Chapter 2

EVALUATION OF EXPERIMENTAL DATA

Asnida Yanti Ani
Faculty of Applied Science (Chemistry)
Universiti Teknologi MARA (Perlis)

1
 2.1
(Significant Figures)

2
 Significant Figures
 The term significant figures refers to the number of important
single digits (0 through 9 inclusive) in the coefficient of an
expression in scientific notation .

 The number of significant figures in an expression indicates

the confidence or precision with which an engineer or
scientist states a quantity.

3
 For example…
 Look at the ruler below

 What would be the measurement in the correct number of

sig figs?
 _______________

4
 Let’s try graduated cylinders
 Look at the graduated cylinder below

 What would be the measurement in the correct number of

sig figs?

 _______________

5
 Rules for Counting s.f.
1. Any digit that is not zero is significant
 845 – 3 s.f.
 1.234 – 4 s.f.

2. Zeroes between non-zero digits are significant

 606 – 3 s.f.
 40,005 – 5 s.f.

3. Zeroes to the left of the first non-zero digits are

not significant
 0.08 L – 1 s.f.
 0.00039 – 2 s.f.
6
 4. If a number is greater than 1, then all zeroes
written to the right of the decimal point is
count as significant figures
 2.0 mg – 2 s.f.
 3.040 dm3 – 4 s.f.

5. If a number is less than 1, then only the zeroes

that are at the end of the number and zeroes
that are between non-zero digits are significant
 0.090 kg – 2 s.f.
 0.3005 L – 4 s.f.

7
 6. For numbers that do not contain decimal points,
zeroes after the last non-zero digits may not be
significant
 400 cm may have 1 s.f. (the digit 4)
 2 s.f. (40)
 3 s.f. (400)

By using scientific notation, we can express the

number 400 as
 4 x 102 – 1 s.f.
 4.0 x 102 – 2 s.f.
 4.00 x 102 – 3 s.f

8
 s.f. in Calculations

Multiplication and
Division: the number of s.f.
in the answer should not be
greater than the factor with
the fewest s.f.

1.827 m × 0.762 m = ?

0.762 has 3 s.f. so the

reported answer is 1.39 m2

9
SIGNIFICANT FIGURES (MULTIPLICATION & DIVISION)

• When multiplying or dividing the answer will have the

same number of significant digits as the least accurate
number used to get the answer. For example, …

2.005 g / 4.95 mL = 0.405 g/mL

• What is 16.874 x 2.6?

10
 Addition and Subtraction: the reported results
should have the same number of decimal places as
the number with the fewest decimal places

NOTE -
 Be cautious of
round-off errors in
multi-step problems.
 Wait until calculating
the final answer
before rounding.

11
 SIGNIFICANT FIGURES (ADDITION & SUBTRACTION)
• When adding or subtracting do NOT extend the result beyond
the first column with a doubtful figure. For example, …

12
 Datum Number of Datum Number of
(grams) Significant (milliliters) Significant
Figures Figures
10,034 150.
1.908 0.705
0.32 0.054
0.00046 5.86 x 10-7
150 3040
0.0000160 0.0000730

13
 2.2
(Tpyes and Sources of Error)

14
 Types of error

Systematic Error Random Error

What is?
Sources ?

15
Systematic /Determinate error

 All measuring devices are sources of determinate

errors.
 Determinate error have a definite source that
usually can be identified.
 They cause all the results from replicate
measurements to be either high or low.
 Because the results are either high or low ,
determinate errors are also called systematic
errors.

16
 3 Sources of systematic/
determinate error

 Operational and personal errors

 Method errors

 Instrument and reagent errors

17
 OPERATIONAL / PERSONAL ERRORS.

Many measurements require personal judgments. Examples include:

 estimating the position of a pointer between two scale division,

 the colour of a solution at endpoint in a titration
 Incorrect reading of miniscus
 Bias is another source of personal error that varies considerable
from person to person
 Inability to correctly followed procedures, such as weighing sample
without waiting for a complete drying in a gravimetric analysis
 Mathematical errors in calculation
 contamination of sample, incomplete dissolution of sample,
 “bumping” of sample during heating (no boiling chips).

How to avoid/correct : Taking extra care and self -discipline

18
 METHOD ERRORS.
Most serious error since difficult to detect. For example :
 the slowness of some reactions,
 the incompleteness of reactions,
 the instability of some species,
 the non-specificity of most reagents,
 and the possible occurrence of side reactions that interfere with the measurement process.

A common method error in volumetric methods results from the small excess of reagent required to
cause an indicator to undergo the colour change that signals completion of reaction.

How to correct/overcome the error:

i) Using standard sample
Analysis using standard reference materials – materials that contain one or more analyses at
exactly known concern.
ii) Performing ‘blank determination’.
 In a blank determination, all steps of the analysis are performed in the absence of the sample.
 (A blank solution is a solution that contains the solvent and all the reagents in an analysis
except the analyte)
iii) Variation in sample size.
19
 INSTRUMENT AND REAGENT ERRORS
•Any glassware used for quantitative measurements (pipettes, burets, etc) are
potential source of systematic error if they are not calibrated. For example:
oImproperly calibrated - pipettes, burettes and volumetric flasks may have
volumes slightly different from those indicated by their graduations
oglassware at a temperature that differs significantly from the calibration
temperature from contaminants on the inner surface of the containers.
oUse of reagents containing known amount of impurities

•For electronic measuring apparatus, systematic errors can come about as

the result of low batteries, poor contacts within the device, sensitivity to
temperature and humidity instrument malfunction.

How to avoid/ correct - Always remember to do calibration of the glassware

or on the instruments.
20
 RANDOM/INDETERMINATE ERRORS.
 Random or indeterminate errors occur when a system of
measurement is extended to its maximum sensitivity (successive
measurements)

 They are cause by the many uncontrollable variables that are an

inevitable part of every physical and chemical measurement.

 The detection of this type of errors is difficult because they are

VERY SMALL and non of them can be positively identified or
measured.

 The indeterminate errors are responsible for DEVIATIONS

that occur in a series of experimental data.

21
 2.3
(Precision & Accuracy)

22
Statistical Procedures to handle analytical data

 Mean
 median
 range
 precise and deviations
 accurate and error

23
 Mean (average) x

 Mean is the average reading of all data or measurements that

are obtained from an experiments.

 Mean can be calculated by dividing the sum of replicate

measurements by the number of measurements in the set.
• Mean for sample n
=x  i 1
xi
n

where n is the number of samples.

24
 Median
Is a value that is in the middle of a set of data.
Median of a set of replicate data is the middle
result when the data are arranged by
increasing in size.

Range (Spread)
The highest value – the lowest value

25
 Precision
 Precision is how close the measured values are to each
other measurements that have been made in exactly the
same way.
 Precision describes the reproducibility of results, that is
the agreement between two or more
 Expressed in terms of DEVIATION
 If the DEVIATION between the measurements is small,
then your works are precise.

Deviation = ( Value obtained - Mean Value ) = ( xi - x )

Relative Deviation (% Deviation) = deviation / mean x 100

= ( xi - x )/ x x 100

26
 ACCURACY
 Accuracy is how close a measured value is to the actual
(true) value or expert value which we believe to be correct.

 Expressed in terms of ERROR/ABSOLUTE ERROR

E = ( Value obtained - True Value )

If the error is small, then our measurements are accurate

and it shows the accuracy of the results obtained. Usually
the errors are reported as % error or relative error.

27
 Absolute error, E

E = ( xi - xt ) The sign of absolute

error indicate whether
where xi value obtained experimentally, xt the value in question is
is the true value. either high or low as
compared to true value.

Relative error , Er Percent relative error, % E

xi - x t
%E=
xi - x t x 100
xt
xt
28
 Example:

The result of an analysis is 36.97g as compared to the

accepted value which is 37.06 g. What is the relative
error ?

Error ( E) =|36.97- 37.06 | =0.09

Relative error ( Er) = 0.09/37.06 x 100 = 0.243%

29
 Exercise

A student carried out an experiment to measure the boiling of an organic

compound at atmospheric pressure. The results obtained are as follows.

Exp.1. Exp.2 Exp.3 Exp.4

54.9 oC 54.4 oC 54.1 oC 54.2 oC

The true boiling for the compound is 54.0 oC. Calculate :

i. Error
ii. Percentage error
iii. Mean
iv. Deviation
v. Percentage/relative deviation
vi. Median.

30
 i) Error = Measured value – true value

Exp.1  54.9  54.0  0.9

Exp.2  54.4  54.0  0.4
Exp.3  54.1  54.0  0.1
Exp.4  54.2  54.0  0.2

ii ) %/Relative error = Error x 100

true value
Exp.1  0.9 100  1.7
54
Exp.2  0.4 100  0.7
54
Exp.3  0.1 100  0.2
54
Exp.4  0.2 100  0.4
54
31
  iii) Mean
54.9  54.4  54.1  54.2
Mean 
4
Mean  54.4#

 iv) Deviation = Measured value – mean

Exp.1  54.9  54.4  0.5

Exp.2  54.4  54.4  0.0
Exp.3  54.1  54.4  0.3
Exp.4  54.2  54.4  0.2

32
 v) % deviation = deviation x 100
mean

Exp.1  0.5  100  0.9

54.4
Exp.2  0
Exp.3   0.3  100  0.6
54.4
Exp.4   0.2  100  0.4
54.4

vi) median = 54.2 + 54.4 = 54.3

2

33
EXAMPLE

High Accuracy Low Precision ??????

Low Accuracy High Precision ????

High Accuracy High Precision ????

34
 Sample & Population
 SAMPLE  POPULATION
 A SMALL set of data <20  LARGE SET of data (>20)

n
 =  xi  µ= N
x i1  xi
i1
n N
where n is the number of (where N is very large for
real population number.)
samples.
(µ is the true mean )
(usually small numbers <20)

35
THE SAMPLE STANDARD DEVIATION,(s)

The Standard
S= n
 i
(x  x )2 Deviation, (s) is a
i 1 more significant
n 1 quantity in that it
measures the
precision or scatter
of sample data set

The n – 1 term represents the degree of freedom. In the

calculation of standard deviation, the degree of freedom is
reduced by ONE.

36
 THE POPULATION STANDARD DEVIATION
().

When N   ,   and s  
x

True Standard Deviation ,  =

N 2
 (x i  μ)
i 1
N

37
 Relative Standard Deviation (RSD) and
% Standard Deviation (% RSD)

RSD = Standard Deviation = S

mean

% RSD = RSD x100

38
 Exercises

Five students obtained % of chlorine in

sample as follows:
19.82 , 19.57 , 19.68 , 19.71 and
19.75
Calculate i) mean
ii) standard deviation

39
 2.4
(Q-test and confidence limit)

40
 THE USES OF STATISTICS
 Q-test :
Deciding whether an outlying value (outliers) in a
set of replicate results should be RETAINED or
REJECTED in calculating the mean for the set of
data.

 T-test :
Determining the number of replicate
measurements required so that the experimental
mean falls in the range where expected true
value lies

41
 Rejection of Data- The Q test.
 In any set data, there may exist an individual value/datum that
is very far away from the mean value and has a strong
influence on the mean of a data set.

Example for this set of data :

5.09, 5.4, 5.5, 5.57, 5.58, 5.59, 5.61

 The value that lies outside the establish confidence limit is

called an outlier and likely to be rejected

42
  The value 5.09 appears to be out of class, or an outlier. It may have been
obtained from a different population. Notice how this number affects the
calculated means:

5.09 + 5.40 + 5.50+ 5.57+ 5.58+ 5.59+ 5.61 = 5.48

7

if this value is removed : 5.4 + 5.5+ 5.57+ 5.58+ 5.59+ 5.61 =5.54
6

 Since the mean obtained using 5.09 is very different from the majority of
the data, we suspect it is out of class and should be rejected

 A method to determine whether an individual value lie outside the

confidence limit or not is a statistical test called the Q-test.

43
  In this test a value for Q is calculated and compared to a table
of Q value that represents certain % of confidence that the
proposed rejection is valid.

The Q value is calculated using:

Q= Gap
Range

suspect value  closest value

Q
highest va lue-lowest value

44
  Gap = 5.4 -5.09 = 0.31
Range = 5.61 -5.09 =0.52

Q = Gap = 0.31 = 0.59

range 0.52

The calculated Q value is then compared with the value of Q

from Q-table

if Qcal < Qtab, the data is retained

if Qcal > Qtab, the data is rejected and a new mean must be
calculated

45
 Q values for rejection of result in Q-test

No of. data 90% 95% 99%

3 0.94 0.97 0.99
4 0.76 0.83 0.93
5 0.64 0.72 0.82
6 0.56 0.62 0.74
7 0.51 0.57 0.68
8 0.47 0.52 0.63
46
 STEPS IN PERFORMING THE Q –test
1. Rearrange the data set in order from the minimum to
maximum.
2. Calculate the mean from the data given.
3. Look for the suspected outlier that is the value which is far
from the mean value.
4. Calculate the Q
Qcal = Gap Q
suspect value  closest value
Range highest va lue-lowest value
5. Check the value of Qcal and Qtab at a particular confidence
limit.
6. If Qcal > Qtab = the suspected value is to be rejected and the
mean has to be recalculate.

47
 T-TEST

 The t-test assesses whether the means of two groups are

statistically different from each other. This analysis is
appropriate whenever you want to compare the means of
two groups

Determining the number of replicate measurements

required so that the experimental mean falls in the
range where expected true value lies (T-test)

48
Probability/chances we called it RANGE.

RANGE which the true value falls (the highest-the

lowest).

The range is called the confidence interval and the limit

of this range is called the confidence limit.

confidence limit = ±

x ts
N
Where t = N-1(Degree of freedom from Table 1)
and N = number of trial

49
 t- value at various confidence interval 90%, 95% and
99%
Degree of
freedom 90% 95% 99%
(N-1)
1 6.314 12.706 63.657
2 2.920 4.303 9.925
3 2.353 3.182 5.841
4 2.132 2.776 4.604
5 2.015 2.571 4.032

50
6 1.943 2.447 3.702
 Exercise

During the standardization of HCl solution with

0.05 M Na2CO3 standard solution, the buret
readings obtained are shown in the table below:

Trial I II III IV V VI VII

Buret 25.75 25.62 25.52 25.21 25.65 25.60 25.71
Vol
.(mL)

The true known value is 25.63 mL

51
 a) Determine:
i. mean ii. Median
iii. Error iv. % error
v. Deviation vi. % deviation vii) range

b) i. Using Q test at 90 %, find out if any data must be

rejected.
ii. Calculate the standard deviation and the volume
of HCl that can be reported at 90% confidence level.

52
 i) Mean = 25.58 mL

ii) Median = 25.62 mL

iii) Error = +0.12 – 0.01, -0.11, -0.42 +0.02, - 0.03, +0.08

iv) % Error = 0.47% - 0.04% - 0.43% -1.64% 0.08%

v) Deviation = +0.17, +0.04, +0.13 -0.06…

53
 vii) % deviation = 0.66% 0.16% 0.51%....

viii) Range = 0.54

b) i) Q test =suspected value = 25.21

nearest neighbor = 25.52
gap 25.52 - 25.21 = 0.31
Q = gap/ range
= 0.57
Qtable 0.51 Q cal > Q tab REJECT

54
 ii) No of reading = 6
new mean = 25.64 mL

= 0.08

ts
  x
N

at 90% CL = 25.64 + (2.015) (0.08)

√6
= 25.64 + 0.067

55