administered:

 38.4°C and 38.2°C

 Is the temperature drop significant? Or in other words: does the drug have an
effect on the change in the temperature?

 It depends on the associated errors.

 (38.2  0.01)C and (38.4  0.01)C significant, the drug possibly had
an effect on the temperature drop

 (38.2  0.5)C and (38.4  0.5)C not significant, it is not possible to


assume that the drug had an
effect on the temperature drop
Random and Systematic Errors

 Random and systematic errors

RANDOM errors only

True Value

Random + Systematic Errors

 A result is said to be accurate if it is relatively free from systematic error.

 A result is said to be precise if the random error is small.


Quoting Results and Errors

 Generally state error to one or two significant figures

 Quote result to same significance as error

 When using scientific notation, quote value and error with the same exponent

 Examples:
 Value 44, error 5  44  5
 Value 128, error 32  130  30
 Value 4.8×10-3, error 7×10-4  (4.8  0.7)×10-3
 Value 1092, error 56  1090  60
 Value 1092, error 14  1092  14
 Value 12.345, error 0.35  12.3  0.4

Don’t over quote results to a level inconsistent with the error


36.678935372  0.5 
Error Analysis

 We are interested in knowing the uncertainty in the final result due to the uncertainties in
the primary measurements.

 Error Analysis on a Commonsense Basis:


Commonsense analysis would combine all the errors in the most detrimental way in
order to determine the maximum error in the final result.

 Consider the calculation of electric power from: P = E I

 where E and I are measured as:


E = 100 V ± 2 V
I = 10 A ± 0.2 A

 The nominal value of power is 100 × 10 = 1000 W. By taking the worst variations in voltage
and current, we could calculate:

Pmaks = (100 + 2) × (10 + 0.2) = 1040.4 W


Pmin = (100 - 2) × (10 - 0.2) = 960.4 W

 Thus, using this method of calculation, the uncertainty in the power is +4.04%, -3.96%.
Error Analysis

 When the plus or minus sign notation is used to designate the uncertainty, the person making
this designation is stating the degree of accuracy with which he or she believes the
measurement has been made. We may note that this specification is in itself uncertain because
the experimenter is naturally uncertain about the accuracy of these measurements.

 To add a further specification of the uncertainty of a particular measurement, the experimenter


should specify certain odds for the uncertainty.

Example: p = 100 kPa ±1 kPa (20:1)

In other words, the experimenter is willing to bet with 20 to 1 odds that the pressure
measurement is within ±1 kPa.

 We wish to estimate the uncertainty in the calculated result on the basis of the uncertainties in
the primary measurements. The result R is a given function of the independent variables x1, x2,
x3, ..., xn. Thus,
R = R (x1, x2, x3, ..., xn)

 Let wR be the uncertainty in the result and w1, w2, ..., wn be the uncertainties in the independent
variables. If the uncertainties in the independent variables are all given with the same odds, then
the uncertainty in the result having these odds is given as:
Uncertainty Analysis

2 1/ 2
 R   R 
2 2
 R  
wR   w1    w2      wn  
 x1   x2   xn  

EXAMPLE:
The resistance of a certain size of copper wire is given as:

R  R0 1   (T  20)

where R0 = 6 Ω ± 0.3% is the resistance at 20 C, α = 0.004 C-1 ± 1% is the


temperature coefficient of resistance, and the temperature of the wire is T = 30 ± 1C.
Calculate the resistance of the wire and its uncertainty.
Uncertainty Analysis

SOLUTION:
The nominal resistance is:
R  6 1  (0.004)(30  20)  6.24 

R
 1   (T  20)  1  (0.004)(30  20)  1.04 wR0  (6)(0.003)  0.018 
R0
R
 R0 (T  20)  (6)(30  20)  60 w  (0.004)(0.01)  4 10 5 C 1

R
 R0  (6)(0.004)  0.024 wT  1 C
T
Thus, the uncertainty in the resistance is:


wR  (1.04) 2 (0.018) 2  (60) 2 (4 10 5 ) 2  (0.024) 2 (1) 2  1/ 2

wR  0.0304  or 0.49%
Uncertainty Analysis

IMPORTANT REMARKS:

• Particular notice should be given to the fact that the uncertainty


propagation in the result wR depends on the squares of the uncertainties in
the independent variables wi.

• This means that if the uncertainty in one variable is significantly larger than
the uncertainties in the other variables, say by a factor of 5 or 10, then it is
the largest uncertainty that predominates and the others may probably be
neglected.

• It is just as unfortunate to overestimate uncertainty as to underestimate it.

• An underestimate gives false security, while an overestimate may make


one discard important results, miss a real effect, or buy much too
expensive instruments.
Uncertainty Analysis
EVALUATION OF UNCERTAINTIES FOR COMPLICATED DATA REDUCTION

In many cases data reduction is a rather complicated affair and is often performed with a
computer routine written specifically for the task.

Suppose a set of data is collected in the variables x1, x2, ..., xn and a result is calculated. At the
same time one may perturb the variables by ∆x1, ∆x2, etc., and calculate new results. We would
have
R( x2 )  R ( x1 , x2 ,, xn )
R ( x2  x2 )  R ( x1 , x2  x2 ,, xn )

For small enough values of ∆x the partial derivatives can be approximated by


R R( x1  x1 )  R( x1 )

x1 x1
R R( x2  x2 )  R ( x2 )

x2 x2
and these values could be inserted in the equation to calculate the uncertainty in the result.
Uncertainty Analysis
EVALUATION OF UNCERTAINTIES FOR COMPLICATED DATA REDUCTION
EXAMPLE:
Calculate the uncertainty of the wire resistance in the previous example using the result-
perturbation technique.
Solution: In the previous example, we have already calculated the nominal resistance as
6.24 Ω. We now perturb the three variables R0, α, and T by small amounts to evaluate the
partial derivatives. We shall take R  0.01
0   110 5 T  0.1
Then R ( R0  R0 )  (6.01)1  (0.004)(30  20)  6.2504
R R ( R0  R0 )  R 6.2504  6.24
and the derivative is approximated as    1.04
R0 R0 0.01
Similarly,
R (   )  (6.0)1  (0.00401)(30  20)  6.2406
R R (   )  R 6.2406  6.24
   60
  110 5

R (T  T )  (6.0)1  (0.004)(30.1  20)  6.2424


R R(T  T )  R 6.2424  6.24
   0.024
T T 0.1
Uncertainty Analysis
EVALUATION OF UNCERTAINTIES FOR COMPLICATED DATA REDUCTION
EXAMPLE:
Calculate the uncertainty of the wire resistance in the previous example using the result-
perturbation technique.
R0  0.01   110 5
T  0.1
Solution:
Previous solution results:

R ( R0  R0 )  6.2504 R R
 1.04  1.04 wR0  0.018 
R0 R0
R R
R (   )  6.2406  60  60 w  4 10 5 C 1
 
R R
R (T  T )  6.2424  0.024  0.024 wT  1 C
T T

All the derivatives are the same as in the previous example, so the uncertainty in R would be
the same, or 0.0304 Ω.
Uncertainty Analysis

REMARKS:

 For a measurement, the uncertainty of an instrument is generally given by


the manufacturer.

 This value is mostly presented as the percentage of the full scale.

 For example,

If the accuracy is said to be ±1.0 kPa for a pressure transducer; this


instrument with a measurement range of 0-100 kPa will have 10%
uncertainty when measuring a pressure of 10 kPa.
Statistical Analysis

In the nineteenth century, Gauss' assistants were doing astronomical


measurements. However, they were never able to exactly repeat their results.
Finally, Gauss became angry and stormed into the lab, claiming he would
show these people how to do the measurements once and for all. The only
problem was that Gauss wasn't able to repeat his measurements exactly
either!

After he recovered his composure, Gauss made a histogram of the results of


a particular measurement and discovered the famous Gaussian or bell-
shaped curve.

Many people's first introduction to this shape is the grade distribution for a
course.

 Gauss Distribution – demo – YouTube

http://www.youtube.com/watch?v=xDIyAOBa_yU&NR=1
Statistical Analysis
Statistical Analysis
Statistical Analysis

STATISTICAL ANALYSIS OF EXPERIMENTAL DATA

First, it is important to define some pertinent terms.


1 n
The arithmetic mean: xm   xi
n i1
The deviation di for each reading is defined by

di  xi  xm
We may note that the average of the deviations of all readings is zero since::

1 n 1 n
di   di   ( xi  xm )
n i 1 n i 1
1
 xm  (nxm )  0
n
Statistical Analysis

The average of the absolute values of the deviations is given by

1 n 1 n
di   di   xi  xm p(x)
n i1 n i1
Note that this quantity is not necessarily zero.

The standard deviation or root-mean-square


deviation is defined by
1/ 2
1 n

    ( xi  xm ) 2  Large or small deviations from the mean
 n i 1 
This is sometimes called the population or biased standard
deviation because it is strictly applies only when a large 1/ 2
 n 2
 i m 
number of samples is taken to describe the population.
( x  x )
For small sets of data (n < 20) an unbiased or sample    i1 
standard deviation is defined by  n 1 
 
Statistical Analysis

The square of the standard deviation, σ2 is called the variance.

There are other kinds of mean values of interest from time to time in statistical analysis.
The median is the value that divides the data points in half.

For example, if measurements made on five production resistors give 10, 12, 13, 14, and
15 kΩ, the median value would be 13 kΩ. The arithmetic mean, however, would be

10  12  13  14  15
Rm   12.8 
5
Statistical Analysis

EXAMPLE:
The following readings are taken of a certain physical length.

Compute the mean reading, standard deviation, variance, and average of the absolute
value of the deviation, using the “biased” basis.

Calculate also the best estimate of standard deviation for the same data based on the
“sample” or unbiased basis.
reading x, cm
1 5.30
2 5.73
3 6.77
4 5.26
5 4.33
6 5.45
7 6.09
8 5.64
9 5.81
10 5.75
Statistical Analysis
SOLUTION:
1 n 1
The arithmetic mean xm   xi  (56.13)  5.613 cm
n i 1 10
1/ 2 1/ 2
1 n
 1 
    ( xi  xm ) 2    (3.533)  0.5944 cm
 n i 1  10 

 2  0.3533 cm 2 reading x, cm di = xi-xm (xi-xm)2×102 |di|


1 5.30 -0.313 9.797 0.313
2 5.73 0.117 1.369 0.117
1 n 1 n 3 6.77 1.157 133.865 1.157
di   di   xi  xm 4 5.26 -0.353 12.461 0.353
n i 1 n i 1 5 4.33 -1.283 164.609 1.283
6 5.45 -0.163 2.657 0.163
1
 (4.224)  0.4224 cm 7 6.09 0.477 22.753 0.477
10 8 5.64 0.027 0.073 0.027
9 5.81 0.197 3.881 0.197
10 5.75 0.137 1.877 0.137

Sum = 56.13 3.533 4.224


Statistical Analysis

reading x, cm di = xi-xm (xi-xm)2×102 |di|


1 5.30 -0.313 9.797 0.313
2 5.73 0.117 1.369 0.117
3 6.77 1.157 133.865 1.157
4 5.26 -0.353 12.461 0.353
5 4.33 -1.283 164.609 1.283
6 5.45 -0.163 2.657 0.163
7 6.09 0.477 22.753 0.477
8 5.64 0.027 0.073 0.027
9 5.81 0.197 3.881 0.197
10 5.75 0.137 1.877 0.137

Sum = 56.13 3.533 4.224


Average = 5.613
1/ 2
σ= 0.627  1 
  (3.533)  (0.3929)1/ 2  0.627 cm
10  1 
The resulting standard deviation value will be different when unbiased basis for small sets
of data is adopted.
Statistical Analysis
PROBABILITY DISTRIBUTIONS
Suppose we toss a horseshoe some
distance x. Even though we make an
effort to toss the horseshoe the same
distance each time, we would not
xm always meet with success.
x
p(x) The better the player is at the game, the
more closely the throws will be grouped
around the mean.
p(x)

x x+x xm x
Also, we would expect fewer and fewer tosses for those x distances which are farther and
farther away from the target. For a good player the maximum probability is expected to
surround the distance xm designating the position of the target.
The curve shown is called a probability distribution. Each value of the ordinate p(x) gives
the probability that the horseshoe will land between x and x+∆x, where ∆x is allowed to
approach zero.
Statistical Analysis

HISTOGRAMS
Distance from Number of
Target, cm Throws
We have noted that a probability distribution is
obtained when we observe frequency of occurrence 0-10 5
over a large number of observations. When a 10-20 15
limited number of observations is made and the raw 20-30 13
data is plotted, we call the plot a histogram. 30-40 11
40-50 9
EXAMPLE: 50-60 8
60-70 10
These data are plotted using increments of either 70-80 6
10cm or 20cm in ∆x and the plots are shown on the 80-90 7
next page. A cumulative frequency diagram could 90-100 5
also be employed for these data. The resulting 100-110 5
curve is given on the following page. 110-120 3
>120 2

Total 99
Statistical Analysis

30

20

10
Number of throws

0
0-20 20-40 40-60 60-80 80-100 100-120 >120

Distance from target, cm


Histogram with ∆x = 20cm
Statistical Analysis

15

10

5
Number of throws

Distance from target, cm


Histogram with ∆x = 10cm
Statistical Analysis

1.00
0.90
0.80
0.70
0.60
0.50
Fraction of throws less than x

0.40
0.30
0.20
0.10
0.00
0 20 40 60 80 100 120 140

Distance from target, cm


Cumulative frequency diagram
Gauss Distribution – Bell Curve
THE GAUSSIAN OR NORMAL ERROR DISTRIBUTION
Assuming that there are many small errors that contribute to the final error and that each
small error is of equal magnitude and equally likely to be positive and negative, the
gaussian or normal error distribution may be derived. If the measurement is designated
by x, the gaussian distribution gives the probability that the measurement will lie between
x and x + dx and is written
1  ( x  xm ) 2 / 2 2
p( x)  e
 2
In this expression xm is the mean reading and
σ is the standard deviation. Some may prefer
to call P(x) the probability density. The
standard deviation is a measure of the width
of the distribution curve; the larger the value
of σ, the flatter the curve and hence the larger
the expected error of all measurements. The
equation is normalized so that the total area
under the curve is unity. Thus,


 p( x)dx  1.0

Gauss Distribution – Bell Curve

At this point we may note the similarity between the shape of the normal error curve and
the experimental distribution for tossing horseshoes. The better the player is at the game,
the more closely the throws will be grouped around the mean and the more probable will
be the distance xm.
Here are some of the properties of
this special distribution:

 The mean is at the peak of the


distribution. p(x)

 The plot is commonly called a


“bell curve” because of its shape.

 It is symmetric about the mean.

 The actual shape depends on the


magnitude of the standard
deviation. Namely, if the standard
deviation is small, the bell will be
tall and skinny, while if the
standard deviation is large, the
bell will be short and fat, as
sketched.
Gauss Distribution – Bell Curve

We may quickly anticipate the next step in the analysis as one of trying to determine the
precision of a set of experimental measurements through an application of normal error
distribution. For sets of data where a large number of measurements is taken,
experiments indicate that the measurements do indeed follow normal error distribution
when the experiment is under control.

1  ( x  xm ) 2 / 2 2
p( x)  e
 2
By inspection of the gaussian distribution function, we see that the maximum probability
occurs at x = xm, and the value of this probability is

1
p ( xm ) 
 2
Gauss Distribution – Bell Curve

The probability that a measurement will fall within a certain range x1 of the mean reading is

xm  x1
1
P( x)   e  ( x  xm ) 2 / 2 2
dx
xm  x1  2

Making the variable substitution

1
x  xm 1
 e
 2 / 2
 P ( )  d
 2  1

Since the distribution is symmetric around η = 0, it may be observed that


1 1
1 1
 e e
2
  2 / 2
P ( )  /2
d  2 d
2  1
2 0
Gauss Distribution – Bell Curve

TABLE-2 Integrals of the Gaussian normal error function

1
1 P
e
 2 / 2
Values of the integral d  are given for different values of the
2 0
2
argument η1. The values are related to error function since
1
1
 e
 2
erf 1  d
  1

1
So that the tabular values are equal to erf(  1 2 ) . Each figure in the body of
2
the table is preceded by a decimal point.
Gauss Distribution – Bell Curve

TABLE-2 Integrals of the Gaussian normal error function


Gauss Distribution – Bell Curve

EXAMPLE:
Calculate the probabilities that a measurement will fall within one, two and three
standard deviations of the mean value.

SOLUTION:
We perform the calculation with η = 1, 2, and 3. The values of the integral may be
obtained from Table-2. We observe that
1 1

 e d  2  e
2
 / 2  2 / 2
d
 1 0

so that
P (1)  2 0.34134   0.6827
P (2)  2 0.47725  0.9545
P (3)  2 0.49865  0.9973
Gauss Distribution – Bell Curve

P (1)  2 0.34134   0.6827


P (2)  2 0.47725  0.9545
P (3)  2 0.49865  0.9973
Therefore we can state that

The probability that a measurement lies within  one standard deviation from the mean is 68%
The probability that a measurement lies within  two standard deviation from the mean is 95%
The probability that a measurement lies within  three standard deviation from the mean is 99%

Confidence level

For the above case, the confidence level is 68%. We are 68.26% confident that any random
measurement of x will lie within  one standard deviation from the mean value.

You might also like