Example 2

Data for the analysis of calcium in rock:

14.35%, 14.41%, 14.40%, 14.32% and 14.37%
Calculate the confidence interval at 95% confidence
level.

ts
μ x 
N

Min, x = 14.37
Standard deviation, s = 0.04

From t table (At 95% confidence level):

Degree of freedom, (N - 1 = 4); t = 2.78

Confidence interval,

2.78  0.04
x
5
 14.37  0.05
At different confidence level,

Confidence Level Confidence Interval

1
 90%  = 14.37%  0.04
95%  = 14.37%  0.05
99%  = 14.37%  0.08

Summary:

If the confidence level increased, the confidence

interval is also increased. The probabilities of 
appear in the x interval increased.

2
TO DETERMINE NUMBER OF REPLICATES

 If  is known (s  ),

z
μ x 
N

2
 
 z 
N 
  x
 

 If  is unknown,

ts
x 
N

2
 ts 
N 
  x

3
Example 3

Calculate the number of replicates needed to

reduce the confidence interval to 1.5 g/mL at 95%
confidence level. Given, s = 2.4 g/mL.

At 95% confidence level, t = 1.96 (d.o.f = ∞),

2
 ts 
N 
  x

2
 1.96  2.4 
N   10
 1.5 

4
TO DETERMINE SYSTEMATIC ERROR

Example 4

Absorption reading of a standard solution at a

particular wavelength is 0.470. Ten measurements
on a sample gave a mean value of 0.461 and
standard deviation of 0.003. Show whether
systematic error exists in the measurements at 95%
confidence level.

At 95% confidence level, N – 1 = 9, t = 2.26,

ts
x
N
 0.461  2.26
0.003
10
 0.461  0.002

The calculation gives confidence limit of,

0.459 <  < 0.463

Does the true mean 0.470 belong to the interval?

Since  is not belong to the interval, so systematic
error is present.

5
STEPS FOR t TEST

1. Comparing two mean values  and x

i) If  is not known,

ts
x 
N

N
 t  x -  
s
ii) If  is known,

  N
 z   x  μ
  σ

iii) Calculate t or z (tcalc) from the data.

iv) Compare tcalc and ttable

6
v) If tcalc > ttable

Reject Null Hypothesis (Ho) i.e.   x .

The different is due to systematic error.

vi) If tcalc < ttable

Accept Null Hypothesis (Ho) i.e. (   x ).

The different is due to random error.

7
Example 5

The sulphur content of a sample of kerosene was

found to be 0.123%. A new method was used on
the same sample and the following data is
obtained:

%S: 0.112; 0.118; 0.113; 0.119

Show whether systematic error is present in the

new method.

Null Hypothesis, Ho:  = x

 = 0.123%

x = 0.116%, s = 0.004

8
 N
 t  x   
s
4
 0.116  0.123
0.004
 3.5 (t calc )

t table = 3.18 (95 %, N - 1 = 3)

Since tcalc> ttable, Ho is rejected.

The two means are significantly different and thus

systematic error is present.

9
2. Comparing two mean values x 1 and x 2

Normally used to determine whether the two

samples are identical or not.

The difference in the mean of two sets of the same

analysis will provide information on the similarity of
the sample or the existence of random error.

Data: x 1 , x 2 and s1, s2

Ho : x1 = x2

We want to test whether x 1 - x 2 = 0

ts
  x1  1
1
N
1

ts
  x2  2
2
N
2

Assume,  1   2 and  1   2 ,

10
Calculate the value of t;

x x 1 2
NN
t = s 1 2

N N
calc
p 1 2

Compare tcalc with ttable; if tcalc< ttable, Ho is accepted.

The pooled standard deviation, sp is calculated

using:

(N1 - 1)s12+(N2 - 1)s22

sp =
N1 + N2 - Ns

where,
N1, N2 are numbers of data in sets 1 and 2
Ns is the number of data sets

11
Example 6

The calcium content (%w/w) of a powdered

mineral sample was analyzed by two methods. The
results are shown below.

Method 1 Method 2
2.71 2.71
2.82 2.68
2.79 2.63
2.71 2.74

Is there any significance different between the

mean values for these two methods at 95%
confidence level?

x1 = 2.76, s1 = 0.06
x2 = 2.69, s2 = 0.05

Ho: x 1 = x 2

12
 (N1 - 1)s12 + (N 2 - 1)s 22
sp=
N1 + N 2 - N s


4 - 10.062  4 - 10.052
44-2
 0.06

xx NN 1 2
t = 1 2
N N
calc
s p
1 2

2.76 - 2.69 4 4

0.06 44
 1.65

ttable = 2.45

tcalc < ttable; Accept Ho.

There is no significance different between the mean

values for these two methods

13
Example 7

The determination of CO in a mixture of gases

using the first procedure gave an s value of 0.21
(10 replicates). The method was modified giving s
of 0.15 (11 replicates).

Is the second method more precise than the first?

Ho : s1 = s2

s22 0.212
Fcalc    1.96
s12 0.15 2

From the F table (One-tailed test):

Degree of freedom for numerator = 9
Degree of freedom for denominator = 10
The critical F value = 3.02

Since Fcalc < Ftable, Ho is accepted.

Conclusion:
The first method has the same precision as the
second method.

14
Example 8

Assume that you are developing a procedure for

determination of cadmium in spirulina tablets. The
analysis was carried out using wet digestion and
measured by atomic absorption spectrometer.

Is there any significant difference in the precision

between the two methods at 95% confidence level?

Cadmium Concentration (ppb)

Your Method Reference Method
20.10 18.89
20.05 19.20
18.85 19.00
18.95 19.70
19.40 19.40
19.99

Ho: s1 = s2
s1 = 0.57, s2 = 0.32

(s 2 )2 (0.57)2
Fcalc  2
 2
 3.17
(s1 ) (0.32)

Two-Tailed Test:
Ftable = 9.36
Fcalc < Ftable. Accept Ho.

There is no significant difference in the

precision of the two methods.

15
Example 9

Data: 10.05, 10.10, 10.15, 10.05, 10.45, 10.10

By inspection, 10.45 seems to be out of the data

normal range. It is easier to see it when the
numbers are arranged in a decreasing or increasing
order.

10.05, 10.05, 10.10, 10.10, 10.15, 10.45

Should this data be eliminated?

x q - xn
Q exp t 
w

where,
xq is the questionable data
xn is its nearest neighbour
w is the difference between the highest and
the lowest value (range).
The Qexpt or Qcalc will be compared with the Qcrit or
Qtable, and the null hypothesis is checked.

16
 10.45 - 10.15
Q exp t 
10.45 - 10.05

= 0.75

Qcrit (95%, n = 6) = 0.625

Qexpt > Qcrit or Qcalc. > Qtable

Data (10.45) can be eliminated.

17