Scribd
Upload
33 views

Lab 4 .

The document discusses statistical evaluation of analytical results including calculating the mean, standard deviation, and coefficient of variation from replicate measurements. It also covers determining the validity of results by comparing means using the student's t-test and using the Q-test to identify outliers.

Uploaded by

w.balawi30
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
33 views

Lab 4 .

The document discusses statistical evaluation of analytical results including calculating the mean, standard deviation, and coefficient of variation from replicate measurements. It also covers determining the validity of results by comparing means using the student's t-test and using the Q-test to identify outliers.

Uploaded by

w.balawi30
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Lab Four

Statistical Evaluation of Analytical Results

Introduction:
Analytical results usually reflect a situation from which a decision will be made such as in
medicinal and environmental measurements. Thus, these results will be subjected to certain
rules before accepted or rejected. Statistics give us tools to accept conclusions that have a high

1
probability of correct and to reject conclusions that do not. This experiment applies the basic
statistical rules that are widely used in analytical chemistry.
For an analytical method, usually we collect finite number of measurements xi, where i = 1 to
“n”, and “n” is the number of replicates. From these measurements we calculate the mean; x
(also called average) Which represents the sum of the measured values divided by “n”:

The standard deviation, s, measures how closely the replicates are clustered around the mean.

Usually, it is more meaningful to represent the percent relative standard deviation known as
the coefficient of variation (CV) rather than the absolute standard deviation, s:

• Coefficient of variation: CV = (s/x) x 100


From the limited number of measurements, we cannot find the true population mean µ. The
confidence interval is an expression stating that the true mean, µ, is likely to lie within a certain
distance from measured mean, x:

• Confidence interval = (x + ts/N1/2) to (x - ts/N1/2)


o Where “t” is the student’s “t”, taken from the following table:

2
Validity of Analytical Results:
Validity of analytical results is determined by comparing mean of measured values with the true
value, comparing means of measured results obtained by different methods, instruments,
persons....etc. With each other and comparing individual differences between two methods,
instruments, persons...etc. All these methods are based on comparison of means with the
student’s t. A calculated t-value is compared with the table t-value with the same confidence
level (95%) and same degree of freedom. Lower t- calculated values than tabulated values
means "no significant difference and absence of determinate errors"

Case 1: Comparing Replicate Measurements:


In this case we have two sets of replicates for the same analyte in the same method obtained
by different methods, persons, or instrument, etc. For the first set we calculate x1, and s1, and
for the second set we calculate x2 and s2. Note that it is not necessary that n1=n2

• Spooled = [S12(n1-1)+S22(n2-1)/ n1 + n2]1/2


Spooled is a pooled standard deviation making use of both sets of data, tcalculated is compared with
the tabulated one at (n1+n2-2) degrees of freedom.

3
Case 2: Q Test for Bad
Sometimes one datum is inconsistent with the remaining data. You can use
the Q test to help decide to retain or discard a suspect datum:
Where gap is the difference between suspect and next one and range is the
difference between highest and lowest
Q calculated= gap/range
If Q calculated > Qtable, the suspect point should be discarded. otherwise, it
should be retained

Objectives:

• Understand basic statistical measures of uncertainty


• Learn when standard deviation vs. standard deviation of the mean is appropriate
• Test for a statistically significant difference from an expected value.

Laboratory Procedure:

• Your instructor will supply you with two set of objects. The set is consisted from four
tablets Weigh each tablet on balance one and balance two and record the weights on
the data sheet.

4
5
6

You might also like