Chem 26: Analytical
Chemistry
Chapter 2
Errors and Evaluating of
Analytical Data
Department of Chemistry
College of Mathematics and Natural Sciences
Caraga State University
Lecture Outline:
A. Types of Errors: Systematic random and Gross Errors
B. Determining Errors
1. Mean and Median
2. Accuracy and Precision
ü Absolute error, Relative error, Standard
Deviation, Variance and Relative Standard
Deviation
Intended Learning Outcomes:
Perform data analysis using statistical method.
Show the difference of the three types of error.
Illustrate the ideas of the types of error.
Compare the difference between accuracy and
precision.
Evaluate by calculating the mean and median of
given samples.
Solve problems relating to precision and accuracy.
Types of Error
No analysis is free of error or “uncertainty”
1. Systematic Error (determinate error)
à the error is reproducible and can be
discovered and corrected.
àaffect the accuracy of results.
à caused by the wrong use of
instruments.
à lead to bias in measurement.
2. Random Error (indeterminate error)
à caused by uncontrolled variables, à data are tending to be scattered
which cannot be defined or eliminated. more or less symmetrically around a
à affect measurement precision. mean value.
4
Systematic (determinate) Errors
1. Instrument errors
Failure to calibrate, degradation of parts in the instrument, power
fluctuation, variation in temperature, etc.
Can be corrected by calibration or proper instrumentation maintenance
2. Method errors
- Errors due to no ideal physical or chemical behavior
-Completeness and speed of reaction, interfering side reactions,
sampling problems
Can be corrected with proper method development
3. Personal errors
Occur where measurements require judgement, result from prejudice,
color acuity problems
Can be minimized or eliminated with proper training and experience
5
Systematic (determinate) Errors
Detection of Systematic Errors
1. Analysis of standard samples
2. Independent analysis : Analysis using a reference
method or reference laboratory
3. Blank determination
4. Variation in sample size : detects constant error only
6
Random (indeterminate) Errors
Ø No identifiable cause: Always present, cannot be
eliminated; the ultimate limitation on the determination
of a quantity (sources of uncertainty).
Example: reading a scale on an instrument caused by the finite
thickness of the lines on the scale; electrical noise
Ø The accumulated effect causes replicate measurements
to fluctuate randomly around the mean; Give rise to a
normal or Gaussian curve; Can be evaluated using
statistics
7
Random (indeterminate) Errors
Sources of Random (indeterminate)
Errors
• Small, undetectable uncertainties produce a detectable
random error
Example
– Four small random errors combine to give an overall error
– Each error has an equal probability of occurring, and that
each can cause the final result to be high or low by a fixed
amount ±U
8
How to determine Error?
Accuracy – closeness of measurement to its
true or accepted value.
Systematic or determinate errors affect accuracy
Precision – agreement between 2 or more
measurements of the sample made in
exactly the same way
Random or indeterminate errors affect precision.
9
How to determine Error?
10
How to determine Error?
Mean – average or arithmetic mean.
Median – arrange results in increasing or
decreasing order.
Precision – S, RSD, CV
Accuracy – Absolute Error, Relative error
11
How to determine Error?
å
n
v Mean i =1
X
è the mean, is the numerical average X =
obtained by dividing the sum of the n
where Xi is the ith measurement,
individual measurements by the number
and n is the number of independent
of measurements. measurements.
è The most widely used measure of
central value.
v Median
è the median, Xmed, is the middle value è When the data include an odd
when data are ordered from the number of measurements, the
smallest to the largest value. median is the middle value. For an
even number, the average value of
è Provides a more robust estimate of
the middle pair is used.
central tendency since it is less
sensitive to measurements with
extreme values.
12
How to determine Error?
v Accuracy
Absolute error (E) – diff. between true and measured value.
E = xi - xt
Where xi = experimental value and xt = true value
Example:
xi = 19.78 ppm Fe and xt = 20.00 ppm Fe
E = 19.78 – 20.00 ppm = -0.22 ppm Fe
(-) value too low; (+) value too high
Relative error (Er) – is often a more useful quantity than the
absolute error.
13
Example
Consider the replicate measurement of Iron
(Fe) concentration: 19.4, 19.5, 19.6, 19.8,
20.1, and 20.3 ppm Fe. The accepted value
Answer:
is 20.0 ppm Fe. Calculate:
1. -1% and -10 in ppt
2. E(19.4 ppm) = -0.6
1. The relative error in percent and in part
E(20.3 ppm) = 0.3
per thousand.
2. The E at 19.4 and 20.3 ppm Fe
A method of analysis yields masses of gold
that are low by 0.4 mg. Calculate the Answer:
percent relative error caused by this result if (a)-0.08%
the mass of gold in the sample is (a.) 500 (b)(b.) -0.16%
mg (b.) 250 mg
14
How to determine Error?
v Precision
l Describes the reproducibility of measurements—in other
words, the closeness of results that have been obtained
in exactly the same way.
l standard deviation (s), variance (s2), and
coefficient of variation (RSD). These three are
functions of how much an individual result xi differs from
the mean, called the deviation from the mean di.
15
How to determine Error?
v Precision
Standard Deviation (s) for small data set: Standard error
of the mean
Standard Deviation (σ) of population: for finite/large set of data
Where µ is the mean or average of the population
16
How to determine Error?
v Precision
Variance (s2)
The square of the standard deviation.
Relative Standard Deviation (RSD)
The standard deviation expressed as a percentage of the mean value.
s
RSD (st ) =
X
s
RSD (st ) in % = ´ 100
X
s
RSD (st ) in parts per thousand (ppt) = ´ 1000
X
17
Example
18
Example
Richards and Willard determined the molar mass of lithium
and collected the following data:
Find its mean and median and calculate its precision (s, sm, s2
and RSD) and accuracy (E and %Er).
19
Improving Data Precision
• Averaging results
– Improvement gained by averaging is limited
– To increase precision by a factor of 10 times
requires 100 times as many measurements
• Decrease s, since sm is directly proportional to s:
– More precision in individual operations
– Change procedure
– Use more precise measurement tools
20
Improving Data Precision
• Several subsets of data can be pooled (combined) to
get a better estimate of the population standard
deviation than using only one data set
• Assumption: same sources of random error in all
measurements
– Samples have similar compositions
– Samples have been analyzed in exactly the same
way
– Samples are randomly drawn from the same
population and thus have a common value of s
21
Improving Data Precision
Pooling Data to Improve the Reliability of s
The pooled estimate of s, which we call spooled, is a weighted
average of the individual estimates.
)!
)" *
* )
" 𝑥& − 𝑥( +& 𝑥+ − 𝑥* + ∑,'(
#
𝑥, − 𝑥- * +. . . . .
+'(
&'(
S!""#$% =
𝑁( + 𝑁* + 𝑁- +. . . . . . . −𝑁.
Where N1 is the number of results in Set 1, N2 is the number in Set 2, and so forth. The term Nt is the total
number of data sets that are pooled.
Example: Sample 1 Sample 2 Sample 3
Determine the spooled. % Heroin 2.24 1.07 4.30
2.27 1.02 4.20
1.03 Answer: 0.0415
22
END OF
CHAPTER
