ANALYTICAL CHEMISTRY
_______________________
LESSON 3
Errors, Random Errors,
and Statistical Data in Chemical Analysis
INTRODUCTION
A student analyzed six equal portions of an aqueous
solution known to contain exactly 20.00 ppm of Iron (III)
in exactly the same way
Note that the results range
from a low of 19.4 ppm to a
high of 20.3 ppm of Iron (III).
The average or mean = 19.78
ppm
DEFINING TERMS
REPLICATES MEAN, ARITHMETIC MEAN, AND
• are samples of about the same size that AVERAGE (ഥ
𝒙)
are carried through an analysis in exactly
the same way • are synonyms for the quantity
obtained by dividing the sum of
• Chemists usually carry two to five
portions (replicates) of a sample through replicate measurements by the
an entire analytical procedure. number of measurements in the set:
MEDIAN
• is the middle result when replicate data
are arranged in order of size.
• For an odd number of data points, the Where 𝑥𝑖 represents the individual values
median can be evaluated directly. of x making up a set of N replicate
• For an even number the mean of the measurements.
middle pair is used.
DEFINING TERMS
Example 1: Calculate the mean and the median for the data shown in Figure 3-1
19.4 + 19.5 + 19.6 + 19.8 + 20.1 + 20.3
mean = 𝑥ҧ =
6
= 19.78 ~19.8 𝑝𝑝𝑚 𝐹𝑒
Because the set contains an even number of measurements, the median is the average of
the central pair:
19.6 + 19.8
𝑚𝑒𝑑𝑖𝑎𝑛 = 2
= 19.7 𝑝𝑝𝑚 𝐹𝑒
DEFINING TERMS
PRECISION
ACCURACY
• describes the reproducibility of measurements
• indicates the closeness of the
measurement to its true or
• the closeness of results that have been obtained accepted value and is expressed
in exactly the same way. by the error.
• Three terms are widely used to describe the
precision of a set of replicate data:
1. standard deviation,
2. variance, and
3. coefficient of variation
• All these terms are a function of the deviation
from the mean di . or just the deviation, which is
defined as:
Accuracy vs Precision
ACCURACY
• measures agreement between
a result and its true value.
• can never determine exactly
because the true value of a
measured quantity can never
PRECISION
be known exactly.
• must use accepted value
instead. • describes the agreement among
several results that have been
• expressed in terms of either
obtained in the way.
absolute or relative error
• May be determined by
replicating, or repeating, a
measurement.
ERROR The absolute error, E is given by the equation:
𝑬 = 𝒙𝒊 − 𝒙𝒕
… in mathematics
Where:
It means the magnitude of a 𝑥𝑡 𝑖𝑠 𝑡ℎ𝑒 𝑡𝑟𝑢𝑒 value of accepted value.
Absolute number ignoring its sign.
Error
… in chemistry
It is the difference between
an experimental result and
an accepted or true value
including its sign
→ If the measurement value is low,
the sign is negative; 𝑥𝑡 = 20.0 𝑝𝑝𝑚 Fe
→ if the measurement result is @ 𝑥𝑖 = 19.8 ppm, E = 19.80 – 20.00 = - 0.20 ppm Fe
high, the sign is positive @ 𝑥𝑖 = 20.1 ppm, E = 20.10 – 20.00 = + 0.10 ppm Fe
Note: retain the sign in stating the error
Your turn!!
Calculate the absolute error
Table 5-1
Results from six replicate determination of iron in aqueous
samples of a standard solution containing 20.0 ppm iron (III)
xi E
19.40
19.50
19.60
19.80
20.10
20.30
ERROR
The relative error of a
Relative measurement is the absolute
Error value divided by the true
value
→ May be expressed as 𝑥𝑡 = 20.0 𝑝𝑝𝑚 Fe
percent, ppm, ppt @ 𝑥𝑖 = 19.8 ppm
𝟏𝟗.𝟖−𝟐𝟎.𝟎𝟎
𝑬𝒓 = 𝟐𝟎.𝟎𝟎
x 100% = -1%
The percent relative error, 𝑬𝒓 is given by the
equation:
𝒙 𝒊 − 𝒙𝒕 𝟏𝟗.𝟖−𝟐𝟎.𝟎𝟎
𝑬𝒓 = 𝒙𝒕
x 100% 𝑬𝒓 = x 1000 = - 10 ppt
𝟐𝟎.𝟎𝟎
If expressed in ppt;
𝒙 𝒊 − 𝒙𝒕
𝑬𝒓 = 𝒙𝒕
x 1,000
Classification of Experimental Errors
This figure summarizes the
results for the determination of
nitrogen in two pure
compounds:
• Benzyl isothiourea
hydrochloride
• Nicotinic acid
The dots show the absolute
errors of replicate results
obtained by four analysts
Classification of Experimental Errors
Analysts 1:
High precision
High accuracy
Analysts 2:
Poor precision
Good accuracy
Analysts 3:
High precision
Poor accuracy
Analysts 4:
Poor precision
Poor accuracy
Classification of Experimental Errors
1. RANDOM ERROR 2. SYSTEMATIC ERROR 3. GROSS ERROR
• Also known as • Also known as determinate • Usually occur only
indeterminate error error occasionally
• It affects the precision • It affects the accuracy of • Often very large
of measurement the results
• May cause the
• It causes data to be • It causes the mean of a set of result to be either
scattered data to differ from the very high or very
accepted value. low
• Causes data to be less
symmetrically around a • It causes the result in a series • Leads to outliers
mean value of replicate measurement to that obviously
be all high and all low differs significantly
from the rest of
the result
Classification of Experimental Errors
Compare the random errors and systematic errors of the analysts
The results of
Random error analysts 1 and 2
for analysts 1 have little
and 3 is systematic
significantly less error.
than analysts 2
and 4 Analysts 3 and
4 show
systematic
errors of about
-0.7 and -1.2%
nitrogen
SYSTEMATIC ERRORS
• Have a definite value BIAS
• Have assignable cause • it measures the systematic error
• About the same magnitude for replicate associated with an analysis
measurement made in the same way • It has a negative sign if it causes the
• Leads to BIAS in measurement technique results to be low and positive if it causes
results to be high
Types of systematic error
Instrument error Method error Personal error
Imperfections Arise from Result from the
in measuring nonideal carelessness,
chemical or inattention, or
devices and
physical personal
instabilities in limitations of
behavior of
their analytical the
components systems experimenter
RANDOM Errors
It causes the data
from a set of replicate
measurement to
fluctuate randomly
around the mean of
the set.
Table 3.1 shows
possible ways the four
errors can combine to
give the indicated
deviations from the
mean value.
This ratio of 1:4:6:4:1
is a measure of the
probability for a
deviation of each
magnitude.
Figure 3-4b shows the
theoretical distribution for
ten equal-sized
uncertainties.
Again we see that the most
frequent occurrence is zero
deviation from the mean
if we make a sufficiently When the same procedure is
large number of applied to a very large
measurements, we can number of individual errors, a
expect a frequency bell-shaped curve like that
distribution like that shown in shown in Figure 3-4c results.
Figure 3-4a.
Such a plot is called a
Gaussian curve or a normal
error curve.
TREATING RANDOM ERRORS WITH STATISTICS
I. Sample Mean and Population Mean
POPULATION MEAN, 𝝁
ഥ
SAMPLE MEAN, 𝒙 • The true mean for the population
• The mean of a limited sample drawn • In the absence of any systematic error,
from a population data the population mean is also the true
value for the measured quantity
σ𝑁
𝑖 𝑥𝑖
ഥ =
𝒙
𝑁 σ𝑁
𝑖=1 𝑥𝑖
𝝁=
𝑁
TREATING RANDOM ERRORS WITH STATISTICS
II. Standard deviation
SAMPLE STANDARD DEVIATION, s POPULATION STANDARD DEVIATION, 𝝈
Thank You!!!
