Basic Concepts of Sampling-

Brief Review; Sampling Frame

Dr. A.C. Kulshreshtha

U.N. Statistical Institute for Asia and the Pacific (SIAP)

Second RAP Regional Workshop on

Building Training Resources for Improving Agricultural & Rural Statistics
Sampling Methods for Agricultural Statistics-Review of Current Practices
SCI, Tehran, Islamic Republic of Iran
10-17 September 2013
 Sampling – Basic concepts and definitions

➢ Surveys and Sampling

➢ Population and Sampling unit
➢ Population Parameters and Statistic/ Estimators
➢ Random Samples and Probability Sampling
➢ Estimator and estimate
➢ Unbiasedness, Consistency, Efficiency
➢ Sampling Frame

Sept.2013 2
Basic Concepts and Definitions- Surveys
Survey: A set of statistical activities designed to obtain data
Types of survey: Census surveys, Sample surveys
Census survey: A complete enumeration of the population of
interest. Data are collected from all elements of the
population
Examples: Population Census, Agriculture Census, Economic Census,
Census of housing, Census of establishments, Livestock Census
Sample survey: A survey of a subset of population employing
selection procedure
Examples: Labour Force survey, Household Income & Expenditure
survey, Survey of Establishments and Enterprises, Demographic health
survey, Living conditions survey

Sept.2013 3
Basic Concepts - Sampling
Sampling is selection process of a part (sample) of an
aggregate material (technically termed as Population) to
represent the whole aggregate
The process of sampling provides information about the
population characteristics on the basis of the
representative sample observations
Advantages of sampling:
Reduced cost (economy)
Greater Speed and Timeliness
Feasibility (if observations are destructive)
Greater Quality and accuracy
Detailed/specialized information, etc.

Sept.2013 4
Basic Concepts- Units and Population
• Unit is an element on which observations can be made.
These are the units of analysis.[Examples: households,
farms/ plots of agriculture crop]
• Reporting unit is one that actually supplies the required
statistical information
• Observation unit is one about which data are reported
• Population (or universe) is defined as totality or
collection of all units (elements) under study
• Finite population has limited number of elements and
an infinite population has unlimited elements

Sept.2013 5
Basic Concepts- Sampling Units
Sampling Unit is an element of the population selected in
the sampling process on which we collect data
Example:
• When we select a sample of households , target units of
observations may be persons living in the households
• Female members of household at age 15-49 in reproductive
health survey
• In multi-stage sampling plan, one has first stage sampling unit
(fsu), second stage sampling unit (ssu), etc.

Sept.2013 6
Basic Concepts- : Characteristic
Characteristic: Different kinds of information on elements of the
population are collected in a survey. Each of these items of
information is called a characteristic
Each characteristics has different values for different individual units
Observations on several characteristics of the units are collected in
a survey
Characteristic can be a quantitative variable like income of a
household, number of cattle on a farm, area of land under rice crop
in an agricultural holding
or an attribute or categorical variable like gender, employment
status of a person, economic activity code of a production unit

Sept.2013
7
 Basic Concepts- Population(Contd.)
• Variable: a characteristic of an element under study
whose measure called value, differs from unit to unit
• Uni-variate population: one where only one
characteristic is observed on an element
• Multi-variate population: one where more than one
characteristic is observed on an element (for two
characteristics, it is bi-variate population)
• Target Population: the set of elements about which
information is sought and estimates of the population
parameters are obtained on the basis of the sample.
Also known as coverage universe
[some units may be excluded (e.g. the homeless)]
Sept.2013 8
 Basic Concepts- Random Sample, Probability Sampling
• Random sample is a sample based on a probability scheme
• Probability sampling theory aims to make statistical inference about
the population parameters on the basis of random samples. Every
population element has a known, non-zero, probability of selection
for the sample
• Probability Sampling Ingredients
– Defined set of samples that are possible to obtain with the sampling procedure
– A known probability of selection is associated with each possible sample
(sampling design)
– A known nonzero probability of selection (inclusion probability) of each element
in the population
– One of the possible samples is selected by a random mechanism according to
the sampling design
• Non-probability Sampling makes use of purposive sample (Quota,
Convenience, Snowball, etc.), results are not statistically valid
[We will focus only on probability sampling] 9
Sept.2013
 Basic Concepts- Parameter, Statistic, Estimator, Estimate
▪ A population parameter is a numerical summary of a population, a
function of elements in the population (Pop mean µ, Pop variance σ2 )
A Statistic is a function of elements in the sample (a subset of the
population) ➔ It is called estimator if indicating parameter

Population Sample
Parameter () Mean
Statistic

Sample

▪ Estimate: numerical value of an estimator that is obtained from a

particular sample of data and used to indicate the value of a parameter
Sept.2013 10
 Example
The values of X and Y shown in the table below
are the actual values (not known to the sampler)

# milch average
Milk Milk output
animals yield
Producers (Y)
(X) (R)
A 3 145 48.3
B 6 260 43.3
C 5 245 49.0
D 5 290 72.5
E 2 140 70.0
F 4 180 45.0

Sept.2013 11
 In the Example
sample
Samples sample values of X sample values of Y ratio -
estimate
1st unit 2nd unit 1st unit 2nd unit mean ( x ) 1st unit 2nd unit mean of R
C D 5 5 5 245 290 267.5 53.5
A B 3 6 4.5 145 260 202.5 45.0

Estimates
Sample Sample
mean ratio
Estimators

Sept.2013 12
 Qualities of Estimators: Unbiased, Consistent, Efficient
• Unbiased estimator of a population parameter is an estimator whose
expected value is equal to that parameter
In the example, Sample means of ‘average number of milch animals’ and
‘average output’ are unbiased estimators of the respective population parameters
(Population means) But, sample yield rate (which is a ratio) is not an unbiased
estimator of the corresponding population parameter
• Consistent estimator is one where the difference between the
estimator and the parameter grows smaller as the sample size grows
larger
Sample ratio (in the example) is not unbiased but is a consistent estimator
• Efficiency is defined as the reciprocal of sampling variance
If there are two unbiased estimators of a parameter, the one whose
variance is smaller is said to be relatively efficient
Sept.2013 13
Sept.2013 14
 Properties of Estimators- Sampling Error
Sampling Error: represents the difference between the estimate
ˆ and the value of the population parameter θ
(ˆ —θ), the error in a sample estimate that owes to the selection of only
a subset (sample) of the total population rather than the entire population.
All sample estimates are subject to sampling error
The most commonly used measure of sampling error is
  

Sampling Variance: V ( ) = E[ − E( )]2

Where E denotes expected value, Sampling variance depicts Precision of
the estimator
 

Bias of estimator : B( ) = E( )−    

Efficiency of the estimator: MSE( ) = E( −  )2 = [Bias ( )]2 + Var( )

15
Sept.2013
 Properties of Estimators- Sampling Variance
• Sampling Variance is the average of squares of (value of the
survey estimator obtained from a sample minus the value of the
population parameter) over all possible samples that can be drawn from
the population   
V ( ) = E[ − E( )]2
• The variance of an estimator contains information regarding how close
the estimator is to the population parameter
• Sampling variance is a measure of sampling error
• Sampling error reflects the difference between an estimate derived from
a sample and the “true value”

Sept.2013 16
Properties of Estimators (Cond.)
Sampling Distribution: A frequency distribution of the values of an
estimator for each sample that can possibly be drawn from the
population
• The sample design and sample size remaining unchanged, the
higher the population variance (measure of variability of a
population) the higher is the sample variance
• The sample design and population (variance) remaining
unchanged, the higher the sample size the lower is the sample
variance
• For a given population and sample size, the sample variance
depends on the sample design adopted

Sept.2013 17
Properties of Estimators- Standard Error, RSE
• Standard error of an estimator Ŷ is the squared root of
the sampling variance of the estimator
 
s.e.(Y ) = Var(Y )

• Relative Standard Error / Coefficient of Variation of an

estimator is Standard Error / value of Parameter Y

SEY 
RSE =  
Y

Sept.2013 18
 Measure of Efficiency - Design Effect (Deff)

• The relative efficiency of a sample design w.r.t. the simplest

sample design – SRSWOR – is measured by Design effect (Deff)
• Design effect of a sample design, say D, is defined as the ratio
the standard errors of D and SRSWOR
s.e.(design D)
Deff =
s.e.(SRSWOR)
• Estimates of Deff are often used for determining the required
sample size for a given design

Sept.2013 19
Survey Design – Issues involved
1. Determining survey objectives and data requirements
2. The population of interest or the target population
3. Reference period; Geographic and demographic
boundaries
4. Sampling frame and sampling unit
5. Sample design
6. Selection of the sample (at different stages)
7. Survey management and field procedures
8. Data collection
9. Summary and analysis of the data
10. Dissemination

Sept.2013 20
 Sampling Frame
• Sampling frame is a list or device that delimits,
identifies, and allows access to the elements of the target
population. Sample elements are selected from frame
using appropriate probability scheme
• Frame Units: Examples
• Area units • Non-area units
– Administrative Subdivisions – Housing Units

– Households
– Census Enumeration Areas
– Persons
– Areas Delineated on Maps
– Enterprises
– Agricultural Holding
– Establishments

Sept.2013
21
 Universe Sample
Sampling FRAME: Concept
• Survey Population has to be represented in a physical form
from which sample can be selected
• Sampling Frame is such a representation
– Explicit list of the elementary units to be surveyed
– Account for all units in some implicit form
Access to Population
 Set of source materials from which the sample is selected
 Provides a means for selecting the elements of the target population
that are to form the sample
Sept.2013 22
 Population Sample

Sampling FRAME: Importance

• Related to cost of survey: availability of
sampling frame suited to best option among
alternative sample designs
• Affects data quality: ‘faulty’ sampling frames
lead to non-sampling errors

Sept.2013
23
 Types of Frames
• List Frames
• Area/ Area-based Frames
• Multi-Stage Frames
• Frames for Series of Surveys
• Multiple Frames (combination)
• Master Sampling Frame:
– Multi-purpose
– Multi characteristics

Sept.2013
24
 Considerations for choosing a Frame
• Long term view of all possible surveys
• Based upon available infrastructure
• Possibility of updating
• Statistical Registers: a type of Frame
– difficult to keep them updated, if not linked to an
administrative process

Sept.2013
25
Sampling Frames: List Frame
• List frame enlists all the elements of the population
Each frame unit is identified and listed with its
identification particulars, such as--
–Listing of household addresses from Census
– Listing of names, addresses and telephone numbers in
telephone directory
– Listing of establishment names, addresses and type of
business produced from a business register
List Frame Sources: Population Registers, Census Listings,
Administrative Lists (e.g., Business Registers)
Listing by operations (e.g., Listing of households within selected
enumeration areas in 2-stage sampling)

Sept.2013 26
List Frames (Contd.)

• Completeness is most critical requirement

• Accurate information of the size and characteristics of
individual units
• Different situations require separate treatment
– Large and few units for which good lists may be available
– Medium sized units covered by a list frame that is more difficult to
construct and maintain
– Medium to small units which may require a combination of list and
area based frames
– Small units which can only be covered with area based frames
– For unevenly distributed units, specially constructed area frame,
taking account of the patterns of concentration (e.g., fishing
households, mining worker households); Rare and dispersed
population also require special methods (tribal households)

Sept.2013 27
 Problems with List Frames
• Blanks (L=0): Listing represents no real unit
• Duplications (L, L) =U: Same unit is represented by more
than one unit
• Clustering of elements L= (U, U): More than one unit
represented by same listing
• Under coverage U=0: Units not represented in the frame
• Failure to locate units L=?: Failure to identify which unit a
selected listing represents
• Change in units and unit characteristics L = U*: Unit itself
(or characteristics of the unit associated with the unit) has
changed

28
Sept.2013
 List Frames
• Strengths:
• can use inexpensive data collection methods (mail, telephone)
• can target specific or rare commodities
• can reduce variability due to sampling
• cost-efficient: could be built on available resources
• Weaknesses:
• does not cover entire population (threshold criteria)
• goes out-of-date quickly
• increased non-sampling errors due to data collection methods
• requires on-going maintenance
➢ build
➢ update
➢ remove duplication
➢ remove out-of-scope records

Sept.2013 29
 Sampling Frames: Area-based
• Area frame is a geographic frame consisting of area units. Every
population element belongs to an area unit
• Frame units are the geographical units in a hierarchical arrangement
– Cover the entire country
– Boundaries are well delineated
– Population figures are available
– Units are mapped
• For household surveys, in developing countries, frame generally
consists of
– one or more stages of area units
– followed by the list of households or enterprises within the selected
ultimate area units
• Durability of frame declines as we move down the hierarchy of the units
Sept.2013 30
 Sampling Frames: Area-based (Contd.)

• Choice of the type of units to be used as the primary stage

units (PSU) is important
• Coverage errors in area based frames arise from
– failures to define boundaries of area units correctly
– poor quality of the lists of ultimate units
• Sources for area-based frames are generally the
– Census blocks or enumeration areas
– Maps of administrative sub-divisions
– Aerial photographs of housing units
– Satellite images of land cover

Sept.2013 31
 Area Frame…More..
• Land is divided in segments on the basis of land use
• A unique ID is assigned to each segment
• List of segments
• Segments (or blocks) are Stratified
(Stratum: a group of homogeneous units)
• Area Frame: linking of segment with the holding is
possible

Sept.2013 32
[National Agric. Statist Service] NASS Area Frame
constructed using satellite imagery, digital maps, GIS software, aerial photography: (i) divide
land area into strata based on land use, (ii) subdivide land use strata into strata blocks, (iii)
select a sample of strata blocks (iv) subdivide selected strata blocks into segments

>50% cultivated
15-50%
cultivated
<15% cultivated
agri urban
commercial
non agricultural
water 33
 Common Problems with Area Frames
• Though areas as units are larger, more easily identified
and more stable than dwellings, households which
appear as units in lists, there are problems, like
• Failure to cover the population exhaustively
• Errors in area boundaries
• Inappropriate type and size of units
• Lack of auxiliary information
• High cost (useful only for repetitive surveys)

Sept.2013 34
 Area Frame (Contd.)

Strengths: Weaknesses:
• complete coverage • expensive (frame
• reduced non-sampling errors construction & data
• estimates well for commonly collection)
produced commodities • difficult to target specific or
• versatility rare commodities
• longevity (desertification/ • sensitive to outliers
• can be inefficient
greening)
• requires definable physical
boundaries

Sept.2013 35
Multi-stage Frame
• Usually area based
– Primary sampling frame (frame for the first stage of sampling)
has to cover the entire population
– Following first stage selection, the list of units at any lower
stage is required only with in the larger units selected at the
preceding stage
• Frame from a single source / number of sources
• Different types of frames may be used for different parts
of the population

Sept.2013 36
 Multi-stage Frame (Contd.)

• Used in Multi-Stage Sampling

– Primary sampling frame for first stage selection
– Intermediate sampling frames for intermediate
stages of selection
– PSU, SSU, TSU,. ..... ,FSU

Sept.2013 37
 Example: 2-stage Sampling
• First stage frame
– Sampling frame units are areas (e.g., blocks, EAs)
[Sampling frame is an area frame]
• Second stage frame
– Sampling frame units are elements (e.g., households; informal
sector enterprises)
[Sampling frame is a list frame of households for each of the
area units sampled in the first stage]

Sept.2013 38
 Multiple Frame
• Technique for combining several incomplete frames in
order to capture completeness of the frames
• Union of all the frames constitutes the entire population
• Samples are selected independently from each frame
• Optimum combination of estimates coming from non-
overlap
• Overlap domains provides an overall estimate for the
multiple frame situation

Sept.2013 39
 Multiple Frame (Contd.)
• Overlap between frames provides independent estimates from
samples coming from different frames
• When overlapping frames used, ensure that unit’s probabilities
of selection remain definite & known
• Various approaches:
– Make frames non-overlapping
– If a list and an area frame are used in combination, any unit selected
from area frame must be excluded from list (whether or not selected)

Sept.2013 40
 Frame for Series of Surveys-
Master Sample
• In multi-stage sampling design, each stage involves task
of frame preparation and sample selection, till finally a
sample of ultimate units is obtained
• For economy/convenience one or more stages of these
tasks may be combined or shared among a number of
surveys
– Sample resulting from shared stages is called a
Master Sample

Sept.2013
41
 Master Sampling Frame
• Master Sampling Frame is basically a list of area units
that covers the whole country
• For each frame unit there may be information on
– urban/rural classification
– identification of higher level units (for example, the district and
province to which the unit belongs)
– population counts
– stratification variables

Sept.2013 42
 Master Sampling Frame- Concept
• From a master sampling frame, it is possible to select the
samples for different surveys entirely independently
• In many instances, there are substantial benefits
resulting from selecting one large sample, a master
sample of area units for the first stage of sampling, and
then selecting sub-samples of this master sample to
service different (but related) surveys
• It is a sample from which sub-samples can be selected to
serve the needs of more than one survey or survey round

Sept.2013 43
 Master Sample-
Objectives
• A common sample of units down to a certain stage from
which further sampling done for individual surveys
• Economize, by sharing among surveys, on costs of
developing sampling frame, design and selection
• Facilitate linkages between different surveys or between
successive rounds of a continuing survey
• Control drawing of multiple samples for various surveys
from the same frame

Sept.2013 44
 Master Sample (Contd.)
• Master sample for household surveys (multi-stage)
– Same sample of primary sampling units (usually, area

units) are used for different surveys

– Next stage units may be same or different

• Periodic (monthly, quarterly, annual) surveys of

establishments make use of a sample of establishments
maintained over some interval of time
• Sample of units designed for multiple use (different or
same surveys repeated over time)

Sept.2013 45
 Master Sampling Frame- Advantages
• Cost of developing a good sampling frame is usually
high and the establishment of a continuous survey
programme makes it possible for the NSO to spread
the costs of construction of a sampling frame over
several surveys
• Facilitates quick and easy selection of samples for
surveys of different kinds and it could meet different
requirements for the sample from the surveys

Sept.2013 46
Integrated Household Survey Programme
• Concept
– long range plan rather than ad hoc
– coordinated planning
– integration of survey design and operations
• Integration can be achieved by
– Use of same concepts and definitions for variables
occurring in several surveys
– Sharing of survey personnel and facilities among the
surveys to secure effective use of staff and facilities
– Use of common sampling frames and samples for all
the surveys in the survey programme
Sept.2013 47
 Desirable Frame Properties
 Quality-related
 Efficiency-related
 Cost-related

Sept.2013 48
 Frame Quality
• Criterion: as a result of the frame, every member of the
target population has a known, non-zero chance of being
selected
• Completeness: all elements are covered
• Accuracy: each element is included once and only once
• Up-to-date: updating of frame must be regular plan
activity

Sept.2013 49
 In Practice …
Target population Frame population

Not reachable
Not Sampled
Refusals Not eligible
included in
sampling population for survey
frame Other non-
responses

Out-of-scope
Under-coverage
Non-response
Sept.2013 50
 Frame Imperfections

Target population Frame population

SAMPLE

Under-coverage: Over-coverage:
Population units not in frame Out-of-scope units

51
Sept.2013
 Frame Imperfections (Contd.)

• Over-coverage
• Duplications
• Under-coverage, non-coverage, incomplete
coverage
• Not enough information to provide access to
some survey population units

Sept.2013 52
 Efficiency Cost
• Inclusion of accurate and up-to-date Low cost of acquisition /
auxiliary information preparation
• Good quality maps of units available Low cost of use
Low cost of
• Easy access; Easy to process
maintenance
• Choice of sampling units available (if
to be used for more than one survey
or survey round)
• Enabling production of summary
statistics

Sept.2013 53
 Sampling Frames: Contents
 Key contents
 Operational Considerations
 Construction
 Administration and Maintenance

Sept.2013 54
 Sampling Frames: Contents
• One record per frame unit
– Primary identifier
Identifiers
– Secondary identifier(s)
– Stratification variable(s)
Unit Characteristics
– Measure(s) of size
– Sample selection indicator
Operational data
– Change indicator(s)

55
Sept.2013
 The Choice of Frame
• Depends upon related on going activities in other
sectors
– Resource availability
• Nature of agriculture
– Extensive, mono-crop, or
– Intensive multi-crop
• Scope of surveys
– just the crop area, crop monitoring, land degradation or
– many economic characteristics e.g. fertilizers, cost of
production

56
Sept.2013
Introduction to Survey Design
Sample is a subset of the population on which
observations are taken for obtaining information about
the population.
Since studying a sample we wish to draw valid
conclusions about the population, sample should
desirably be ‘representative’ of the target population.

Sample design specifies how to select the part of the

population to be surveyed

Sept.2013 57
Probability Sampling Designs

The most common sampling techniques used for

official surveys are :
• Simple Random Sampling
• Systematic Sampling
• Stratified Sampling
• Probability Proportional to Size (PPS) sampling
• Cluster Sampling
• Multi-Stage Sampling
All are examples of probability sampling

[These will be discussed in next session]

Sept.2013
58
Thanks

59