4 ประชากร การสุ่มตัวอย่าง และวิธีการเก็บ Eng

Population and
Sampling
Associate Prapon
Sahapattana, Ph.D.
Graduate School of
Public Administration
Topics
 Meaning of population
 Census
 What are needed for sampling?
 Type of sampling
 Non-probability sampling
 Probability sampling
 Sample size
 Data collection
Population
Question
Identification
Research
Result Literature
Presentation Review
Data Hypothesis
Analysis Selection
Creation of
Data Research
Collection Design
Tool and Operational- outline
Collection ization and Population
Indicator and
Placement Sampling
Research Methodology Process

Population
 All units of the things studied
 Ex. A study of voter opinion: what is the
population?
 All voters
 Need to identify that population include whom
and not include whom. Ex. Voters include
those who living outside the country or not?
Population
 Limit to human only?
 Ex. A study of role of media on violence in
children. Population may be each piece of
newspaper news
 Population may be organization, ex. School,
police station, province, village, etc.
 Population may be every individual human.
Population
 Population identifies the scope of research
whether it very large or specific to some
groups.
 Population reveals the sampling frame which is
the list of every single name building up to the target
population.
Census
 A study of every unit of what we are studying.
 Do not usually done because consumes high
labor and cost, makes data not up-to-date,
takes time, and difficult to control the quality
of data.
 However if population is small, census should
be done, ex. A study of opinion of female
investigators on law regarding juvenile; a
study of university students who are disable,
etc.
Sampling
What are needed for sampling?
 Objectives of the study and the objectives to
use the study findings which derives from
hypothesis and operationalization.
 Things which in the scope of population to be
studied
 Sampling frame which contains up-to-date
data
 Sampling method and sample size
 Data collection method, ex. Interview, focus
group, observation, etc.
Types of Sampling
 Non-probability sampling
 Probability sampling
Non-probability sampling
Non-probability sampling
 Sampling which the chance of each sample to
be selected is unknown
 The accuracy or reliability of the statistic from
sample estimation to the population cannot be
measured.
 Not use when need to estimate the population
parameter.
 Use when required samples are rare in
population, Ex. Hacker, villages without
internet, etc.
Non-probability sampling (Cont.)
 Accidental or convenience sampling

 Use when researchers cannot manipulate the
sample’s life but collect data from the real setting

as cooperation provided by sample
 Collect data from those who are willing to
cooperate or who happen to be in the place

 Ex. A study of immigration of labors into industrial
zone.
 Caution to interpret the representative of the
population
Quota sampling: Use when researcher want samples

with some specific characteristics
 Ex. Men using wrinkle-free lotion and men
who are not using. A researcher collects data
from each group for a specific number planned
and analyze.
 After a quota is set, any way of data collection can
be used.
 A representative issue to be considered if needed.
Purposive sampling: Use when data is needed from an

expert or experienced person. Ex. Interview high
ranking management of an organization
 No representativeness
 Or a researcher believes that the data can be

representative of the population
 May be wrong if the researcher does not familiar
with the population
Snowball sampling: Use when the subject is difficult to
access. Ex. male juveniles in sex service or students use
drug
Method:
 Begin with a first sample then ask the sample to
identify other 1 to 3 samples. Go collect data from

them and ask each sample for other 1 to 3 samples
till collect enough samples.
 Popular in qualitative research
 No representativeness
Probability sampling
Probability sampling
 The chance of each sample to be picked from
the population can be identified
 Generally, the chance for each sample is equal.
However, it can be weighted. Ex. Villages
with more households have more chance to be
picked into sample than those with less
households.
Probability sampling (Cont.)
 Sample must accurately reflect the population,
or having the same combinations.
 Probability sampling allows the use of statistic
to estimate the parameter of the population.
 Ideally, sample should be as small as possible
to hold the characteristics of the population in
order to save the cost of data collection.
Types of Probability Sampling
Probability Sampling
Simple random sampling: Basic of the probability sampling and
also be used for other finalized way of sampling
 Each unit has equal opportunity to be picked
 Begin with sampling frame, then give number
from 1 to N
 Pick samples in two ways: number drawing or
table of random numbers
Example of table of random numbers
Probability Sampling (Cont.)
Systematic sampling: A solution for reducing the drawing times.
 Begin with a number smaller than k, k = N/n. (N =
population, n = sample size)
 Then count by k to the next number, pick 1 sample
and count by k, then pick 1 sample, and so on till
get enough sample.
 Use when subjects sorted randomly otherwise
samples may have some sort of relationship and
not represent the population. Ex. Data about
factories registered with ministry of industry was
sorted by size (if apply this method, sample may
compose of only large or small factories)
Stratified sampling: Use when characteristics of population vary
widely. Ex. Income of rural people in regions across country,
rice productivity per rai in Thailand, etc. of
 Sampling directly from the population may result
in large error especially a small sample.
 Stratified sampling can solve this problem. Begin
with dividing samples into subgroups with most
homogeneous within subgroup and heterogeneous
between subgroup. Then pick samples from each
subgroup.
Stratified sampling: Ex. A study of farmer income in the northern
part of Thailand. A researcher may divide population into 4
subgroup: farmers in large irrigation system, medium, and small
irrigation systems, and those outside the irrigation system. Then
pick samples from each subgroup by using simple or systematic
random sampling. After adjust for the weight, the samples will
represent the farmers in the northern part of Thailand.
Cluster sampling: Use when population can be divided into
groups with diversity within the group and different between
groups. It’s like shrink the population.
 The division of group may use geographic, like
province, district, subdistrict, or organization, level
of students, etc.
 This method saves more time and labor compared
to simple random sampling.
 When pick samples from each subgroup, simple
random sampling or systematic random sampling
may be used.
Multistage sampling:
 In cluster sampling, a researcher may sampling
deeper into sublevel.

 Ex. Province may be divided into district,
subdistrict, village, household; university divided

into school, department; ministry into department,
division; etc.
Multistage sampling:
 One strength: no need for sampling frame of the
households, use only in upper level: village,

subdistrict, district, and province.
 Ideally, more samples in the beginning stages of
sampling and fewer samples in the later stages of

sampling. However, it will cost more for the process.
Sample Size
 The bigger size of sample, the better
representative of sample to population.
 The sampling error depends on the sample
size, not the ratio of population to the sample.
 To set sample size, one needs to set the
confidence level first, e.g., .90, .95, .99.
 Then calculate the sample size or use table.
Sample Size
Issues to be considered for setting sample size:
 Error accepted or confidence level
 Variance of population
 Objectives of the study
 Statistical technique used
 Budget
Data Collection Process
Data Collection Process
 Data collection: collect data by researcher.
The data called primary data. Ex.
Questionnaire, interview, focus group,
observation, and experiment.
 Data compilation: use data already collected or
reported. The data called secondary data. Ex.
Statistics from organizations, archives,
publications, etc.
Type of Data
 Primary and secondary data
 Qualitative and quantitative data:
 Qualitative data: not number, cannot be process by statistic
 Mostly, from observation, note, conversation, interview, focus
group, brain storming
 Quantitative data: data in number
 Field research should use both qualitative and quantitative
data.
 In a big research project, qualitative data should be used for
pilot study, then the collection tool can be constructed.
 A good qualitative research should use quantitative data to
support the study findings or conclusion.
Question or Comment?

4 ประชากร การสุ่มตัวอย่าง และวิธีการเก็บ Eng

Uploaded by

Copyright:

Available Formats

4 ประชากร การสุ่มตัวอย่าง และวิธีการเก็บ Eng

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

4 ประชากร การสุ่มตัวอย่าง และวิธีการเก็บ Eng

Uploaded by

Copyright:

Available Formats

Population and

Research Methodology Process

 Accidental or convenience sampling

sample’s life but collect data from the real setting

cooperate or who happen to be in the place

Quota sampling: Use when researcher want samples

Purposive sampling: Use when data is needed from an

 Or a researcher believes that the data can be

identify other 1 to 3 samples. Go collect data from

deeper into sublevel.

subdistrict, village, household; university divided

households, use only in upper level: village,

sampling and fewer samples in the later stages of

 Objectives of the study

 Statistical technique used

You might also like