You are on page 1of 40

Business Research Methods

(MGP6010A)

Research Methodology
MGP6010A
MIT-WPU: SoM(PG)

Unit 4
Sampling
PPTs by

Dr. Swapnil Undale

This work is licensed under the Creative Commons Attribution-ShareAlike 4.0 International
License. To view a copy of this license, visit http://creativecommons.org/licenses/by-sa/4.0/.
Business Research Methods
(MGP6010A)

Sampling
4.1 Basic concept of Sampling: universe/population, sample,
sampling frame, sampling unit & characteristics of good
sample.

2
Business Research Methods
(MGP6010A)

Sampling-
Process of selected samples for the study
Population/Universe-
Population refers to any group of people or objects that form the subject of
study in a particular survey.
“The universe of units from which the sample is to be selected.” –
Bryman & Bell
“The total collection of elements about which we wish to make
some inferences” – Cooper & Schindler

Sample-
It is a subset of population
“Segment of population that is selected for investigation.”
– Bryman & Bell
“Some of the elements of population” – Cooper &
Schindler
Business Research Methods
(MGP6010A)

Sampling-
Population Element/Unit
Single member of the population
“A individual participant or object on which the measurement
is taken”
It is the unit of study
Sampling Element/Unit
Single member of the sample
Census
An examination/study of each and every population element
Sampling frame
“List of all population elements from which the sample will be
drawn is called as Sampling frame”
Business Research Methods
(MGP6010A)

Characteristics of Good Sample Design


• Sample should truly represent population.
• Sample size should be adequate.
• Sample design should result in a small sampling error.
• Sample design should be viable in the context of resources (time,
money and manpower) available for research.
• Sample design should be such that the systematic bias should be
controlled. In other words, it means that all items of the sample
should be selected independently of one another and all items of
the population should have the same chance of being selected in
the sample.
• Sample design should be such that the result of the sample study
can be applied, in general, for the population with reasonable level
of confidence.
Business Research Methods
(MGP6010A)

Sampling Process
1) Define Population/Universe

2) Get Sampling Frame/Source List

3) Decide Census or Sampling

4) Decide Sample Size

5) Decide Sampling Method/Procedure


Business Research Methods
(MGP6010A)

Sampling
4.2. Probability sampling: different types of sampling.

7
Business Research Methods
(MGP6010A)

Probability & Non probability Sampling-


Probability sampling-
Also called as ‘random’ or ‘chance’ sampling.
Every item has equal chance of being included in the
sample.
Lottery method, individual units are picked up from the
group not deliberately but by some mechanical process.

Random sampling ensures the law of Statistical Regularity – “If on


an average the sample chosen is a random one, the sample will
have the same composition and characteristics as the universe”

Best technique to select representative samples.


Business Research Methods
(MGP6010A)

Probability & Non probability Sampling-


Non-Probability sampling-
Also called as ‘deliberate’ or ‘purposive’ or ‘judgment’
sampling.
Items are selected deliberately.
Researcher’s choice of selecting element and include them
in sample.
No assurance that every element will get equal chance to
be selected as sample
Business Research Methods
(MGP6010A)

Sampling Methods/Techniques
➢Probability Sampling Methods
1. Simple Random Sampling Method
2. Systematic Sampling Method
3. Stratified Random Sampling Method
4. Cluster/area Sampling Method
5. Multistage Sampling Method

➢Non-Probability Sampling Methods


1. Convenience Sampling Method
2. Judgment/Purposive Sampling Method
3. Quota Sampling Method
4. Snowball Sampling Method
Business Research Methods
(MGP6010A)

Probability Sampling Methods


1. Simple Random Sampling Method:

• Samples are selected randomly from Population


• Randomness means without any logic/system/pattern
• All the units have equal chance of selection
• Methods:
o Lottery Method
o Random Number Table
o Random Number Generating Software
Business Research Methods
(MGP6010A)

Simple random sample


A probability sample in which each element of population
has a known & equal chance of selection

E.g. Population = Total students in MITWPU (1200)


Sample size = 50 students
Sample size
Probability of selection = -----------------------
Population size
= 50/1200
= 0.042 or 4.2%
Business Research Methods
(MGP6010A)

Probability Sampling Methods


2. Systematic Sampling Method:
• Both System and Randomness together
• Each unit is given a unique number and arranged in
order
• Calculate Sampling/Skip Interval (SI)
• SI= Population Size/Sample Size = ‘k’
• Select one unit randomly from first ‘k’ units from
sampling frame
• Go on adding ‘k’ (SI) (System) to the previously
selected unit to find out next unit
Business Research Methods
(MGP6010A)

Systematic sampling
A probability sample drawn by applying a calculated skip
interval to a sample frame.

Formula to calculate skip interval-

Population sample frame (N)


Skip interval(k) = ------------------------------------------------
Sample Size (n)
Business Research Methods
(MGP6010A)

Systematic sampling
E.g. Population = Total students in MITWPU (1200)
Sample size = 50 students

Population sample frame (N)


Skip interval(k) = ------------------------------------------------
Sample Size (n)
= 1200/50
= 24

Select any number randomly between 1 – 24, and then select rest
49 numbers with a gap of 24 numbers.
Business Research Methods
(MGP6010A)

Systematic sampling

k = 24
Select any number randomly between 1 – 24, and then select rest
49 numbers with a gap of 24 numbers.

23 143 263 383 503

47 167 287 407 527

71 191 311 431 551

95 215 335 455 575

119 239 359 479 599


Business Research Methods
(MGP6010A)

Probability Sampling Methods


3. Stratified Random Sampling Method:
• Population is divided (Stratified) into homogeneous
groups called as Strata
• Homogeneity within Group
• Heterogeneity amongst Groups
• Researcher decides Stratum Sample Size
• Samples are selected from all the strata randomly
• Most Accurate Method
• Requires More Resources
Business Research Methods
(MGP6010A)

Stratified Random sampling


A probability sampling technique in which the population
is divided into different sub-homogeneous groups or strata and
samples are randomly selected from such sub-groups or strata.
E.g. Population = Total students in MITWPU e.g. 1200
Sample size = 50 students
Sub groups, Select from each of following groups -
MBA Mkt = 240 MBA Fin = 160
MBA HR = 120 MBA Opr = 60
MBA Gen = 60 MCA = 90
ETC…

Samples from each stratum can be selected proportionately or


disproportionately ….
Business Research Methods
(MGP6010A)

Probability Sampling Methods


4. Cluster/area Random Sampling Method:
• Population is divided into groups called as Clusters
• Heterogeneity within Group
• Homogeneity amongst Groups
• Clusters are selected Randomly
• All the Units from the selected Clusters are selected as
Sample
• Not a Single Unit from remaining clusters is selected
• Most Cost Efficient Method
• Less Accurate Method
Business Research Methods
(MGP6010A)

Cluster sampling
A probability sampling technique in which the population
is divided into several small sub groups and some groups selected
randomly for study.

e.g. MBA Mkt = 240 MBA Fin = 160


MBA HR = 120 MBA Opr = 60
MBA Gen = 60 MCA = 90
ETC…
Business Research Methods
(MGP6010A)

Cluster sampling
A probability sampling technique in which the population
is divided into several small sub groups and some groups selected
randomly for study.

e.g. MBA Mkt = 240 MBA Fin = 160


MBA HR = 120 MBA Opr = 60
MBA Gen = 60 MCA = 90

Area Sampling is same as cluster sampling where different


geographical areas are considered and clusters
Business Research Methods
(MGP6010A)

Probability Sampling Methods


5. Multistage sampling-
A probability sampling technique which involves the
selection of sample units in mote than one stage.
PSUs- Primary sampling units (first stage units) from
population are selected, from PSUs second stage units are
selected, the process continues until the selection of final
sampling units (final stage units).

e.g. Tea consumption Sample size = 7000 across country


1st stage Select district – (PSU) = 25
2nd stage Select cities = 4 from each city = 100 cities
3rd stage Select families = 70 from each city = 7000
(Finale stage/sampling unit)
Business Research Methods
(MGP6010A)

Sampling
4.3. Non-probability sample- different types of non-probability sampling,

23
Business Research Methods
(MGP6010A)

Probability & Non probability Sampling-


Non-Probability sampling-
Also called as ‘deliberate’ or ‘purposive’ or ‘judgment’ sampling.
Items are selected deliberately.
Researcher’s choice of selecting element and include them
in sample.
No assurance that every element will get equal chance to
be selected as sample
Business Research Methods
(MGP6010A)

Probability & Non probability Sampling-


Non-Probability sampling-
1. Convenience-
A non-probability sampling technique where researcher
use any readily available individuals as participants
Most cheapest & easiest to conduct
Least reliable
Researcher is free to select anybody as sample
Business Research Methods
(MGP6010A)

2. Judgment sampling-
A purposive sampling where the researcher arbitrarily
selects sample units to conform to some criterion.
An experts may be asked to select sample.
e.g. Problem – Sales de-growth across all territories (500+)
Study of all territories not practical.
Expert in Sales management will be asked to select sample
territories for study.
On the basis of certain criterions (applying his judgment)
he select sample territories
Criteria – market conditions, seniority of employee etc
Business Research Methods
(MGP6010A)

3. Quota sampling-
A purposive sampling where relevant characteristics are
used to stratify the samples.
This is useful to increase the representativeness of
samples
e.g.
Programme Tot Students Proportion Sample Size
MBA Mkt 240 32.88 16
MBA HR 120 16.44 8
MBA Gen 60 8.22 4
MBA Fin 160 21.92 11
MBA Opr 60 8.22 4
MCA 90 12.33 6
Total 730 100.00 50
Business Research Methods
(MGP6010A)

4. Snowball sampling-
A non- probability sampling in which subsequent
participants are referred by current sample elements.

Initial individuals are discovered/identified


These initial individuals refer others who are similar to them and
so on.
Like a snowball rolls along.
Business Research Methods
(MGP6010A)

Sampling
4.4. Sampling Errors, non-sampling errors, methods to
reduce errors, sample size constraints.

29
Business Research Methods
(MGP6010A)

Sampling-
Sampling error
Error created by sampling process.
Sample is not representative of population (Difference between
sample characteristics and population characteristics)

Non-sampling error
All errors other than sampling errors
Wrong answers
wrong data entry
Errors in coding, tabulation and computation
Error in defining Population
Unavailability of respondents
Refusal to answer fully or partially

Methods to reduce errors…. ????


Business Research Methods
(MGP6010A)

Methods to Reduce Sampling Errors


• Get correct and valid sampling frame
• Decide sample size scientifically using a statistical
formula
• Use large sample size: The law of ‘Inertia of Large
Numbers’ (other things being equal larger the size of sample more
accurate the results are likely to be)
• Random selection of sample: The law of ‘Statistical
Regularity’ (if a sample is taken at random from a population it is likely to
possess the characteristics as that of the population)
• Draw a representative sample
Business Research Methods
(MGP6010A)

Methods to Reduce Non-Sampling Errors


• Define research problem correctly and clearly.
• Formulate hypothesis scientifically.
• Select appropriate type of data.
• Use valid source of data.
• Use appropriate data collection method.
• Use properly selected and logically designed data collection instrument.
• Get complete data.
• Get accurate responses from the respondents.
• Do proper selection, training and supervision of manpower involved in
research.
• In case of non-response from few sampling units substitute the sampling units.
• Use appropriate tools for data analysis.
• Draw valid inferences.
Business Research Methods
(MGP6010A)

“How much sample should one select?”


“How large my sample be?”

Practical considerations, constraints & methods of determining


sample size

1. Unaided judgment
2. All you can afford
3. Average size for samples for similar studies
4. Required size per cell
5. Use of traditional statistical models
6. Use of Bayesian statistical model
Business Research Methods
(MGP6010A)

1. Unaided Judgment:
Client feels the particular sample size is required.

e.g. Client: “ we want a sample of 150 persons for this study”


Researcher: “ why you think this is a appropriate sample
size?”
Client: “For this problem that is about the size we need”
This is an arbitrary approach arriving at sample size.
It does not consider precision of results or cost of the sample.
This should be avoided.
Business Research Methods
(MGP6010A)

2. All you can afford:


Generally budget for conducting the research is fixed.
Researcher has to conduct the study within the budgetary limits.
Fixed cost of project is estimated (Research design, questionnaire, data
analysis, report, staff)
Fixed cost is deducted from the budget. Remaining is allotted to sampling. This
reminder of budget is divided by per unit cost of sample which returns the
sample size.
e.g. Research budget = Rs. 10,000/-
Fixed cost = Rs 5,000/-
Budget for sampling = Rs. 5,000 (10000-5000)
Cost per sample = Rs 50/-
Sample size (n) = 5000/50=100
Business Research Methods
(MGP6010A)

3. Average size for samples for similar studies-

4. Required size per cell-


Depending on the groups or subgroups (strata) and how many samples are
required from each group.

e.g. MBA Mkt = 25


FIN = 25
HR = 25
BFS = 25
TMM = 25
IEM = 25
OPR = 25
IB = 25
SPORTS = 25
HM = 25
CSR = 25
GEN = 25
Total = 300
Business Research Methods
(MGP6010A)

5. Use of traditional statistical models


For determining sample size using statistical techniques we must know-
i. The estimate of standard deviation () : The variance in the
population.
Level of homogeneity or heterogeneity of population
Lesser the SD higher the homogeneity
ii. The extent of error (e) allowed/acceptable (precision or reliability of
estimates):
How close the estimate of sample are to population parameter?
Standard error(e): standard deviation of sample means
iii. The desired level of confidence in estimate
How much confidence we expect (90%, 95%, 99%) that the
actual error will be within allowable limit i.e. less than or equal
to ‘e’(standard error)
Business Research Methods
(MGP6010A)

5. Use of traditional statistical models


Formula

z2 X 2
n = ————
e2
Where-
n = Sample size
z = standard variate at given confidence level
 = Standard deviation of population
e = acceptable error
Business Research Methods
(MGP6010A)

6. Use of Bayesian statistical model


Steps
Find out Expected Value of Sample Information (EVSI) for
each possible sample size (n)
Calculate cost of each possible sample size (n)
e.g. n = 30/40/100/200/500 etc
Find out Expected Net Gain from Sampling (ENGS) for
each sample size
ENGS = EVSI – (Cost of Sample)
The sample size with largest positive EMGS is selected.
Determining Value of Sample Information is very difficult hence this model
is rarely used.
Business Research Methods
(MGP6010A)

THANK YOU

40

You might also like