The document discusses various sampling methods used in research. It describes probability sampling methods like simple random sampling, systematic sampling, stratified sampling, cluster sampling, and multistage sampling. It also discusses non-probability sampling methods such as purposive sampling, volunteer sampling, snowball sampling, quota sampling. The document provides examples and formulas for calculating sample sizes for both categorical and continuous variables to achieve the desired level of precision and confidence level. It emphasizes the importance of random selection and adequate sample sizes to obtain results representative of the population.
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
SAMPLING methods d p singh .ppt
1. Types of Sampling
• 1. Probability Sampling
• 2. Non Probability Sampling
1
Dr. D. P. Singh, School of Research Methodology, TISS
2. Probability Sampling Methods
1. Simple Random Sampling
2. Systematic Sampling
3. Stratified Sampling
4. Cluster Sampling
5. Multistage Sampling
2
Dr. D. P. Singh, School of Research Methodology, TISS
3. 1. Simple Random Sampling : A Simple Random Sample of n
subjects from a population is one in which each possible
sample of that size has the same probability of being selected.
How to Select a Simple Random Sample?
• Before selecting a Random Sample, we need a list of all
subjects in the population. The list is called Sampling Frame.
The most common method for selecting a simple random
sample from the sampling frame is the use of a random
number table.
• A random number table is a table containing a sequence of
numbers that is computer generated according to a scheme
whereby each digit is equally likely to be any of the integers
0,1,2, ... ,9
3
Dr. D. P. Singh, School of Research Methodology, TISS
5. 2. Systematic Sampling: If a sample of size n is
to be selected from a population size by N and let
k = N/n. A systematic random sample
1. Selects a subject at random from the first k names in the
sampling frame, and
1. Selects every kth subject listed after that one. The
number k is called the skip number.
k = 3
Random Start
5
Dr. D. P. Singh, School of Research Methodology, TISS
6. 3. Stratified Sampling : A stratified random
sample divides the population into separate
groups, called strata and then selects a simple
random sample from each stratum.
The population is divided into k homogeneous
strata with stratum size N1, N2, …, Nk such
that N1 + N2 + …+ Nk=N
N1 N2
Nk
+ +……+ = N
n1 n2
nk
+ +……+ = n
6
Dr. D. P. Singh, School of Research Methodology, TISS
7. Comparison of SRS, STRS and SYRS
Sample 1 2 3 4 5 6 7 8 9 mean
Population 2 2 2 5 5 5 8 8 8 5
Simple Random Sampling
2 2 2 2 Underestimate
5 5 5 5
8 8 8 8 Overestimate
1 2 5 8 5
2 2 5 8 5
Dr. D. P. Singh, School of Research Methodology, TISS
7
8. 4. Cluster Sampling: Simple, systematic and stratified
random sampling are very expensive or ever impossible to
implement in many situations particularly when a
complete and up to date sampling frame is not available.
In cluster sampling, the population is divided in to
large number of groups, called clusters. A cluster
sample is one for which the sampling units are the
subjects in a random sample of the clusters.
8
Dr. D. P. Singh, School of Research Methodology, TISS
9. 5. Multistage Sampling: Multi-stage sampling methods
use combinations of various sampling techniques. For
example, to study various characteristics of adult residents
in Maharashtra state, one could treat districts as clusters
and select a random sample of a certain number of them.
Now within each district selected one could take cluster
sample of villages. Within each village selected, one could
systematically sample every 10th household. Within each
household selected, one could select one adult at random
from the sample.
9
Dr. D. P. Singh, School of Research Methodology, TISS
10. Non-Probability Sampling Methods
Social research is often conducted in situations
that do not allow the kinds of probability
sampling discussed so far, for large-scale social
surveys. Suppose we want to study
homelessness. Neither a list of all homeless
individuals is available nor it can be created.
Moreover, there are times when probability
sampling wouldn’t be appropriate. Such
situations call for non-probability sampling.
10
Dr. D. P. Singh, School of Research Methodology, TISS
11. Methods of Non-Probability Sampling
1. Purposive or Judgement Sampling;
2. Volunteer Sampling;
3. Snowball Sampling;
4. Quota Sampling; and
Selecting Informants
11
Dr. D. P. Singh, School of Research Methodology, TISS
12. 1. Purposive or Judgement Sampling
Sometimes it is appropriate to select a sample on the basis of knowledge of
a population and the purpose the study. This type of sampling is called
purposive or judgement sampling.
2. Volunteer Sampling
One of the most common non-probability sampling methods is volunteer
sampling. In this method subjects volunteer themselves for the sample. A
good example of volunteer sampling is visible almost any day on television.
Some T.V. programmes ask viewers to offer their opinions on any issue or
vote for any celebrity for his/her performance through SMS or calling a phone
number. The danger in this method is that the sample will poorly represent the
population and may yield misleading conclusions.
12
Dr. D. P. Singh, School of Research Methodology, TISS
13. 1. Snowball Sampling
Snowball sampling method is appropriate when the
members of a special population or individuals with a
rare characteristic are difficult to locate such as
persons in a village who were bitten by snake. In this
method, the researcher collects data from the few
members of the target population; then asks those
individuals to provide information needed to locate
other members of the target population whom they
happen to know.
13
Dr. D. P. Singh, School of Research Methodology, TISS
14. 1. Quota Sampling
Quota sampling begins with a matrix, or table,
describing the relevant characteristics of the target
population. Depending on your research purposes, you
may need to know what proportion of the population is
male and what proportion female as well as what
proportion of each gender fall into various age
categories, educational levels, ethnic groups, and so
forth.
Once you have created such matrix and assigned a
relative proportion to each cell in the matrix, you
proceed to collect data from people having all the
characteristics of a given cell. You then assign a weight
to all the people in a given cell that is appropriate to
their total population. When all the sample elements are
so weighted , the overall data should provide a
reasonable representation of the total population.
14
Dr. D. P. Singh, School of Research Methodology, TISS
15. Sampling Error
The sampling error of a statistic is the error
that occurs when a statistic based on a sample
estimates or predicts the value of a
population parameter.
1. Other Sources of Variability/Error
2. Under-coverage
3. Response Bias
4. Non-response
5. Missing Data
15
Dr. D. P. Singh, School of Research Methodology, TISS
16. Sample Size
• Feasible and Desirable
• Time, Money, Manpower and transport
• More variation more samples
• Estimates, confidence, margin of error
Dr. D. P. Singh, School of Research Methodology, TISS 16
17. Sample size calculation
• Proportion
• N = (1.96)2 * p * (1-p) / E2
• Where 1.96 is t test value at 95% confidence level
• P is proportion of prevalence taking .5 gives optimal sample size
• Changing p from .5 will lead to reduction in sample size
• E is margin error reducing it from .05 to .02 will increase sample
size
• An illustration with 95% confidence level, margin of error 5 %, p
value .5
Dr. D. P. Singh, School of Research Methodology, TISS 17
18. One Sample, Dichotomous Outcome (p)
• Confidence Level (e.g., 95%) 95%
• Estimate of true proportion (p),
• if unknown, enter 0.5 0.5
• Margin of Error (E) 0.05
• Sample Size Required 385
• Attrition? Enter Percentage Expected to be Lost/Missing Data
0%
• Sample Size Required, Accounting for Attrition
• n = 385
Dr. D. P. Singh, School of Research Methodology, TISS 18
19. Example : p=0.1, E=.05, CI-95%
• Confidence Level (e.g., 95%) 95%
• Estimate of true proportion (p),
• if unknown, enter 0.5 0.1
• Margin of Error (E) 0.05
• Sample Size Required 139
• Attrition? Enter Percentage Expected to be
Lost/Missing Data 0%
• Sample Size Required, Accounting for Attrition
• n = 139
Dr. D. P. Singh, School of Research Methodology, TISS 19
20. Population Size Sample size for permissible Error (proportion)
0.05 0.03 0.02 0.01
100 79 91 96 99
200 132 168 185 196
500 217 340 414 475
1000 278 516 706 906
10000 370 964 1936 4899
100000 382 1056 2345 8762
1000000 384 1066 2395 9513
Dr. D. P. Singh, School of
Research Methodology,
TISS 20
21. Estimate for continuous variable
Dr. D. P. Singh, School of Research Methodology, TISS 21
Where z is the value of standard normal distribution based on pre
selected confidence level (level (for example 1.645 for 90%
confidence level, 1.96 for 95% confidence level and 2.64 for 99%
Confidence level). σ is Standard deviation of outcome variable,
and E is margin of error – 5% or 1%
Remember smaller margin of error would require more sample
and similarly higher confidence level require larger sample
22. Example 1
• Suppose a sample of persons suffering from blood pressure in a industrial town is to
be selected by simple random sampling to estimate the level of diastolic blood
pressure. The precision is to be ± 5 with 95% confidence. An earlier survey found
the standard deviation as 20. Find the sample size for a population size of 1000 and
10,000.
• Substituting the information in relation no = t² S²/ d²
•
• no = (1.96)² * (20)² / (5)² = 61.46 or 62 approx.
•
• Since no/N=.062 Using equation (iv) we get 62/{1+(62/1000)}= 59
•
• For a population of 10,000, the ratio no/N is negligible and one can take sample of
62 as optimum sample.
•
Dr. D. P. Singh, School of Research Methodology, TISS 22