METODE DAN DISTRIBUSI SAMPLING

Makalah ini di susun untuk memenuhi salah satu tugas Statistik Ekonomi I

Dosen Pembimbing : Thomas Nadeak SE., MM

Disusun oleh :

Anisa Ramayanti (16416262201031)

Irna Ratnasari (16416262201042)
Lusi Yuli Yani (16416262201083)
Monita Dewi R.N (16416262201036)
Nuryati Fadilah (16416262201175)
Sasa Saldiah (16416262201028)
Syifah Fauziah (16416262201131)
Wahyu Widodo (16416262201053)

AK 16 A

PROGRAM STUDI AKUNTANSI

FAKULTAS BISNIS DAN ILMU SOSIAL

UNIVERSITAS BUANA PERJUANGAN

2017/2018

1
 Preface
First of all, thanks to Allah SWT because of the help of Allah, writer finished writing
the paper entitled “Sampling Methods and Distribution” right in the calculated time.

The purpose in writing this paper is to fulfill the assignment that given by Mr.
Thomas Nadeak SE., MM as lecturer in samantics major.

In arranging this paper, the writer trully get lots challenges and obstructions but with
help of many indiviuals, those obstructions could passed. writer also realized there are still
many mistakes in process of writing this paper.

Because of that, the writer says thank you to all individuals who helps in the process
of writing this paper. hopefully allah replies all helps and bless you all.the writer realized tha
this paper still imperfect in arrangment and the content. then the writer hope the criticism
from the readers can help the writer in perfecting the next paper.last but not the least
Hopefully, this paper can helps the readers to gain more knowledge about samantics major.

Karawang, 4th of January 2018

2
 Table List of Content

Preface .............................................................................................................................. i

Table List of Content ....................................................................................................... ii

Sampling Methods and Distribution

11.1 Introduction .............................................................................................................. 1

11.2 Understanding Population and Sample .................................................................... 3

11.3 Sampling Method .................................................................................................... 5

11.3.1 Simple Random Sampling ..................................................................................... 6
11.3.2 Structured Random Sampling Withdrawal .......................................................... 11
11.3.3 Cluster Sampling ................................................................................................. 14
11.3.4 Systematic Random Sampling ............................................................................. 16
11.3.5 Quota Sampling (Quote Sampling) ..................................................................... 16
11.3.6 Purposive Sampling (Purposive Sampling) ........................................................ 16

11.4 Sampling Error (sampling Error) ........................................................................... 17

11.5 Distribusi Sampel Rata-Rata dan proporsi ............................................................ 19

11.5.1 Using Average Sample Distribution ................................................................... 24
11.5.2 Proportional Sampling Distribution .................................................................... 26
11.5.3Using Sample Distribution of Propottions ............................................................ 29

11.6 Distribution of sample of average difference and proportion ................................ 31

11.7 Correction factor until population is limited ......................................................... 34
11.8 postulate the middle boundary ............................................................................... 35

3
 SAMPLING METHODS AND DISTRIBUTION

11.1 Introduction

In book 1, we have studied the descriptive statistics used to describe or describe the
data that has been collected into an information, such as the size of concentration, the size of
the spread, the index number, as well as the presentation of good and proper data. While in
this book II, will be studied Inductive statistics used to determine the characteristics of
populations based on the characteristics of the sample and analyze and interpret the results
into a conclusion or information useful for decision-making. Information such as probability
of an event, hypothesis testing, and parameter estimation are the main studies of inductive.

Many economic and financial events must be known characteristics, to be useful in

business decision making. some events in economics and finance as well as their relation to
business decisions can be illustrated in the following example.

a. Increasing Revenue. "Want Rich Must Have Side Job", Kompas news reported on
April 29, 2014. Social Economic Status Survey is an effort to know how society can
increase income to enter the upper middle class. The survey was conducted by
selecting 500 respondents who are productive age communities aged between 18-50
years. Sample selection method was done by random sampling method through
telephone interview. Survey results show that side work can increase income between
45 and 65%. So want rich? Start looking for side jobs.
b. Influence of Interest Rate Decline of Bank Indonesia. The decline in Bank
Indonesia interest rates to below 8% followed by the decline in lending rates in the
banking sector, provides an opportunity for investors to expand their business by
utilizing credit from banks. To get credit from the bank of course we must know
some interest rates on credit, how long the repayment of credit, how long the grace
period of payment, and how the requirements to get credit. In 2013 there are 120
national banks including branches of foreign banks and regional development banks
operating in Indonesia, and more than 1,700 people's credit banks, each of which
implements different interest rates, repayment periods and credit terms depending on
policy of each bank. To choose which bank is best for credit, is it necessary to survey
all banks, or just a few banks?
c. Reksa Dana Products. In the last six years (2008-2013), Indonesia's macroeconomic
conditions are less stable due to the impact of the global economic crisis, rising
prices for fuel oil and gas, and the demands for higher wages. At that time, fuel
prices rose twice and gas prices once, the average inflation rate of 5.9% with the
highest inflation rate of 11.06% in 2008 and the lowest inflation rate of 2.78% in
2009, the average economic growth during the last 6 years is 6.37%, the bank interest
rate is relatively low, and the rupiah exchange rate against the US dollar is in the
range of Rp.9600. When deposit rates are low, many public funds are transferred in
the form of investments other than deposits, one of which is mutual funds. Mutual

4
 funds are an investment option with a higher yield rate on deposits, but also have a
greater risk. Therefore, investors should be careful in choosing mutual fund products.
Mutual fund products usually consist of stocks, bonds, SBIs, and other financial
products. Currently there are 82 mutual fund companies and each company has
several mutual fund products. To know the level of investment return and risk level,
is it necessary to know all mutual products and from 82 companies or enough from
some companies only?
d. Advertise on Television. "Driver Bajaj Jokowi-JK and Jokowi KW-2 Being a
Herbal Medicine". Inspired by the success of the "Green Gang" advertisement that
can boost Sampoerna Hijau's sales by up to 200%, the Sido Herbalist company also
tried to advertise on television because, according to Wiharto, Marketing Director of
PT Sido Muncul, the recent increase in sales did not show any significant
improvement . Therefore, there are several combinations between herbal forms and
advertising models that want to be released, such as herbal powder product with
advertising model Eeng Saptahadi and Ine Febriyanti as well as liquid packing herbs
with advertising model of Sophia Latjuba and Rhenald Kasali. The latest herbal
advertisement in 2014 was inspired from the presidential election campaign, then
featured Jokowi Kw-2 and dopir bajaj Jokowi-JK. To decide which product to
launch, is it necessary to gather opinions from all customers, or just enough
customers who are considered representative of lower, middle, and upper class
society?

The above four cases illustrate the two trends in obtaining the characteristic value,
first involving all members of the community or population to obtain parameters, such as
interest rates from all banks, investment returns of all firms reksa dana, all herbal customers
of Sido Emerging, and the whole community of the capital market; secondly, involving some
of the community members or samples such as some of the banking companies and some
herbal customers tolak angin Sido Muncul, and from some of the capital market
communities.

If in obtaining the parameters used tendency to involve all members, then the costs
incurred will be very high or even often impossible to implement. To find out the loan
interest rate, the return on investment, and the credit terms of the entire 120 Banks, will
require greater time and cost. So also to know the level of investment returns from various
products and product combinations from the entire company reksa dana. In the case of Sido
Muncul, it is also very impossible to collect opinions from all customers. In addition to the
list of unavailable customers, as well as cost and time issues. In the case of the capital market
community, the constraint also limits the definition of analyst understanding so that some
questionnaires are considered invalid, as well as the limits of the capital market community.
It is also unclear whether that is active in the stock, or includes investors who only deposit
funds through investment managers or reksa dana.

5
 Given some of the difficulties that occur when using all members of the population to
obtain parameters, an inductive statistics is developed, to draw general conclusions about the
entire population based on statistical information obtained from some populations commonly
referred to as sampe. To find out population parameters, for example 10-15 sample banks to
know the interest rate and credit terms, 5-10 mutual fund companies for investment returns,
2,000-3,000 corporate customers Sido Muncul, and 500 people from productive community
groups. By using a good sample, it is hoped that the conclusions drawn will be close to the
population.

In this chapter, we will first discuss the terms of population, samples and probability
samples, sampling methods, sampling errors, and the distribution of samples and the
boundary postulate.

11.2 Understanding Population and Sample

In the book I have discussed understanding of population and sample which is an

important aspect in studying inductive statistics. Population is a collection of all possible
people, objects, and other measures, which become the object of attention or collection of all
objects of concern. While the sample is a part of a particular population of concern.
Population and sample relationships can be described as follows.

The population can in fact be grouped into two parts, namely the limited population
and the infinite population. The limited population is a population whose element is limited
to size N. While the infinite population is a population undergoing a continuous process so
that size N becomes unlimited change in its value. The example of a limited population is in
the case of banking, which is the number of all banks there are 120 banks, the population of
the company reksa dana as many 82 companies. Meanwhile the sample of unlimited
population is the customer of Sido Muncul herbalism and the capital market community. The
number of customers of herbal medicine at any time can change, increase or decrease, as well
as in the case of investors in pasr moda. Another conto is the production of cars, the number
of car population at any time change because the production process runs continuously.
Therefore, in the case of unlimited population it is difficult to be able to record the members
of the population.

6
 The sample is part of the population. By using samples, we can obtain a measure
called statistics. In inductive statistics there is a tendency to make general conclusions based
on information from samples or statistics. In this case need to be distinguished understanding
of parameters and statistics. Parameters are the size of the population, while the population is
the size of the sample. In conditions where good sample selection is expected, the expected
statistics are expected to be equal or close to the same parameters.

Samples can be divided into two groups, ie probability samples and nonprobability samples.

"The probability sample is a sample selected in such a way from the population that each
member of the population has the same probability or opportunity to sample."

So in principle, if each member is given the same opportunity is called a probability sample.
In the case of banking sample selection, if from 120 banks have equal opportunity, that is
0.0083 (1/120), while in industry reksa dana there are 82 companies and each company has
the same chance, that is 0.0122 (1/82 ), then the probability sample is named. The probability
samples typically include random sampling, area sampling, and stratified.

The use of samples to draw conclusions from the population becomes increasingly
important, as often the use of the population is deemed less viable by considering the
following.

a. Collecting and contacting the entire population is often time consuming. In case to
know the rate of return on investment, it would be very inefficient to register from 82
companies reksa dana or 120 banks if to register 1 company reksa dana takes half a
day to research the product, the combination of products and risks, interest rates,
behind the investment, and the overall requirements of the credit, it will take 41
working days or more than one month for a company reksa dana and 60 days or 2
months for banking. You may also be able to imagine how long it will take to record
the population of artist Inul Daratista who is rated 35 or viewed by about 35 million
people? therefore, the use of samples will be more time efficient.
b. The cost of studying all the elements in the population is very expensive. If all the
elements of the population must be studied, then the costs incurred will be very
expensive. To find out the rating of an event such as Inul Daratista rating with a
population of 35 million, if each member takes 6 minutes, the required call Rp.750 /
member, or Rp.26.25 billion (750 x 35 million) . For example in the case of banking,
if each questionnaire has a cost of Rp.50.000, then it takes funds Rp.6.000.000.
Therefore, using a smaller sample of the population would be more cost-effective.
c. It is impossible to physically examine all elements of the population. It would be very
difficult to test an unlimited population overall. In the case of bank interest rates, the
rupiah exchange rate and the stock price will change at any time or even every
second. Therefore, it will be difficult to calculate the population because the number
is unlimited. Therefore, the use of samples such as daily, monthly, and yearly interest
rates is relatively easy in its statistical analysis. In the case of a physical nature such

7
 as populations of birds, fish, and others who are always moving, population
calculations are also difficult to do.
d. The destructive nature of testing. This property applies to an object population. In the
case of the standard test of food quality, it is impossible to test the quality of all food
products, because if all tested, the production will be tested and nothing is sold.
Therefore, to test the quality of food used production samples such as daily samples
or certain hours.
e. Based on these reasons, the use of the sample develops with the development of
inductive statistics, so that the conclusions of the samples are used for generalization
in the population.

11.3 Sampling Method

There are several methods that can be used to select samples from the population. Of these
methods, there is no "best" method for choosing the same objective, which provides an
opportunity to require elements or members of the population to be included in the sample.
Basically sampling methods can be grouped into two parts, namely probability samples
(probability sampling) and samples non probability (non probabilitiy sampling). The
probability sampling method is a method that gives equal opportunity to members of the
population to be sampled. Meanwhile, the non probability sampling method, in contrast, is
that not every member of the population has the same probability. This occurs because the
sample is taken with special consideration or a systematic sampling arrangement.
Schematically, the sampling method can be illustrated as follows.

Sampling Method

Probabilitiy Sampling Non Probabilitiy

1. Simple random sampling 1. Systematic sampling

2. Stratified random sampling 2. Quota sampling
3. Cluster sampling 3. Purposive sampling

8
11.3.1 Simple Random Sampling
Simple random sampling is defined as follows.
"A simple random sampling is a random sampling of the population regardless of the strata
present in the population and each member of the population has the same opportunity to
sample."

To describe a simple random method, used the case of mutual fund companies. In
2013 there were 82 mutual fund companies in Indonesia, taken as sample only 10%, ie 8.2 or
rounded to 8 companies. How to choose 8 companies from 82 mutual fund companies? A
simple random sample method ensures that 82 companies have the same probability. There
are two ways that can be done, namely:

1. Shake system. the system is the same as the arisan system, which is to make a paper
roll as much as 82 nuah and written each sheet of paper for one company. Rolls of
paper were put into a whisk and shaken until 8 samples were obtained. This method is
easy to do, but it is not practical to use for large populations, for example with
thousands of members, this method will be less efficient.

2. Using random random. to facilitate selecting samples especially for large populations,
can use random tables (see Appendix 1). this random table is derived from a random
process using a computer where the probability of a number from 0, 1 ..., 9 is the
same.to use a random table, it must be determined first starting point (starting point).
you are free to determine where the starting point is, or use the starting point
according to the random table. Following an example of random table usage and 8
sampling of 82 mutual fund companies (Table 11.1), where in this example a random
number table is drawn from The Rand Corporation. A million Random Digits with
100,000 Normal Deviates (New York; The Free Press, 1955)

At the starting point, you are free to start from anywhere, and for this case starting from the
1st row point of the 10th column fan with a random value of 88344, then proceed in sequence
following the row (to the right) or the column (down). in this example, the sequence is done
following the column (down). since the population is 82, only two digits of the random
number are used, and the value is below or equal to 82. Therefore, the number 88344 is taken
only 88, but this number is greater than 82, it is searched by row sequence . in the 10th
column the value is 88.89, and 96, the value above 82, this is not the sample number. next to
the shaded numbers there are 8 sample numbers, namely 72, 29, 68, 45,15, 58, 56, and 21.
this number is the sample number.

1 2 3 4 5 6 7 8 9 10

1 13962 70992 65172 28053 02190 83634 66012 70305 66761 88344

2 43905 46941 72300 11641 43548 30455 07686 31840 03261 89139

3 00504 46658 38051 59408 16508 82979 92002 63606 41078 86325

9
4 61274 51238 47267 35303 29066 02140 50867 39847 50968 96719

5 43753 21159 16239 50595 62509 61207 86815 29902 23395 72040

6 83503 51662 21636 68192 84294 38754 84755 34053 94582 29215

7 36807 71420 35804 44862 23577 79551 42003 58684 09271 68396

8 19110 55680 18792 41437 16614 83053 00812 16749 45347 88199

9 82615 86984 93290 87971 60022 35415 20852 02909 99476 45568

10 05621 26584 36493 63013 68181 57702 49510 75304 38724 15712

11 06936 37293 55875 71213 83025 46063 74665 12178 10741 58362

12 84981 60458 16194 92403 80951 80068 47076 23310 74899 87929

13 66354 88441 96191 04794 14714 64749 43097 83976 82381 72038

14 49602 94109 36460 62353 00721 66980 82554 90270 12312 56299

15 78430 72391 96973 70437 97803 78683 04670 70667 58912 21883

After obtaining a random number for the sample, it can be matched on the list of members of
the population, so as to know which mutual fund company is the sample. A list of the
population of mutual fund companies (mutual fund investment managers) can use data at
Kontan Data Center (http://pusatdata.kontan.co.id//reksadana/).

No Mutual Funds

1. AAA ASSET MANAGEMENT 42. MANDIRI MANAGEMENT

INVESTASI

2 AIM TRUST 43. MANULIFE ASSET

MANAGEMENT INDONESIA

3. AMCI MANAJEMEN 44. MAYBANK GMT ASSET

INVESTASI INDONESIA MANAGEMENT

4. ARCHIPELAGO ASSET 45. MEGA ASSET MANAGEMENT

MANAGEMENT

5. ASHMORE ASSET 46. MEGA CAPITAL INVESTAMA

10
 MANAGEMENT INDONESIA

6. AVRIST ASSET 47. MILENIUM DANATAMA

MANAGEMENT INDONESIA

7. AXA ASSET MANAGEMENT 48. MILENIUM DANATAMA

INDONESIA INDONESIA

8. BAHANA TCW INVESTMENT 49. MINNA PADI ASSET

MANAGEMENT MANAGEMENT

9. BATASA CAPITAL 50. MNC ASSET MANAGEMENT

10. BATAVIA PROSPERINDO 51. NARADA KAPITAL INDONESIA

ASSET

11. BNI ASSET MANAGEMENT 52. NET ASSET MANAGEMENT

12. BNP PARIBAS INVESTMENT 53. NIKKO SECURITIES INDONESIA

PARTERS

13. BRENT ASSET MANAGEMENT 54. NISP ASSET MANAGEMENT

14. CIMB PRINCIPAL ASSET 55. OPTIMA KHARYA CAPITAL

MANAGEMENT MANAGEMENT

15. CIPTADANA ASSET 56. OSO MANAGEMENT

MANAGEMENT INVESTMENT

16. CORFINA CAPITAL 57. PACIFIC CAPITAL INVESTMENT

17. DANAREKSA INVESTMENT 58. PANIN SECURITAS,TBK.

MANAGEMENT

18. EASTSPRING INVESTMENT 59. PARAMITRA ALFA SEKURITAS

INDONESIA

19. EKOKAPITAL SEKURITAS 60. PAVILLON CAPITAL

20. EMCO ASSET MANAGEMENT 61. PG ASSET MANAGEMENT

21. EQUATOR INVESTMENT 62. PHILLIP SECURITIES INDONESIA

22. EQUITY SECURITIES 63. PNM INVESTMENT

INDONESIA MANAGEMENT

23. EUROCAPITAL PEREGRINE 64. PRATAMA CAPITAL ASSET

SECURITIES MANAGEMENT

11
24. FALCON ASIA RESOURCES 65. PRIME CAPITAL
MANAGEMENT

25. FIRST STATE INVESTMENT 66. PROSPERA ASSET

INDONESIA MANAGEMENT

26. GANI ASSET MANAGEMENT 67. RECAPITAL ASSET

MANAGEMENT

27. GAP CAPITAL 68. RELIANCE MANAJER

INVESTMENT

28. GOLDMANY ASSET 69. RHB OSK ASSET MANAGEMENT

MANAGEMENT

29. HARVESTINDO ASSET 70. SAMUEL ASSET MANAGEMENT

MANAGEMENT

30. HENAN PUTIHRAI ASSET 71. SARIJAYA PERMANA

MANAGEMENT SEKURITAS

31. INDO PREMIER INVESTMENT 72. SCHRODER INVESTMENT

MANAGEMENT MANAGEMENT INDONESIA

32. INDOSURYA ASSET 73. SEMESTA INDOVEST

MANAGEMENT

33. INSIGHT INVESTMENT 74. SINARMAS ASSET

MANAGEMENT INTRU MANAGEMENT
NUSANTARA

34. INTRU NUSANTARA 75. SUCORINVEST ASSET

MANAGEMENT

45. INVESTINDO NUSANTARA 76. SURYA TIMUR ALAM RAYA

SEKURITAS

36. JAKARTA INVESTMENT 77. SYAILENDRA CAPITAL

37. JISAWI FINAS FUND 78. SYNERGY ASSET

MANAGEMENT COMPANY MANAGEMENT

38. KHARISMA ASSET 79. TIGA PILAR SEKURITAS

MANAGEMENT

39. KIWOOM INVESTMENT 80. TREASURE FUND INVESTAMA

MANAGEMENT INDONESIA

12
40. KRESNA ASSET 81. TRIMEGAH ASSET
MANAGEMENT MANAGEMENT

41. LAUTANDHANA 82. VALBURY CAPITAL

INVESTMENT MANAGEMENT MANAGEMENT

Based on the random number in the previous stage, it can be seen that the mutual fund
companies that are sampled are PT Ciptadana Asset Management, PT Equator Investment,
PT Harvestindo Asset Management, PT Mega Asset Management, PT Oso Investment
Management, PT Panin Sekuritas, Tbk., PT Realiance Investment Manager, and PT Sonroder
Investment Management Indonesia (Samples are sorted by sample number). Such sampling
process is called simple random by using random table.

EXAMPLE 11.1
Bank Indonesia wants banking performance in the first semester of 2013. For this purpose, 10
banks will be taken as sample. In the year, 2013, the number of ank there are 120 pieces.
Using a random table, what number is the sample, if the starting point, for example, starts
from the 3rd column, the second line.

Tabel 11.3 Contoh Pengacakan dengan Tabel Angka Mengikuti Baris

Titik awal --- --- --- --- --- dst

1 2 3 4 5 6 7 8 9 10

1 13962 70992 65172 28053 02190 83634 66012 70305 66761 88344

2 43905 46941 72300 11641 43548 30455 07686 31840 03261 89139

3 00504 46658 38051 59408 16508 82979 92002 63606 41078 86325

4 61274 51238 47267 35303 29066 02140 50867 39847 50968 96719

5 43753 21159 16239 50595 62509 61207 86815 29902 23395 72040

6 83503 51662 21636 68192 84294 38754 84755 34053 94582 29215

7 36807 71420 35804 44862 23577 79551 42003 58684 09271 68396

8 19110 55680 18792 41437 16614 83053 00812 16749 45347 88199

9 82615 86984 93290 87971 60022 35415 20852 02909 99476 45568

10 05621 26584 36493 63013 68181 57702 49510 75304 38724 15712

13
11 06936 37293 55875 71213 83025 46063 74665 12178 10741 58362

12 84981 60458 16194 92403 80951 80068 47076 23310 74899 87929

Answer:

1. For these purposes the first step is to register and assign the number to 120 banks.

2. The starting point on the 3rd column of the 2nd row we get a random value of 72300,
and since the population is 120, only three digits are used, that is 723, and since it is
over 120, it is not the population number

3. If forwarded by line and value below 120, got 10 sample number, that is
116,76,32,5,21,92,29,56,69.

4. If there is known list and population number of banking, for example from directory
of Banking Indonesia at www.bi.go.id, then 10 bank which become sample is Bank
Rakyat Indonesia Agroniaga, Tbk; Bank BNI Syariah, Bank SBI Indonesia; Bank
Muammalat Indonesia; Bank OCBC NISP Tbk; Bank Panin Syariah; Bank Jabar
Banten Syariah; Bank DBS Indonesia; JP. Morgan Chase Bank, NA; and BPD North
Sulawesi. We can assess the performance of the 10 banks through asset value, ktedit
given, public funds, operating profit, current year earnings, value of CAR, NPL and
LDR as a measure of the performance of the banking sector.

11.3.2 Structured Random Sampling Withdrawal

The drawing of structured acaak samples is known as stratified random sampling. this
sampling is used for populations that have members or non-homogeneous elements.
Structured random definitions and samples are as follows

"Structured random sampling is done by dividing the population members into several
subgroups called strata, then a sempel is selected from each stratum."

The stratification process can be described as follows.

14
In the unequal population there are 24 members (5 round, 7 triangle, and 12 boxes). In the
process of sampling directly, one of the tjga shapes and colors can not be drawn. Therefore,
the stratification method is used to improve the estimation of the characteristics of the
population by first grouping the Population on the same element to obtain a high degree of
homogeneity in each group or strata. In this example the population is grouped into three
groups of rounded group with 5 members, triangle group with 7 members, and 12 lotak
cotton groups. To take samples from each group can use a simple random method.

How to determine the number of samples on a structured method? To find the number of
samples there are two ways, namely structured and disproportionate structures. Proportional
sampling requires a proportional sample of each stratum. for sample monitoring. of the 24
members of the population will be taken 10 samples, then the number of samples per stratum
as follows.

Jumlah
Stratum Kelompok Proporsi Jumlah sampel per stratum
Anggota
N1 = (N1/N) x n = 0,21 x 10
1 Bulat N1 = 5 N1N = 0,21
=2
N2 = (N2/N) x n = 0,29 x 10
2 Segitiga N2 = 7 N2N = 0,29
=3
N3 = (N3/N) x n = 0,50 x 10
3 Kotak N3 = 12 N3N = 0,50
=5
Jumlah
N = 24
Total

15
From the table above shows that the number of samples per stratum is based on
The proportion of the percentage of each stratum to its total. Determination of the number of
samples
which is known as proportionate stratified random sampling method.
In addition to a structured proportional structural drawing method, there are also ones
Nonproportional, Nonproportional method is used to determine the number of samples
the population is structured, but the amount is not or less proportional, As
example in cases where round is only 1, number of squares 4, and number of segitlga 19
so the number of samples per stratum is:

Stratum Kelompok Jumlah Proporsi Jumlah Sampel per Stratum

Anggota (Dibulatkan)
Proporsional Nonproporsional
1 Bulat 1 0,04 0,04 x 10 = 0 1
2 Kotak 4 0,17 0,17 x 10 = 2 2
3 Segitiga 19 0,79 0,79 x 10 = 8 7
Jumlah 24
10
Total

In the round stratum it is seen that the number of samples is 0, therefore it can be policy is
taken that a set of strata must be represented, for example round 1 sample, frog 2 sample, and
triangle 7 samples. Such a method is known as nonproportionate stratified random side.

Structured sampling in some cases has the advantage of being able to reflect more accurately
the population parameters than the random method as early as possible. This happens when,
in the case, the amount of a strand of sanqat is small, it will not sampled using a simple
random method. With a structured method, then each strata has a minimum number of
samples, so that all structures can represented so that the results are more accurate.

Example 11-2
Seorarng investors want to know the performance of the company in the financial sector
located in Indonesia Stock Exchange (BEI) in 2008. The financial sector that exist in BEI
consisting of 27 banking companies, insurance and financing companies as many as 19
companies. etek companies there are 9 companies, and financial companies other sebanvak 5
companies. If 15 samples will be taken with the sample the method is structured
proportionally, then how many samples each group?

Answer
To know the number of samples of each group, it must be known first
the proportion of each group to the total firm. After getting the value of proportion. then
we can know the number of samples by multiplying the proportion by the number of samples.

Jumlah Sampel per

Kelompok
Jumlah Perusahaan Proporsi Kelompok
(Subsektor)
(Dibulatkan)
Perbankan 27 0,44 0,44 x 15 = 7
Asuransi dan 19 0,31 0,31 x 15 = 5

16
pembiayaan
Efek 9 0,15 0,15 x 15 = 2
Keuangan lainnya 6 0,10 0,10 x 15 = 1
jumlah 61 15

11.3.3 Cluster Sampling

Cluster withdrawal is a technique of selecting samples from groups of small units (clusters)
of a relatively large and widespread population, Members in each cluster are not
homogeneous, in contrast to structured drawing members. Cluster members are similar with
members of the population. but in smaller quantities. Difference between population stratum
and cluster in the sampling can be described as follows.

Populasi

Sampel terstruktur Sampel Kluster

In structured samples, from which group unit or stratum units are created Each stratum has
the same member or element. In the cluster sample it appears that. The formation of a unit is
not based on member equality, but a unit Groups with smaller number of members, for
example, if the assessment will be held each semester against a financial institution. In the
case of structured samples, then the whole institutions in Indonesia are grouped, for example
into a banking institution (round), insurance and financing institutions (triangles), securities
agencies (pentagon), and institutions other finance (rectangular). Each institution has a
homogeneous member, However, in the case of different cluster samples, the units become
regions, eg Cluster I Kalimantan region (with 21 financial institutions, 10 banks, 6 insurance
and financing, 3 securities, and 2 other financial institutions), Cluster Il for Sumatra (with 17
institutions finance; 6 banks, 6 insurance and financing, 3 securities, and 2 other financial
institutions), and Cjuster Ill Java area (with 20 financial institutions; 8 banks, 7 insurance and
financing, 2 securities, and 2 other financial institutions).

17
 How to samples after pang-cluster-an? Sample selection done by simple random method as
has been stated previously.
The benefit of this cluster method is to reduce the cost of sample population sampling
spread over too large a geographical area For financial institution prosecutions

All of Indonesia, created 3 clusters then mamilih 1 of 3 clusters by random method

Simple. If we use a random table (as in simple random instance), then Cluster 1 is selected as
a sample. Therefore, enough research is done on the Cluster
1. Selection of members from Cluster 1 to be used as research samples can also be done
with a simple random method using simple random tables.

Strata-1 Strata-2 Strata-3 Strata-4

Diambil secara acak Diambil secara acak Diambil secara acak Diambil secara acak
sesuai jumlah sampel sesuai jumlah sampel sesuai jumlah sampel sesuai jumlah sampel
di strata-1 di strata-1 di strata-1 di strata-1

Cluster-1 Cluster-2 Cluster-3

Cluster diambil secara acak misalnya terpilih cluster-1

Cluster-1 sebagai sampel

Seluruh anggota cluster yang terpilih

sebagai sampel (cluster-1) menjadi
unit pengematan

18
11.3.4 Systematic Random Sampling

Systematic sampling is a sampling technique based on the order of members populations that
have been assigned a sequential number that can be defined as follows.

"Sampling is said to be systematic when every or a member in populations arranged in a

certain way-alphabetically. from big to small or vice versa - then a random point is chosen
every K-member of the population is selected. As a sample. "

For example, apebila will be selected 5 mutual fund companies from 82 companies existing
mutual funds, then the company to be a sample using systematic methods. Some of the
necessary steps are:

a. Provide a sequence number, for example based on alphabetical. from A to Z

b. The population number is 82 and the sample number is 5, then the distance between
the samples is 16
c. The sample numbers were 1, 17, 33, 49, and 65 (each sample spaced systematically ie
16). Based on the table of mutual fund companies (Investor, June 2008), can be
known the names of sample companies, namely AAA Asset Management Danareksa
Investment Management, Insight Investments Management. Minna Rice Asset
Management, and Prime Capital (See Table 1 1.2).

11.3.5 Quota Sampling (Quote Sampling)

Quota sampling is the return of a characteristic population certain to the desired amount or
quota. The purpose of quota sampling is to improve all components in the population. The
purpose of representation This is so that the characteristics in the population can be and are
represented and illustrated with good. For example, if will be conducted research on the
performance of the banking from population of 120 banks and determined quota is 50 banks,
If collection data has not reached 50, then the research is considered unfinished and research
should be continued until it reached the quota. If in the study should describe representation
of the population from banks, then each group must have a quota. banks can be grouped into
three sections, namely 15% of large bank assets, 36% medium asset, and 49% small asset.
The quota of each group became 7 major banks (15% X 50), 18 medium banks (36% x 50),
and 25 small banks (49% x 50). Sampling each group must conform to a predetermined
quota.

11.3.6 Purposive Sampling (Purposive Sampling)

purposive sampling is the withdrawal of sanpel with certain consideration. Such

considerations are based on research interests or research, Withdrawal with purposive
samples divided into two ways, namely (a) convenience sampling, ie sampling based on the
researcher's desire in accordance with the research of the research. and (b) judgment
sampling, ie the sampling based on the assessment of the characteristics sample members
tailored to the study. This method is usually done for qualitative research, For example,
research on Bedouin culture, then a researcher will choose the sample intentionally, ie
Bedouins; research about flour, then the sample that was purposely selected was PT Bogasari
on the grounds or judgment that PT Bogasari is the largest and dominant company within
flour industry with more than 40% market share.

19
11.4 Sampling Error (sampling Error)

Inductive statistics relate to the effort to draw conclusions about the population based on the
sample analysis. Therefore, the indicator of the sample, ie the statistic, is expected to be the
same as the population indicator, ie the meter. Attempts to get closer the statistical value with
the meter is done by obtaining the appropriate sample and certifying each member of the
population. Several methods can be used, ranging from simple random, structured, clustered,
systematic, purposive, to obtain statistical values that approximate parameters. However,
since the sample size is only a percentage of the population, it can not be avoided that neither
the average value of the count nor the standard deviation of the sample will be equal to the
average value of the count and the standard deviation of the population. The difference
between the statistical value and the meter is known as the sampling error (smpling error).
Misconceptions of the sample dancers are devised as follows.

"Sampling error is the difference between the sample statistical value and the paramenter
value of the population"

Example 11-3

Bank Indonesia wants to see the performance of 5 regional banks (BPD) by looking at the
profit earned per December 2012 by taking samples of only 2 banks. Calculate sample errors
that may occur as a result of the sampling process?

Bank BPD Value of 2012 (In Millions)

Bank DKI 339.284
BPD Yogya 102.631
BPD Jateng 563.106
BPD Jatim 724.639
BPD Sumut 421.776

Answer:

Bank Profit value Combination amount Average Sample Error

(X) ( X -𝜇)
Bank DKI 339.284 1. DKI 441.915 221.958 -209.330
Yogya
BPD Yogya 102.631 2. DKI 902.390 451.195 20.908
Jateng
BPD Jateng 563.106 3. DKI 1.063.923 531.902 101.057
Jatim
BPD Jatim 724.639 4. DKI 761.060 380.530 -49.757
Sumut
BPD Sumut 421.776 5. Yogya 665.737 332.869 -97.419
Jateng
6. Yogya 827.270 413.638 -16.652

20
 Jatim
7. Yogya 524.407 262.204 -168.084
Sumut
8. Jateng 1.287.745 643.873 213.586
Jatim
9. Jateng 984.882 492.441 62.154
Sumut
10. Jatim 1.146.415 573.208 142.921
Sumut
𝜇 = ∑ 𝑋/5 403.287

1. Keep in mind that there are 5 BPD Banks and as a sample in the select 2 banks alone
so there will be a combination of 10 samples remember that the formula combination
is

𝑵! 𝟓! 𝟓!
𝑪𝑵
𝒏= = 𝟐!(𝟓−𝟐)! = 𝟐!𝟑! = 𝟏𝟎
𝒏!(𝑵−𝒏)!

2. Having found all combinations then it is necessary to calculate the average of any possible
combinations () and the mean of the population (μ)

3. After finding the average earnings as per December 2012 (in millions of rupiahs) from the
sample and population, it can be calculated sampling error with the formula (-μ). In the table
can be seen the value of sampling error in combination 1 is -209.330; combination 2 for
20,908; combination 3 of 101,657; and so on up to a combined 10 of 142,921.

Example The above shows that by using the sample, we can find the possibility of sampling
error. The sample results are used to estimate the parameters of a population. If the
combination of 1 and 7 is selected, the sample error is small, otherwise if the selected
combination of 8 and 10, then the sample error is quite large. The next question is how the
level of confidence in the results of the conclusions from the sample to predict the population
parameters about the performance of the banking and the effectiveness of an ad with the
model and other problems. To answer that question, it should be understood by constructing a
sample distribution (sampling distribution) and the average sample count (sample means)

21
11.5 Distribusi Sampel Rata-Rata dan proporsi

Penarikan kesalahan sampel dapat di ketahui apabila rata-rata hitung sampel sangat bervariasi
dari sampel satu kesampel lain. Selain itu terdapat juga permasalahan lain, yaitu apakan nilai
rata-rata sampel (X) dan standar pariasi sampel (sd) dapat mencerminkan dan sesuai dengan
nilai rata-rata populasi (𝜇) dan stansdar deviasi populasi (𝜎). Untuk sebuah populasi dapat
dipastikan bahwa hanya ada nilai rata-rata dan standar deviasi tunggal, sehingga di sebut
dengan deterministic variable.

As for the sample, there are many combinations of samples in which each sample
combination has a statistical value. Since the sample has many statistical values of each of its
combinations is pariabel of a random nature, the statistical values can have a probability
distribution.

The sample distribution of the calculated mean can be defined as follows

"The sample distribution of the sample hit rate is a probability distribution comprising all the
probability of the sample count of a given sample size selected from the population, and the
probability of occurrence in relation to each sample count average."

Example 11-4

To see whether an asset is considered productive, a return on assets (ROA) is usually used.
Here are the ROA until December 2012 from some relatively large banks in Indonesia

Bank ROA (%)

BRI 5,15
MANDIRI 3,55
CIMB NIAGA 3,11
DANAMON 3,18
BII 1,49

Based on the above data

a. Calculate what the average value of the population is

b. Calculate the average value of the samples if taken 2 banks from 5 existing banks as
samples! What is the conclusion of the average value of the population from the
sample mean?
c. Make a sample distribution of the calculated average and make the polygon diagram
for the population and sample?
d. Calculate the standard deviation for the population and sample!

22
Answer

a. The average value of the population

∑ x 5,15 + 3,55 + 3,11 + 3,18 + 1,49 16,48

b. 𝜇 = = = = 3.30
𝑁 5 5

To know the average value of the sample count, it must be known that all possible
combinations of 2 sample sizes without recovery for each sample are taken. For that we need
to know first erapa combination, the average value of each sample of the average nilani each
sample.

1) Combination
𝑁! 5! 5!
𝐶𝑛𝑁= = 2!(5−2)! = 2!3! = 10
𝑛!(𝑁−𝑛)!

2) Perhitungan rata-rata dari setiap sampel

No Combination Combination ROA

Count Average ( X )
1 BRI-MANDIRI 5,15 + 3,55 (8,7/2) = 4,35
2 BRI-CIMB NIAGA 5,15 + 3,11 (8,26/2) = 4,13
3 BRI-DANAMON 5,15 + 3,18 (8,33/2) = 4,17
4 BRI-BII 5,15 + 1,49 (6,64/2) = 3,32
5 MANDIRI-CIMB NIAGA 3,55 + 3,11 (6,66/2) = 3,33

No Combination Combination
Count Average ( X )
ROA
1 MANDIRI-DANAMON 3,55 + 3,18 (6,73/2) = 3,37
2 MANDIRI-BII 3,55 + 1,49 (5.042) = 2,52
3 CIMB NIAGA-DANAMON 3,11+ 3,18 (6,29/2) = 3,15
4 CIMB NIAGA-BII 3,11 + 1,49 (4,60/2) = 2,30
5 DANAMON-BII 3,18+ 1,49 (4,67/2) = 2,34

3) The average value of the sample count average

1
x = ∑x
𝐶𝑛𝑁

23
 x= 1
(4,35 + 4,13 + 4,17 + 3,32 + 3,33 + 3,37 + 2,52 + 3,15 + 2,30 + 2,34 ) = 3,30
10

From problem a, it is known that the average value of population hit (μ) = 3.30 and from
problem b, it is also known that the mean value of the sample count () = 3.30, so it can be
concluded that the pata menter value is equal to the value statistics.

c. c. Population Distribution and Sample from Average Value of Count

internal Middle Population Sample

Value Frequency Probability Frequency probability
1,49-2,22 1,9 1 (1/5) = 0,20
2,23-2,96 2,6 3 (3/10) =0,30
2,97-3,70 3,3 3 (3/5) = 0,60 4 (3/10) =0,40
3,71-4,44 4,1 3 (3/10) =0,30
4,45-5,18 4,8 1 (1/5) = 0,20
Amount 5 1,00 10 1,00

Here is the probability distribution in the form of a polygon..

Population Sample
0,7
0,7
0,6
0,6
0,5
0,5
0,4
0,4
0,3
0,3
0,2
0,2
0,1
0,1
0
0

1,9 3,33 4,8 1,9 2,6 3,3 4,1 4,8

d. The population standard deviation is formulated as follows.

24
 (𝑋 − 𝜇)2
𝜎 = √∑ 𝑁

Standard Deviation

X (X - µ) (X - µ)

5,15 1,85 3,42

3,55 0,25 0,06

3,11 -0,19 0,04

3,18 -0,12 0,01

1,49 -1,81 3,28

∑X = 16,48
∑(X - µ)² = 6,81

𝜎=
µ = 16,48 = (𝑋−𝜇)2
6,81
3,3 √∑ 𝑁 =√ = 1,17 5

Standard deviation (variance) of the sample mean

1
𝑆(𝑋̅) = √ ∑(𝑋̅ − 𝑋̿)²
𝐶𝑛𝑁

𝑋̅ (𝑋̅ − 𝑋̿) (𝑋̅ − 𝑋̿)²

4,35 1,05 1,111

4,13 0,83 0,696

4,17 0,87 0,755

3,32 0,02 0,001

3,33 0,03 0,001

3,37 0,07 0,005

2,52 -0,78 0,602

25
3,15 -0,15 0,023

2,30 -1,00 0,992

2,34 -0,96 0,924

∑(𝑋̅ − 𝑋̿)² =5,11

∑𝑋̅ = 32,96

1 5,11
𝑆(𝑋̅) = √ 𝑁
∑(𝑥̅ − 𝑥̿ )2 = √
µ = 32,96/10=3,30 𝐶𝑛 10
= 0,71

it is seen that between the standard deviation of the population (-1.17) and the mean sample
(s-0.71) is not the same. Thus, the relationship between the standard deviation of the
population and the sample is not the same as the relationship between the mean value of the
population and the sample.when viewed from the average standard deviation value of the
sample which is relatively smaller than the population, it can be concluded that the average
member value of the sample is relatively more centered on the middle value than the
population. This occurs because the population spread is wider than the sample. However,
there is a correlation between the mean standard deviation of the sample and the standard
with the standard deviation of the population for both limited and unlimited populations.

The relationship between s and 𝜎 for a finite population is expressed as follows

𝜎 𝑁−𝑛
𝑠= √
√𝑛 𝑁 − 1

where N is the number of population members and n the number of sample members

The relationship between s and 𝜎 for an unlimited population is expressed as follows.

𝜎
𝑠=
√𝑛
Example 11-5

The standard deviation of the sample (s) if the population standard deviation (𝜎) is known is
1.17, the number of population members (N) is 5 and the number of sample members (n) is 2

Answer

A. the number of samples is 5, then included in the limited population

B. the relationship between s and 𝜎 for a limited population

26
 𝜎 𝑁−𝑛 1,17 5−2
𝑆(𝑋̅) = 𝑛 √ 𝑁−1 = √ = 0,83𝑥0,87 = 0,71
√ √2 5−1

so, the standard deviation of the sample can be known by not having to calculate one by one
from the sample combination

11.5.1 Using Average Sample Distribution

By knowing the mean value of the sample (𝑥̅ ) and the standard deviation (s) of a sample and
by transforming the equation Z (see book 1-chapter 9), it is easy to know the probability of an
event. This can be done because the average sample distribution is the normal distribution
with the mean (𝑥̅ ) and standard deviation (s) can be written X~N (𝑥̅ , 𝑠). The normal mean
sample distribution form can also be seen from the sample polygon graph and the normal
population. Thus the value of Z can be formulated as follows.

𝑋̅ − 𝜇
𝑍=
𝑠
Z also has a normal distribution but with a mean (𝑍̅) = 0 and standard deviation (S (z)) = 1 or
written z-N (0,1). From Example 11-4, the values Z, 𝑍̅and S (z) can be calculated as follows.

(𝑋̅) (𝑋̅ − (𝑋̅ − 𝑋̿)² 𝑋̅ − 𝜇 Z-z (Z – Z)²

𝑍
𝑋̿) 𝑠

4.35 1.05 1.111 1.47 1,47 2,17

4.13 0.83 0.696 1.17 1,17 1,36

4.17 0.87 0.755 1.22 1,22 1,48

3.32 0.02 0.001 0.03 0,03 0,00

3.33 0.03 0.001 0.05 0,05 0,00

3,37 0.07 0.005 0,10 0,10 0,01

2,52 -0.78 0,602 -1,09 -1,09 1,18

3,15 -0.15 0,023 -0,21 -0,21 0,04

2.30 -1,00 0,992 -1,39 -1,39 1,94

2,34 -0,96 0,924 -1,34 -1,34 1,81

∑𝑋̅ = 32,96 ∑(𝑋̅ − 𝑋̿)

2
∑𝑍 = 0 ∑(𝑍 − 𝑍)2 = 10
= 5,11

27
𝜇 = 𝑋̿ = 3,30 𝑆(𝑋̅) = 0,71 0 10
𝜇𝑧 = 𝑍̅ = S(z)=√(10) = 1
10
=0

If the value of Z is known by using the table area under the normal curve, we can know the
probability of an event.

Example 11-6

Throughout the year 2013 there are 31 new issuers listed on the Indonesia Stock Exchange
(IDX). The 31st issuer listed on IDX throughout 2013 is the largest herbal medicine company
in Indonesia, namely PT Industri Jamu and Pharmaceutical Sido Muncul Tbk., With SIDO
stock code. Observations after 23 days, ie seati 18 December 2013 - 24 january 2014, SIDO
stock price continues to fluctuate with the trend of increasing prices. The closing stock price
had decreased to 680 and had risen to 815. During the observation, the last 23 days SIDO
stock average 750 with a standard deviation of 45.5. What is the chance of SIDO stocks
falling below 725, and how many chance the price rises above 800?

Answer:

a. Given n = 23; 𝑥̅ = 750 and 𝜎= 45.5

b. For the mean population and sample is equal = µ = 750
c. To find the value of s using the formula for an unlimited population, this is due to the
process of stock transactions running continuously
𝜎 45,5
𝑠= = = 9,5
√𝑛 √23
d. After the standard deviation and middle values are known, Z can be found and
probability
𝑥−𝜇
𝑧=
𝑠
Z for X <725 = (725-750) / 9.5 = The value of Z -2.63 in the wide table below the normal
curve shows the value of 0.4957, so the probability of the stock price below 725 is 0.5 -
0.4957 = 0 , 0043 or 0.43%

Z for X> 800 = (800 - 750) / 9.5 = 5.26. The value of Z = 5.26 in the table below the normal
curve does not exist, because the maximum Z value is 2.99. Therefore, the probability can be
used the highest value, ie Z 3.99 with probability 0.49997. So the probability of stock price
above 800 is 0,5-0,49997 = 0,00003 or 0,003%.

The calculation results show that the probability of Sido Muncul stock price to fall below 725
and increase above 800 is very small. Sido Muncul stock price will be relatively stable at its

28
middle value, which is 750. Therefore, for investors who simply take gian term, this stock is
not too suitable.

11.5.2 Proportional Sampling Distribution

In addition to the average sampling distribution, there is also a distribution of proportional

sampling. The word proportion shows the percentage (proportion x 100%) of a part or
element to the total of parts or elements. In everyday life, we often hear things like the
number of employees of the production of 70% are men. The value of 70% shows the
proportion of male employees to the total number of employees is 70%, so it can be
concluded that the remaining 30% is female employees. Other notes such as banking assets.
According to Investor, June 2008, the number of banks with assets above 50 trillion is
estimated at 10%, between 25-50 trillion of 50% and below 1 trillion of 32%. In the field of
agribusiness for example there is also a proportion such as fruit production in the country
reached 70%, while the remaining 30% is imported fruit and many more examples of
proportions. Basically in every study, the problem of proportion can always be done with
attention to the purpose of research so that the determination of the proportion can be based
on a particular character.

As with the average sample distribution, is the statistical value of the mean value of the count
and the standard deviation of the sample and the population having the same characteristics,
ie p equal P, as well as how the sample deviation standard (Sp) is. For a discussion on this
proportion, the symbol P is for the population parameter and p symbol for the sample
standard. To better understand this issue can be seen in the following example.

Example 11-7

Here are the stock of 5 companies on BEI on November 11, 2013

Insurance Company Stock Price Sheet

Asuransi Bina Dana Arta Tbk. (ABDA) 4.500

Asuransi Ramayana Tbk. (ASRM) 1.040

Lippo General Insurance Tbk. (LPG) 3.200

Maskapai Reasuransi Ind. Tbk. (MREI) 2.750

Panin Insurance Tbk. (PNIN) 720

From above count example.

a. What is the proportion of the insurance company to buy it’s shares (P), if the investor
will buy the stock at a price with a stock price above 2000?

29
 b. What is the middle value of the proportional sampling distribution (𝑃̅𝑝)if the sample
is required above the stock price 2000 if the value of P equals (𝑃̅ 𝑝)?

c. What is the standard deviation (𝑆𝑝 ) of the sample distribustion of that proportion?

Answer:

a. Proportion of insurance company with stock price with stock price above 2000.
Number of companies with stock price 2000 is ABDA, LPG, and MREI. So there are 3
companies from 5 insurance companies. Therefore, the proportion (P) is 3/5 = 0,6.

b. To Calculate P, it should be know : (1) the number of sample combinations (2) the
proportion of stock prices above 2000, and (3) calculate as follows.

1
𝑃̅𝑝 = ∑𝑃
𝐶𝑛𝑁

1) Number of combination of sample

𝑁! 5! 5!
𝐶𝑛𝑁 = = = = 10 𝐶𝑜𝑚𝑏𝑖𝑛𝑎𝑡𝑖𝑜𝑛𝑠
𝑛! (𝑁 − 𝑛) 2! (5 − 2)! 2! 3!

2) Calculates the proportions by their combination

No Sampel Combination Stock Price > Proportion Of

2000 Sample

1 4.500 3.200 3.200 2 2/3

2 4.500 2.750 2.750 2 2/3

3 4.500 1.040 720 1 1/3

4 4.500 3.200 2.750 3 3/3

5 4.500 3.200 720 2 2/3

6 4.500 2.750 720 2 2/3

7 1.040 3.200 2.750 2 2/3

8 1.040 3.200 720 1 1/3

9 1.040 2.750 720 1 1/3

10 3.200 2.750 720 2 2/3

So, the average value of the sampel distribution of proportions is

30
 ̅̅̅̅̅̅̅̅̅̅̅
1 1
𝑃̅𝑝 = 𝑁 ∑ 𝑝 = ( ) = 0,6
𝐶𝑛 10

From answer a and b it is seen that the value of P = p = 0,6. So the average value of the
sample with the average the average value of the population for the sample of the proportion
of magnitude is also the same.

C. To find the standard deviation of samples from sample proportions used the following
formula.

1
𝑆𝑝 = √𝐶 𝑁 ∑(𝑝 − 𝑃̅ 𝑝)2
𝑛

For Comvenieence, a table can be made as follows, and remember that the value p = 0,6
becomes 1-0,6 = 0,4.

Stock Price > 2000 Proportion Of (𝑝 − 𝑃̅ 𝑝) (𝑝 − 𝑃̅ 𝑝)2

Sample (p)

2 2/3 = 0,67 0,07 0,0049

2 2/3 = 0,67 0,07 0,0049

1 1/3 = 0,33 -0,27 0,07

3 3/3 = 1,00 0,40 0,16

2 2/3 = 0,67 0,07 0,0049

2 2/3 = 0,67 0,07 0,0049

2 2/3 = 0,67 0,07 0,0049

1 1/3 = 0,33 -0,27 0,07

1 1/3 = 0,33 -0,27 0,07

2 2/3 = 0,67 0,07 0.0049

∑(𝑝 − 𝑃̅ p)2 0,40

1 √0,04 = 0,2
√𝐶 𝑁 ∑(𝑝 − 𝑃̅ 𝑝)2
𝑛

So, the standard deviation for the proportion sampel is 0,2 which can be interpreted that the
values in the sample spread 0,2 from the mean value of 0,6.

31
Standard deviation of distribution of proportional sampling can be calculated also by other
means, that is :

𝑃(1 − 𝑝) 𝑁−𝑛
𝑆𝑝 = √ 𝑥√
𝑛 𝑁−1

0,6 (1 − 0,6) 5−3

𝑆𝑝 = √ 𝑥√ = 0,2
3 5−1

The result of the standard deviation value using this formula is the same as the previous
formula, is 0,2.

For an unlimited population, or small sample size, ie less than 5% of the population, then the
standard deviation can be formulated as follows.

𝑃 − (1 − 𝑃)
𝑆𝑝 = √
𝑛

This occurs because the value of (N-n)/(N-1) is close to 1, so it does not affect the standard
deviation value.

11.5.3 Using Sample Distribution of Propottions

As with only the average sample distribution, in the proportion sample can also be used the
normal distribution approach.This is based on the fulfillment of terms, namely the value of
n.P and n (1-P) greater than, then it will be close to normal. Therefore, by khowing the
proportion of the sample (p) and population (P) and standar deviation sample (sdp)1, will be
known the value of Z. the value of Z for the distribution of proportions can be formulated as
follows.

𝑃 − ̅̅̅̅
𝑃𝑝 𝑝−𝑃
𝑍= =
𝑆𝑝 𝑆𝑝

Example 11-8

Informatics is a relatively famous information technology education institute in Jakarta. every

day, the agency receives a phone call that displays the program informatics as many as 100
people and 20% of them, that is 20 people then register. calculate the probability of (a) more
than 30 people registering and (b) less than 15 sign up!

32
Answer:

a. Proportion of registering P = 0,2 (20/100). Sehingga proporsi tidak memilih (1-P) = 1

– 0,2 = 0,8.

b. Calculate the standard deviation and the mean value

𝑃(1−𝑃) 0,2.0,8 0,16

Sp = √ =√ = √ 100 = 0,04
𝑛 100

c. Calculate the value of Z for probability above 30 or 30/100 = 0.3

(𝑝 − 𝑃) (0,3 − 0,2)
𝑍= = = 2,5
𝑆𝑝 0,04

The value of Z = 2.5 has a probability of 0.4938, so the probability above 03 is P (p> 0.3) = P
(z.2.5) = 0,5 - 9,4938 = 0.0062 or 0.62%. so, the probability of choosing lebh from 30 people
is 0.62%.

d. Calculate the value of Z for probability less than 15 or 15/100 = 0.15

e. Nilai Z sebesar -1,25 yang memiliki probabilitas 0,3944 sehingga kurang dari 0,15
adalah P (p.0,15) = P (Z, -1,25) = 0,5 - 0,3944 = 0,1056 atau 10,56%. jadi,
probabilitas yang mendaftar kurang dari 15 orang adalah 10.56%, dengan demikian
peluang lebih dari 15 orang adalah sebaliknya, yaitu 100 - 10,56% = 89,44%

Case!!!
Assess the case: Rich People RI Allocate 30% of Investment Portfolio

The share of investment in the property sector is predicted to remain high in line with the
increasing number of rich people in Indonesia. The consultancy firm Knight Frank through
The Weakth Report 2014 shows that 61% of rich Indonesians believe the prospects for wealth
creation in 2014 are still positive. The research results also mention the average wealth of
Indonesia makes the property sector 30% of its investment portfolio. As many as 50% of the
rich increase the investment portion of the property sector in 2013. In addition, the research
states that an average of 24% of wealth of super rich Indonesians is an investment in the form
of houses, both as a first residential or second occupancy. Meanwhile, 17% of rich people
consider or have plans to replace the house for business reasons, quality of life, and
education. Meanwhile, the same research results show that 26% of rich people in Indonesia
have plans to buy a second home within the next 12 months.

Source: http;//m.bisnis.com/property/read/20140321/, 21 Maret 2014, Editor: Linda Teti

Silitonga

33
11.6 Distribution of sample of average difference and proportion

The sample distribution of the mean difference and the proportion is a distribution of two or
more populations. the purpose of the distribution of the difference is to distinguish the
characteristics of one population with another population, whether the same or different. In
the following illustration, it can be seen that there are two populations, one with an average
value and a standard deviation. then take one sample of size n, with average value and
standard deviation x1 sd x1. population two had mean and standard deviation values then
samples were taken with size n2 with mean and standard deviation values x2, sd x1.

Population1 samples
sized n1 x1
𝜇1,𝜎2 sd x1 Whether

X1 – x2 =
𝜇1−𝜇2
Population 2 Sample 2
berukuran
𝜇2,𝜎2
n2 x2 s x2

after obtaining the parameter and statistical value of the two populations. the question is
whether these two populations have the same parameters and statistics ( X1 – x2 = 𝜇1−𝜇2 )

on the basis of every n-sized sample taken from the population is a random variable and tends
to approach normally. therefore, the distribution of the average difference and proportion also
follows the norms distribution pattern. The average value and the proportion have the average
value and standardeviation as follows.

34
* the average value of the sample distribution of the average difference of x1 - x2

𝑋𝑋1−𝑥2= 𝜇1−𝜇2

𝑝𝑝1−𝑝2= 𝑝𝑝1−𝑝2=𝑃
1−𝑃2

* the default value of the deviation of the sample distribution of the average difference of x1 -
x2 )

𝑆2𝑥1 𝑆2𝑥2
𝑆𝑥1−𝑥2= √𝑆2𝑋1+𝑆2𝑥2 = √ +√
𝑛1 𝑛2

*while the standard deviation of sample distribution of data difference

(𝑋1−𝑋2) (𝜇1− 𝜇2)
Z= −
𝑆𝑥1−𝑋2

Example 11-9

today many companies that require the means of infortation technology in rangkamembangun
good corporate governance welcomes the needs of many companies, so also develop
companies provide IT services. in 2013 there are 7 companies subsector computer services

35
and other devices contained in the bell, the PT Astra graphia TBK. (ASGR). as a fellow
company that there dibel, is expected to the 7 companies have kineja samabaiknya. to ensure
it is taken 2sampel company, namely pt centrin, pt online tbk. and pt limas centric indonesia
tbk. observations conducted for 40 days, ie in January and febuari 2014 trading stock pt
centrin online tbk. shows the average stock price 221per shares with the standard deviation of
17.93. Meanwhile, the observation of pt metrodata elektoniks tbk. during 30hari, ie in january
and febuari 2014, the trade shows the price of sahan average 312 per share with the standard
deviation of 16.37.apabila investors want the difference between the two stocks less than 75
with the intent to portofolion stock, how many opportunities the desire is achieved?

Answer :

𝜇1 = 221 ; n1 = 40; 𝜎1 = 17,93

𝜇1 = 321 ; n2 = 30; 𝜎2 = 16,37

How many P(X1- X2) < 75 ) ?

(𝑋2−𝑋1)−(𝜇1−𝜇2) 75−(312−221) −15
Z= = = = -3,68
𝑆𝑥1−𝑥2 (17,93)2 (16,37)2 4,12
√ +
40 30

Withavalue of Z = -3,68 got value probability = 0,499883.

So,P(X1-X2) < 75 = P (Z < -3,28) = 0,5 – O,499883 = 0,00012 or 0,0125 thus, the
probability for the difference of the two stocks of the dark 75 companies is 0,00012 or
0,012% the probability value in I is relatively small and this shows differnce in stock
princesthe relatively large of the company

*The averge value of the sample distribution of the average difference

𝑃𝑝1−𝑝2= 𝑝1−𝑝2=𝑃1−𝑝2

*the default value of the devation of the sample distribution of the avarege differece 𝜎𝑝1 −
𝑝2

𝑃1 ( 1−𝑃1) 𝑃2(1−𝑃2)
𝑆𝑝1−𝑝2= √𝑆2𝑝1+𝑆2𝑝 = √ +√
𝑛1 𝑛2

*
(𝑝1−𝑝2) (𝑝1− 𝑝2)
Z= −
𝑆𝑝1− 𝑝2

36
Example 11-10

seoran investors currently holds shares of various industry groups consisting of machinery
and heavy equipment, automotive, textile, and garment industries. observations over the past
2 months show that 44% of the stock price probability of this group increases. other investors
apparently held a trading group with the probability that the stock price would increase by
67% if the investor had 100 lots for the stock of various industry groups of trade increased by
10% more than the increase of industrial group's stock price?

answer :

trading n1 = 200, p1 = 0.67

various industries n2 = 100, p2 = 0.44
difference of proportion or difference of proportion = p1 - p2 = 0,1
the standard deviation of the difference in proportion is:

𝑃1(1−𝑝1) 𝑃2(1−𝑃2) 0,67(1−0,67) 0,44 (1−0,44)

𝑆𝑃1−𝑝2 = √ + √𝑝 = √ + √𝑝 = 0,06
𝑛1 𝑛2 200 100

the value of Z is obtained by

(𝑝1−𝑝2)− (𝑝1−𝑝2) 0,1−(0,67−0,44)
= = -2,16
𝑆𝑝1−𝑝2 0,06

so p ((p1-p2)> 0.1) = p (z> -2,16) = 0,5 + 0,4803 = 0,9803, so pobability of stock price
difference increased more than 10% is 98,03 %. this shows that there is a large relative
difference between the rise in stock prices of various industry and trade groups

11.7 Correction factor until population is limited

what has been discussed both on the average sample distribution and proportion as well as the
sample distri- bution of the mean difference and proportion, so far based on populations of
very large size or known as unlimited popuplations. but often in reality can be encountered
small or limited populations. if discussed companies that are monopolistic, makajumlah very
small companies such as power generation companies there are only two, namely Indonesia
power and ubhar, railway companies only one, while the television industry set = rta logging
only birapa only. for a population of a limited nature, usually with the number of samples
samples (n) less than 5% of the population (n)

the standard deviation adjustment for the average is :

𝜎 𝑁−𝑛
Sx = √ adjustment of the standard deviation for the population
√𝑛 𝑁−1

37
 𝑝(1−𝑝) 𝑁−𝑛
Sp = √ √
𝑛 𝑁−1

What is the benefit of this adjustment process? this adjustment process uses logic when the
sample is a percentage of the population, so expect any estimates to be of average value or
standard deviation, to be more precise than those obtained with smaller samples. using
adjustments or correction factors will result in a smaller value in estimating the average
standard count population deviciation:

Sample Size percentage dark population correction factor

10 0,010 0,9955

25 0,025 0,9879

50 0,050 0,9752

100 0,100 0,9492

200 0,200 0,8949

500 0,500 0,7075

11.8 postulate the middle boundary

as already presented in polygon graphs in subcharts 11.4 and 11.6, populations are normally
distributed, and so also the distribution of samples. if in our sample we have an average value
of x and standard deviation s, then the value x will be tedistribution normally can be
expressed by X - N (X, Sx2) which can be interpreted that the value of x is normally
distributed with rat and varian Sx2. why is the tendency of a distribution to normal
distribution to be so important that it is based on the mid-point proposition which is one
important proposition, which states: "for the average population and the o2 variance, the
average sampling of the sample distribution of all possible combinations of n-sized samples
obtained from the population would be close to the normal distribution, where the average of
the sample distribution count is the same as the average of the population count and the
sample distribution variance is equal to o2n '' in other words, for large samples of a
population derived from any distribution can be approximated by a normal probability
distribution. Thus, if known values and sample variance, it can be easy to draw conclusions
about population parameters.

38