𝑏𝑦 𝑆𝑢𝑚𝑖𝑡 𝐾𝑢𝑚𝑎𝑟

1. Objective 2.2
Given the following hypothetical population, estimate the average number of
cigarettes a person smokes by selecting a stratified random sample of size n=5 by
proportional allocation.
𝑆𝑡𝑟𝑎𝑡𝑢𝑚 𝐼 (𝑚𝑒𝑛) 𝑆𝑡𝑟𝑎𝑡𝑢𝑚 𝐼𝐼 (𝑤𝑜𝑚𝑒𝑛)
𝑌11 = 20 𝑌21 = 10
𝑌12 = 25 𝑌22 = 12
𝑌13 = 35 𝑌23 = 8
𝑌14 = 30 𝑌24 = 6
𝑌15 = 24
𝑌16 = 26

2. Calculation
For proportional allocation we know that
𝑛
𝑛𝑖 = .𝑁
𝑁 𝑖
𝑛 5 1
Where = =
𝑁 10 2

1
𝑛1 = ×6=3
2
1
𝑛2 = ×4=2
2
Let we select from table

𝑦11 = 25, 𝑦12 = 20, 𝑦13 = 35 𝑓𝑟𝑜𝑚 𝑠𝑡𝑟𝑎𝑡𝑢𝑚 𝐼

𝑦21 = 10, 𝑦22 = 6 𝑓𝑟𝑜𝑚 𝑠𝑡𝑟𝑎𝑡𝑢𝑚 𝐼𝐼

Hence the estimator 𝑦𝑠𝑡 𝑖𝑠

𝑘
1 (𝑁1 𝑦1 + 𝑁2 𝑦2 )
𝑦𝑠𝑡 = ∑ 𝑁𝑖 𝑦𝑖 =
𝑁 10
𝑖=1

(𝑦11 + 𝑦12 + 𝑦13 ) 25 + 20 + 35 80

𝑦1 = = =
3 3 3
𝑦21 + 𝑦22 10 + 6 16
𝑦2 = = =
2 2 2
80 16
6×(3)+4×( 2) 192
𝑦𝑠𝑡 = =
10 10
𝑏𝑦 𝑆𝑢𝑚𝑖𝑡 𝐾𝑢𝑚𝑎𝑟

The population mean (true average) is

𝑘 𝑁𝑖
1
𝑌 = ∑ ∑ 𝑌𝑖𝑗
𝑁
𝑖=1 𝑗=1

1
𝑌= [(𝑌 + 𝑌12 + 𝑌13 + 𝑌14 + 𝑌15 + 𝑌16 ) + (𝑌21 + 𝑌22 + 𝑌23 + 𝑌24 )]
10 11
1
𝑌= [(20 + 25 + 35 + 30 + 24 + 26) + (10 + 12 + 8 + 6)]
10
1 196
𝑌= [160 + 36] =
10 10
𝑌 = 19.6 … … … … (𝑖)

3. Conclusion
Hence the estimate average number of cigarettes a person smokes by selecting a
stratified random sampling of size n=5 by proportional allocation is given by (i) i.e.
19.6.
 𝑏𝑦 𝑆𝑢𝑚𝑖𝑡 𝐾𝑢𝑚𝑎𝑟

1. Objective 2.3
A sampler proposes to draw a stratified random sample. He expects that his field
costs will be of the form
𝑘

∑ 𝐶𝑖 𝑛𝑖 .
𝑖=1

His advance estimate of relevant quantities for the two strata are as follows:
𝑆𝑡𝑟𝑎𝑡𝑎 𝑊𝑖 𝑆𝑖 𝐶𝑖 (𝑖𝑛 ₹)
1 0.4 10 4
2 0.6 20 9

𝑛1 𝑛2
i. Find the value of and that minimises the total field cost for the given value of
𝑛 𝑛
𝑉(𝑦𝑠𝑡), where yst is the stratified sample mean.
ii. Find the sample size required under the optimum allocation to make v(yst)=1.
iii. How much total field cost will be.
2. Calculation
We know
𝑛𝑁𝑖 𝑆𝑖
√𝐶𝑖
𝑛𝑖 = , ∀𝑖 = 1,2,3 … 𝑘
𝑁𝑆
[∑𝑘𝑖=1 𝑖 𝑖 ]
√𝐶𝑖
Where 𝐶𝑖 is the cost per unit in 𝑖𝑡ℎ stratum

𝑊1 𝑆1
10
𝑛1 √𝐶1 0.4 × 1
= = 2 = … … (𝑖)
𝑛 𝑊1 𝑆1 𝑊2 𝑆2 0.4 × 10 0.6 × 20 3
+ +
√𝐶1 √𝐶2 2 2

𝑊2 𝑆2
20
𝑛2 √𝐶2 0.6 × 2
= = 3 = … … . (𝑖𝑖)
𝑛 𝑊1 𝑆1 𝑊2 𝑆2 0.4 × 10 0.6 × 20 3
+ +
√𝐶1 √𝐶2 2 2
𝑏𝑦 𝑆𝑢𝑚𝑖𝑡 𝐾𝑢𝑚𝑎𝑟
Now
2
𝑁𝑖2 𝑆𝑖2 𝑊12 𝑆12 𝑊22 𝑆22
𝑣(𝑦𝑠𝑡 ) = ∑ 2 = +
𝑁 . 𝑛𝑖 𝑛1 𝑛2
𝑖=1

𝑊12 𝑆12 𝑊22 𝑆22

𝑙= 𝑛 + 𝑛
1 2
𝑛 𝑛
3
= (16 × 3) + (144 × ) = 264
2
∵ 𝑛 = 264 … … … (𝑖𝑖𝑖)
𝑛1 1 264
𝑤𝑒 ℎ𝑎𝑣𝑒 = 𝑜𝑟 𝑛1 = = 88
𝑛 3 3
𝑛2 2 2 × 264
= 𝑜𝑟 𝑛2 = = 176
𝑛 3 3
𝑘

𝑡ℎ𝑒𝑛 𝐶 = ∑ 𝐶𝑖 𝑛𝑖 = 𝐶1 𝑛1
𝑖

= 4 × 88 + 9 × 176
= 352 + 1584
= ₹ 1936 … … . . (𝑖𝑣)
Conclusion
i) Hence the value of n1/n is 1/3 and the value of n2/n is 2/3
ii) The sample size required under the optimum allocation to make variance
under stratified sample for 1 is 264
iii) The total field cost will be ₹1936
𝑏𝑦 𝑆𝑢𝑚𝑖𝑡 𝐾𝑢𝑚𝑎𝑟
𝑬𝒙𝒑𝒆𝒓𝒊𝒎𝒆𝒏𝒕 𝟑
1. Case Study of Systematic Sampling
Systematic Sampling is commonly employed technique if the complete and up to date list of the
sampling unit is available. This consists in selecting only the first unit at random the rest being
automatically selected according to some predetermined pattern involving regular spacing of units.

Let us suppose that 𝑁 sampling units are serially numbered from 1 to 𝑁 in some order and a
𝑁
sample of size n is to be drawn such that 𝑁 = 𝑛𝑘 ⇒ 𝑘 = 𝑛

𝐾 = sampling interval

Systematic sampling consists in drawing a random number say 𝑖 < 𝑘 and selecting the unit
corresponding to this number and every 𝑘 𝑡ℎ unit subsequently. This the sample size of n units will
consist of the units

𝑖, 𝑖 + 𝑘, 𝑖 + 2𝑘, … … … … 𝑖 + (𝑛 − 1)𝑘

I. Advantage
 Sample selection is easy
 Personal bias is eliminated.
 Easy to manage large population.
II. Disadvantages
 Not all numbers have the equal chance of selection only the initial respondent us selected
at random basis.
 Not suitable when the population of homogeneous.
 Sample of the same characteristics is selected over time which makes no sense.
3. Notations and Terminology
Let 𝑦𝑖𝑗 denote the observation on the jth unit of the 𝑖 𝑡ℎ sample (𝑖 = 1,2,3, … . , 𝑘 ; 𝑗 =
1,2,3, … . . 𝑛)
 𝑦𝑖 = mean of 𝑖 𝑡ℎ systematic sample
𝑛
1
𝑦𝑖 = ∑ 𝑦𝑖𝑗
𝑛
𝑗=1

 𝑦•• = 𝑡ℎ𝑒 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑚𝑒𝑎𝑛

𝑘 𝑛 𝑘
1 1
𝑦•• = ∑ ∑ 𝑦𝑖𝑗 = ∑ 𝑦𝑖
𝑛𝑘 𝑘
𝑖=1 𝑗=1 𝑖=1

 𝑆 2 = 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑚𝑒𝑤𝑛 𝑠𝑞𝑢𝑎𝑟𝑒

𝑏𝑦 𝑆𝑢𝑚𝑖𝑡 𝐾𝑢𝑚𝑎𝑟
𝑘 𝑛 𝑛 𝑛
1 2 1 2
𝑆2 = ∑ ∑(𝑦𝑖𝑗 − 𝑦•• ) = ∑ ∑(𝑦𝑖𝑗 − 𝑦•• )
𝑁−1 (𝑛𝑘 − 1)
𝑖=1 𝑗=1 𝑖=1 𝑗=1

𝑘 𝑛
2
1 2
𝑆𝑤𝑠𝑦 = ∑ ∑(𝑦𝑖𝑗 − 𝑦𝑖 )
𝑘(𝑛 − 1)
𝑖=1 𝑗=1

𝑁 − 1 2 (𝑛 − 1)𝑘 2
𝑣𝑎𝑟 (𝑦𝑠𝑦𝑠 ) = .𝑆 − . 𝑆𝑤𝑠𝑦
𝑁 𝑁
4. Objective
The data in the table for artificial population which exhibits a fairly steady rising trend. Each
column represents a systematic sample and the rows are the strata. Compare the precision of
systematic sampling, random sampling and stratified sampling.

Data for 10 systematic sampling with

𝑁 = 4, 𝑘 = 10, 𝑁 = 𝑘𝑛 = 40

Systematic Random Sampling

Strata
1 2 3 4 5 6 7 8 9 10
I 0 1 1 2 5 4 7 7 8 6
II 6 8 9 10 13 12 15 16 16 17
III 18 19 20 20 24 23 25 28 29 27
IV 26 30 31 31 33 32 35 37 38 38

Solution: we have
𝑘 𝑘
1 2 1 2 2
𝑣𝑎𝑟(𝑦𝑠𝑦𝑠 ) = ∑(𝑦𝑖 − 𝑦•• ) = [∑ 𝑦 𝑖• − 𝑘𝑦 •• ]
𝑘 𝑘
𝑖=1 𝑖=1

𝑘
1 2 2
= 2 [∑(𝑛𝑦𝑖• ) − 𝑛2 𝑘𝑦 •• ]
𝑛 𝑘
𝑖=1

2 𝑘−1 2
𝑣𝑎𝑟(𝑦𝑠𝑡 ) = . 𝑆𝑠𝑤𝑡
𝑛𝑘

2
1 2
𝑤ℎ𝑒𝑟𝑒 𝑆𝑠𝑤𝑡 = ∑ ∑(𝑦𝑖𝑗 − 𝑦𝑖• )
𝑛(𝑘 − 1)
𝑖 𝑗

1 1
𝑣(𝑦𝑛 ) = ( − ) 𝑆 2
𝑅 𝑛 𝑁
1 2
𝑆2 = ∑ ∑(𝑦𝑖𝑗 − 𝑦•• ) 𝑖𝑠 𝑡ℎ𝑒 𝑡𝑜𝑡𝑎𝑙 𝑚𝑒𝑎𝑛 𝑠𝑞𝑢𝑎𝑟𝑒
𝑁−1
𝑖 𝑗
𝑏𝑦 𝑆𝑢𝑚𝑖𝑡 𝐾𝑢𝑚𝑎𝑟

Strata Systematic Random Sampling

1 2 3 4 5 6 7 8 9 10 Total

I 0 1 1 2 5 4 7 7 8 6 41

II 6 18 9 10 13 12 15 16 16 17 132

III 18 19 20 20 24 23 25 28 29 27 233

IV 26 30 31 31 33 32 35 37 38 38 331

total 50 68 61 63 75 71 82 88 91 88 737

2
= ∑ ∑ 𝑦𝑖𝑗 − 𝑐𝑜𝑟𝑟𝑒𝑐𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟

5,28,529
= (727)2 −
40
𝑆𝑆 = 5313.775
𝑛
𝑇𝑗2
𝐵𝑒𝑡𝑤𝑒𝑒𝑛 𝑠𝑡𝑟𝑎𝑡𝑎 = ∑ − 𝐶. 𝐹.
𝑘
𝑗=1

180415
= − 13213.225
10
= 4828 − 275

2
𝑤𝑖𝑡ℎ𝑖𝑛 𝑠𝑡𝑟𝑎𝑡𝑎 𝑆. 𝑆. 485.5
𝑆𝑤𝑠𝑡 = = = 13.486
𝑛(𝑘 − 1) 36

𝑇𝑜𝑡𝑎𝑙𝑆𝑆 5313.775
𝑆2 = = = 136.251
𝑁−1 39
1
𝑣𝑎𝑟 (𝑦𝑠𝑦𝑠 ) = [54713 − 160 × (18.17)2 ] = 11.807
160
1 1
𝑣𝑎𝑟(𝑦𝑠𝑡 ) = ( − ) 13.486 = 3.034
4 40
1 1
𝑣𝑎𝑟(𝑦𝑛 ) = (4 − 40) 136.251 = 30.656
𝑅

𝒗𝒂𝒓(𝒚𝒔𝒕 ) < 𝒗𝒂𝒓 (𝒚𝒔𝒚𝒔 ) < 𝒗𝒂𝒓(𝒚𝒏 )

𝑹