Scribd
Upload
42 views

2 - Errors and Statistical Data in Chemical Analysis

This document discusses the importance of statistics in analytical chemistry for communicating experimental data, analyzing finite data sets, and designing optimal experiments. It defines key statistical terms like mean, median, population, sample, accuracy, and precision. Accuracy refers to the agreement with the true value while precision refers to the reproducibility of results. Systematic errors are flaws in design or equipment while random errors are uncontrolled variables. Standard deviation, confidence intervals, t-tests, F-tests, and Dixon Q-tests are discussed as statistical methods for analyzing data accuracy and precision and detecting outliers.

Uploaded by

Steven Joshua Driz
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
42 views

2 - Errors and Statistical Data in Chemical Analysis

This document discusses the importance of statistics in analytical chemistry for communicating experimental data, analyzing finite data sets, and designing optimal experiments. It defines key statistical terms like mean, median, population, sample, accuracy, and precision. Accuracy refers to the agreement with the true value while precision refers to the reproducibility of results. Systematic errors are flaws in design or equipment while random errors are uncontrolled variables. Standard deviation, confidence intervals, t-tests, F-tests, and Dixon Q-tests are discussed as statistical methods for analyzing data accuracy and precision and detecting outliers.

Uploaded by

Steven Joshua Driz
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 21

Errors and Statistical Data in Chemical

Analysis
CHM 311 - Analytical Chemistry
Why do we need statistics in analytical
chemistry?

• Scientists need a standard format to communicate


significance of experimental numerical data.

• Objective mathematical data analysis methods


needed to get the most information from finite data
sets

• To provide a basis for optimal experimental design.


What Does Statistics Involve?

• Defining properties of probability distributions for


infinite populations
• Application of these properties to treatment of finite
(real-world) data sets
• Probabilistic approaches to:
• Reporting data
• Data treatment
• Finite sampling
• Experimental design
Some Useful Statistics Terms

• Mean – Average of a set of values


• Median – Mid-point of a set of values.
• Population – A collection of an infinite number of
measurements. N  infinity
• Sample – A finite set of measurements which represent the
population.
• True value (true mean)- (m), mean value for the population.
• Observed Mean –(x), mean value of the sample set
Accuracy v. Precision

• Accuracy: degree of agreement between


measured value and the true value.

• Absolute true value is seldom known


• Realistic Definition: degree of agreement
between measured value and accepted true
value.
Accuracy v. Precision

• Precision: degree of agreement between replicate


measurements of same quantity.
• Repeatability of a result
• Standard Deviation
• Coefficient of Variation
• Range of Data
• Confidence Interval about Mean Value
Accuracy v. Precision
• You can’t have accuracy without good
precision.
• But a precise result can have a determinate
or systematic error.
Systematic Errors

• Systematic error, also called determinate error, arises from a flaw in


equipment or the design of an experiment.

• For example, a pH meter that has been standardized incorrectly


produces a systematic error. Suppose you think that the pH of the buffer
used to standardize the meter is 7.00, but it is really 7.08. Then all your
pH readings will be 0.08 pH unit too low. When you read a pH of 5.60,
the actual pH of the sample is 5.68. This systematic error could be
discovered by using a second buffer of known pH to test the meter
Ways to detect systematic error
• 1. Analyze a known sample, such as a certified reference material. Your
method should reproduce the known answer.

• 2. Analyze blank samples containing no analyte being sought. If you


observe a nonzero result, your method responds to more than you
intend.

• 3. Use different analytical methods to measure the same quantity. If


results do not agree, there is error in one (or more) of the methods.

• 4. Round robin experiment: Different people in several laboratories


analyze identical samples by the same or different methods.
Disagreement beyond the estimated random error is systematic error.
Random Errors

• Random error, also called indeterminate error, arises from uncontrolled


(and maybe uncontrollable) variables in the measurement. Random
error has an equal chance of being positive or negative. It is always
present and cannot be corrected.
Ways of Expressing Accuracy

• Absolute Errors: difference between true value and


measured value
• Mean Errors: difference between true value and mean
value
• Relative Error: Absolute or Mean Errors expressed as a
percentage of the true value
• ((m-x)/m)*100 = % Relative Error
• Relative Accuracy: measured or mean value expressed as
a percentage of true value
((x/m)*100 = % Relative Accuracy
Standard Deviation

• Standard Deviation s of an infinite set of experimental data is


theoretically given by

• s = S(xi – m)2/N

• xi = individual measurement
• m = mean of infinite number of measurements (true value)
• N = number of measurements
Standard Deviation of a Finite Set of
Experimental Data

• Estimated Standard Deviation, s (N < 30)


• s = (S(xi – x)2/(N-1))
• For finite sets the precision is represented by s.
• Standard deviation of the mean smean
• Smean = s/N
• Relative standard deviation rsd: or coefficient of
variation
• (s/mean)*100 = % rsd
Confidence Limit: How sure are you?

• Confidence Limit = x ± ts/N


t statistical factor that depends on the number of degrees of freedom
 degrees of freedom = N-1
T test v. F test

• t Test When an Accepted Value is Known


m = x ± ts/N
It follows
±t = (x- m) N/s

The F Test
• Designed to indicate whether there is a difference between two methods.
• F = s12/s22 degrees of freedoms 1 and 2

If calculated F value exceeds a tabulated F value at a selected confidence level, then there is a
significant difference between the variances of the two methods
• F = s12/s22.
• You compare the variances of two different methods to see
if there is a significant difference in the methods, at the
95% confidence level.
• Select a confidence
level (95% is good) for
the number of samples
analyzed (=
degrees of freedom +1).
• Confidence limit = x ±
ts/√N.
• It depends on the
precision, s, and the
confidence level you
select.
Dixon Q Test: Rejection of a Result

• The Q test is used to determine if an “outlier” is due to a determinate


error. If it is not, then it falls within the expected random error and
should be retained.
• Q = a/w
• a = difference between “outlier” and nearest sorted result
• w = range of results.
• QCalc = outlier
difference/range.
• If QCalc > QTable,
then reject the
outlier as due to a
systematic error.
Errors and Statistical Data in Chemical
Analysis
CHM 311 - Analytical Chemistry

You might also like