FA-Chapt2

BASICS OF STATISTICAL
APPLICATIONS IN DATA
ANALYSIS
CHAPTER 2

Propagation of Uncertainty
•

1
 FA-Chapt2

Measures of Central tendency

• Mean
The mean of two or more measurements is their average
value.

- Median
• The median is the middle result when replicate data are arranged
in increasing or decreasing order.
• Odd number (locate the middle)/ even number (average value of
middle pair)

Measures of Spread
• Range
• Deviation from the mean di
• Standard deviation
• Variance
• The variance is just the square of the standard deviation
• Coefficient of variation (CV)
• Relative Standard Deviation (RSD)

2
 FA-Chapt2

The Population Mean and the Sample Mean

population standard deviation ,

where N is the number of data points
making up the population.

Measure of Spread
• Sample standard deviation, s
N
2
N N
x
 x - x 
2 2 N i
i  di x i
2
- i =1

s= i =1
= i =1 i =1 N
s=
N -1 N -1 N -1

• Relative Standard Deviation (RSD)

• Coefficient of Variation (CV)

3
 FA-Chapt2

Characterization of analytical errors

• Errors associated with the central
tendency reflect the accuracy of the
analysis
• Errors associated with the spread reflect
the precision of the analysis

Accuracy and Precision

4
 FA-Chapt2

Accuracy
Absolute error: E = xi - xt
Where xt is true or accepted value

- If the measurement result is low, the sign is negative;

- If the measurement result is high, the sign is positive.

Relative error:
xi - xt
Er = 100%
xt
The relative error expressed in percent, parts per
thousand (ppt), or parts per million (ppm), depending on
the magnitude of the result.

Errors in Chemical Analysis

Difference between a measured value and the
“true” or “known” value
Estimated uncertainty in a measurement of
experiment

5
 FA-Chapt2

Types of Errors in Experimental data

1. Random (indeterminate) Error
Data scattered approx. symmetrically about a mean
value
Affects precision
2. Systematic (determinate) Error
Several possible sources.
Readings all too high or too low.
Affects accuracy
3. Gross Errors (or blunders)
Lead to outlier’s and require statistical techniques
to be rejected.
Detectable by carrying out sufficient replicate
measurements.

Systematic Errors
1. Instrumental errors
2. Method errors
3. Measurement errors
4. Sampling errors
5. Personal errors

• Bias measures the systematic error associated

with an analysis. It has a negative sign if it causes
the results to be low and a positive sign otherwise.

6
 FA-Chapt2

Random Errors
1. Caused by uncontrollable variables which normally
cannot be defined.

2. The accumulated effect causes replicate

measurements to fluctuate randomly around the mean.

3. Random error give rise to a normal or guassian curve.

4. Results can be evaluated using statistics

5. Usually statistical analysis assumes a normal

distribution.

Random and Systematic errors

7
 FA-Chapt2

Eg. random and systematic errors

• Four students (A–D) each perform an analysis in which
exactly 10.00 ml of exactly 0.1 M sodium hydroxide is
titrated with exactly 0.1 M hydrochloric acid. Each student
performs five replicate titrations, with the results shown

Eg. random and systematic errors

• A standard sample of pooled human blood serum
contains 42.0 g of albumin per litre. Five laboratories (A–
E) each do six determinations (on the same day) of the
albumin concentration, with the following results (g l 1
throughout):

• Comment on the bias, precision and accuracy of each of

these sets of results.

8
 FA-Chapt2

Accuracy and Precision

Precision describes random error,

bias describes systematic error,
and the accuracy, i.e. closeness to the
true value of a single measurement or a
mean value, incorporates both types of
error.

Reproducibility and Repeatability

Replicates: Samples of about the same size that are

carried out through an analysis in exactly the same way.

• Repeatability describes the precision of within-run

replicates
• Reproducibility describes the precision of between-run
replicates
• The reproducibility of a method is normally expected to be
poorer (i.e. with larger random errors) than its
repeatability

9
 FA-Chapt2

Properties of Gaussian Curve

The equation for a Gaussian curve has the form
2 2
e - ( x -  ) / 2
y=
 2

Where:
 – population mean
 - population standard deviation
x – sample mean
s – sample standard deviation

z term
The normal error curve:
variable, z
x-
z=

z is the deviation of a data
point from the mean relative to
one standard deviation.
When x-= , z is 1;
2 2 2 2
Variance, 2 : e - ( x -  ) / 2 e - z / 2
y= =
 2  2

10
 FA-Chapt2

Area under a Gaussian

From equation above, and illustrated by the previous
curves, 68.3% of the data lie within ±  of the mean ,
i.e. 68.3% of the area under the curve lies between ±  of
.

Similarly, 95.5% of area lies between ±2, and 99.7%

between ±3.

There are 68.3 chances in 100 that for a single datum the
random error in the measurement will not exceed ±.

The chances are 95.5 in 100 that the error will not exceed
±2.

Standard Deviation of Calculated Results

11
 FA-Chapt2

Use of Statistics
in Data Evaluation

Confidence Intervals
Finding the Confidence Interval when  is Unknown

We use statistical parameter t (Student’s t), which is exactly

the same way as z except that s is substituted for .
x-
t=
s
For the mean of N measurements,
x-
t=
s Remember:
N is equal to N-1 for a
N small set of data
t
CI for  = x
N

12
 FA-Chapt2

Confidence Intervals
Finding the Confidence Interval when  is Unknown

Detection of Gross Error

The Q-test
xq - xn
Q=
w
Xq is quetionable result, xn is its nearest neighbor, and w is
the spread of data set. Qcrit values are given in the tables. If
Q is greater than Qcrit , the questionable result can be
rejected with the indicated degree of confidence.

13
 FA-Chapt2

Comparison of Two Experimental Means

To compare the results of two different methods, analyst,
conditions e.g. there are two sets of tests N1 and N2. If t
calculated is greater than t in the tabulated one at the 95%
confidence level, the two results are considered to be
different.

sd s 2pooled s 2pooled N1  N 2
=  = s pooled
N N1 N2 N1 N 2

x1 - x2
t=
N1  N 2
s pooled
N1 N 2

Comparison Precision (F-test)

Compare standard deviation of two sets of data.
 one-tailed, i.e. one direction (is Method A better than Method B)
 two-tailed (do Methods A and B differ in their precision).

The F-test considers the ratio of the two samples variances i.e. the ratio
of the squares of the standard deviations.
The quantity calculated (F is always ≥1)

s12
F = 2
s2

14
 FA-Chapt2

Characteristics of an analytical method

Response: The way in which the
result or signal of a method
varies with the amount of
compound or property being
measured.

Calibration curve

Standardization and Calibration

1. Comparison with Standards

2. External Standard Calibration.

- An external standard is prepared separately from the sample.

- Calibration is accomplished by obtaining the response signal

(absorbance, peak height, peak area) as a function of the know
analyte concentration

Parameters used to describe a Calibration Curve:

S – measured signal
c – analyte concentration
Sbl – instrument signal for blank

15
 FA-Chapt2

Standardization and Calibration

Characteristics of an analytical method

Sensitivity:
Calibration sensitivity = slope (m) of calibration curve.
Analytial sensitivity (g)=slope (m)/ standard deviation (Ss)

 ability to discriminate between small differences in

analyte concentration
Slope and reproducibility of the calibration curve.

16
 FA-Chapt2

Limit of Detection
Sm – minimum analyte signal
Sbl- mean blank signal

LmD = S m - S bl / slope

blank signal, yB, plus three standard deviations of the blank, sB

Dynamic Range

Linear region of calibration curve where the lower

limit is ten times the standard deviation of the
blank.
LOQ – Limit of Quantitation
LOL – Limit of Linearity

17
 FA-Chapt2

Significant Figures
 Reflection of a measurement’s uncertainty.

 Number of sig. fig. is equal to the number of digits in the

measurement, with the exception of a zero used to fix the
location of a decimal point.
 How many significant figures are there?
 1.020

 0.020

 100

 1 x 102

 1.0 x 102

Significant figures
• Round-off numbers in Arithmetic
• More than halfway to the next higher digit  round up
• Less than halfway  round down
• Exact the halfway  round up to the nearest even digit
 123.5748  123.57
 123.4560  123.46
 123.5750  123.58

• Addition and Subtraction

• least decimal place (least –certain)
 5.345 + 6.72 = 12.084001 = 1.21 x 101
 7.26 – 6.69 = 0.57
 135.621 + 0.33 + 21.2163 = 157.17

18
 FA-Chapt2

Significant Figures

• Multiple and Division

• number having the fewest sig. figs (least –certain)
8.37  1.345  123.538 = 133.243 = 1,33 102
2.432  106  6.512  10 4 - 0.1227  105
= 2.432 10 6  0.06512  106 - 0.01227 10 6 = 2.37422  106
9.0  1.2000 = 10.8 = 11 = 0.11 102 = 2.374 10 6
100
% X = 0.0748  = 14.2802 = 14.3
0.5238

• When performing calculations, do not round off until the

calculation is completed!

Logarithms and antilogarithms

• The number of digits in the mantissa of log X: number of
digits in mantissa of logx = number of significant figures
in x.

• Number of digits in antilog x (=10x) = number of significant

figures in the mantissa of x

19