These slides are related to statistics. This is an detailed version of the topic. This slide discusses about various methods of sampling and also tells us about the method of planning and executing any particular survey.
3. POPULATION AND SAMPLE
ā¢ An aggregate or individual items relating to a phenomenon under
investigation is called as āpopulationā.
ā¢ When few items are selected for statistical enquiry from a given
population, it is called a āsampleā.
ā¢ A sample is the subset of the population.
4. CENSUS AND SAMPLE METHOD
ā¢ In a census, data about all individual units (e.g. people or households)
are collected in the population.
ā¢ In a survey, data are only collected for a sub-part of the population;
this part is called a sample.
ā¢ Results obtained by census are more reliable than sample survey.
ā¢ But census survey are very costly and time consuming.
5. ORGANIZATION OF STATISTICAL SURVEY
ā¢ It deals with organisation and execution of large scale sample survey.
ā¢ They are considered under two major heads:
1. Planning the survey
2. Execution of the survey
6. PLANNING THE SURVEY
1. Purpose of survey
ā¢ Purpose should be clear
2. Scope of survey
ā¢ Depends on purpose and availability of time and resources
3. Nature of information required
ā¢ Should decide separately
ā¢ Depends on the likely use
7. PLANNING THE SURVEY
4. Units to be used
ā¢ There are two kinds of units: units of collection and units of analysis and
interpretation
ā¢ Units of collection is again divided to 2: simple units and composite units
ā¢ Simple units are single conditions with no restrictions. E.g. day, hour
(mensurational units) , rupee (pecuniary value unit) , house (produced units)
etc.
ā¢ Composite unit is a simple unit with some qualification. E.g. skilled worker. It
enables quicker comparison.
8. PLANNING THE SURVEY
4. Units to be used
ā¢ Unit can be arbitrary or conventional.
ā¢ Properties required:
ļ¶ Simple to understand
ļ¶ Suit the purpose
ļ¶ Should not be unclear
ļ¶ Should be usable throughout the survey
9. PLANNING THE SURVEY
5. Sources of data
ā¢ Primary data refers to the first hand data gathered by the researcher himself.
ā¢ Secondary data means data compiled from published or unpublished sources.
6. Techniques to be adopted
ā¢ Census survey and Sample survey
7. Choice of frame
ā¢ Frame ā list of the units of survey
ā¢ Frame may be inaccurate, incomplete, could have duplication, inadequate or
out of date
10. PLANNING THE SURVEY
8. Accuracy aimed
ā¢ Absolute accuracy is unattainable
ā¢ Degree of accuracy depends upon object of enquiry.
9. Other considerations
ā¢ The investigator should consider whether the enquiry is
i. Official or semi official or non-official
ii. Confidential or not
iii. Regular or ad hoc
iv. Initial or repetitive
v. Direct or indirect
11. EXECUTION OF THE SURVEY
1. Setting up an administration organization
ā¢ Existing organization is to be utilized or new one to set up
ā¢ Central and regional officed can be set as per needs
2. Designing of forms
ā¢ Questionnaires or schedules or other forms
3. Select, train, supervise field investigators
ā¢ Select proper persons for field work, impart uniform training and supervise
their field work
12. EXECUTION OF THE SURVEY
4. Control accuracy of the fieldwork
ā¢ Close watch on progress would help
ā¢ Periodical and sample checks are useful
5. Reduce non-response
ā¢ Steps are to be taken to collect the information
6. Present information
ā¢ Collected information is to presented in statistical tables, diagrams and
graphs.
13. EXECUTION OF THE SURVEY
7. Analyzing information
ā¢ Purpose of the survey is achieved at this stage.
ā¢ Collected data is analyzed to find details
8. Prepare reports
ā¢ Reports show the purpose of the survey, personnel involved, the time, the
mode of collection, the accuracy, the nature, the coverage and source of the
information, etc.
15. PRINCIPLES OF SAMPLING
ā¢ Two important principles upon which the theory of sampling is based:
1. Principle of Statistical Regularity
2. Principle of Inertia of Large Numbers
16. STATISTICAL REGULARITY
ā¢ āThe law of statistical regularity lays down that a moderately large
number of items chosen at random from a large group are almost
sure on the average to possess the characteristics of the large group.ā
ā¢ A sample selected at random will be a representative sample.
ā¢ A sample study saves time and money and its usage is justified by this
principle.
17. INERTIA OF LARGE NUMBERS
ā¢ As per this principle, other things being equal, larger the size of
sample more accurate the results are likely to be.
ā¢ E.g. Tossing 10000 coins could give head almost 5000 times but we
canāt say with the much confidence that we get 5 head out of 10
tosses.
18. METHODS OF SAMPLING
ā¢ Sampling method can be grouped into two viz. random sampling and
non random sampling.
ā¢ Random sampling is known as probability sampling or chance
sampling.
ā¢ The different types of probability samplings are
i. Where each unit has āequal chanceā of being selected
ii. Sampling units have different probability of being selected
iii. Probability of selection of a unit is proportional to the sample size
19. METHODS OF SAMPLING
ā¢ Non probability sampling doesnāt afford any basis for estimating the
probability that each item in the population has of being included in
the sample.
ā¢ Samples are selected according to a fixed sampling rule and not by
assigning probabilities.
20. METHODS OF RANDOM SAMPLING
1. Simple random sampling
2. Stratified random sampling
3. Systematic sampling
4. Cluster sampling
21. 1. SIMPLE RANDOM SAMPLING (srs)
ā¢ Each unit in the population gets an equal chance of being represented
in a sample
1. srs with replacement (srswr)
In this, selected units are replaced. So, a unit can happen more than once in the
sample.
2. srs without replacement (srswor)
In this, selected elements are not replaced before next draw. So, there is no
possibility of same element to occur more than once in the sample.
22. METHODS OF SIMPLE RANDOM SAMPLING
ā¢ Selected by two methods
1. Lottery method
2. Random number method
23. LOTTERY METHOD
ā¢ Useful when population is comparatively small
a. Assign serial numbers like 1,2,3ā¦ to each item
b. Make lots of each number and shuffle
c. Take one lot at random
d. Note the number and select the item
e. Repeat until we get the req. number of items to sample
24. RANDOM NUMBER METHOD
ā¢ Adopted with the aid of random number tables
ā¢ These are table of numbers in which digits are selected by mechanical
process of randomization are tabulated.
i. Assign serial numbers to sampling unit
ii. Open any page of the random number table
iii. Select a number blind-foldedly
iv. Start from the number and proceed along a row, column or
diagonally the successive numbers
25. RANDOM NUMBER METHOD
ā¢ Three such random number tables are noteworthy
a. Tippettās table of random numbers
b. Kendall and Babington-Smith numbers
c. Fisher and Yateās numbers
ā¢ Tippettās number consist of 41,600 digits taken from census reports
and combined by fours to give 10,400 4-digit numbers.
ā¢ Method of Consulting Tippetās Table
26. MERITS
ā¢ Saves money and time
ā¢ Eliminates possibility of biased errors
ā¢ Errors neutralize themselves and becomes more representative of
aggregate
ā¢ No detailed plan needed for selection of the samples
ā¢ Has possibility of testing the accuracy of a sample by examining
another sample from same universe
27. DEMERITS
ā¢ Canāt be applied where some units of universe are so important
where their inclusion is essential
ā¢ Sample should be large, else it canāt be truly representative of the
whole universe
ā¢ Impossible to identify every population member
ā¢ Expensive and time consuming if population is large.
28. STRATIFIED RANDOM SAMPLING
ā¢ If population is heterogeneous but can be divided to homogeneous
sub-groups (strata), this method can be applied.
ā¢ This would make a stratified random sample.
ā¢ It is either proportional or non-proportional.
ā¢ In non proportional sample certain strata are represented in large
proportion to the population than others.
29. MERITS
ā¢ The sample becomes more representative since each stratum is
represented in the sample
ā¢ More precision since strata are homogeneous
ā¢ Leads to administrative convenience
30. DEMERITS
ā¢ Any wrong in the formation of different strata would make results
quite unreliable.
ā¢ Difficult to decide basis and number of stratifications
ā¢ Tedious and time consuming
31. SYSTEMATIC SAMPLING
ā¢ Sample units are arranged in definite order, say, alphabetical,
chronological, geographical etc.
a) Arrange sampling units in definite order
b) Give them serial numbers 1,2.3ā¦
c) Divide them into groups equal to required sample size
d) From first group choose an item using lottery method
e) If we got 7th item, choose 7th item of every group systematically at
equally spaced intervals
32. SYSTEMATIC SAMPLING
ā¢ Systematic sampling is often equal to random sampling in results.
ā¢ Merit: It is useful if similar parts of the population are tend to be
grouped together.
ā¢ Demerit: Not useful if there is some periodic variation in the
population corresponding sampling interval.
33. CLUSTER SAMPLING
ā¢ Cluster sampling is the selection of sample units in two stages.
1. Certain groups or clusters called primary sampling units are
selected from the population
2. Individual items called elementary sampling units are drawn from
each clusters. It is also called sub-sampling.
ā¢ When cluster is contained in geographic area, cluster sampling is
called area sampling.
34. ADVANTAGES
ā¢ Minimizes cost per elementary sample unit
ā¢ Permits grouping of observations for easy coverage
ā¢ Interviewers can also concentrate on few clusters and save time and
expense
36. ADVANTAGES OF SAMPLING
1. Comparatively more economical
2. Ensures completeness and a high degree of accuracy
3. Possible to obtain more detailed information
4. Advocated when census is not necessary or desirable
5. In some cases, sampling is the only feasible method.
6. More scientific that census since the reliability of results can be
known
37. LIMITATIONS OF SAMPLING
1. A sample survey must be properly planned and executed else it
could give incorrect and misleading results.
2. Sampling requires services of experts
3. Not useful when one needs minute details of individual constituent
of a universe
4. Various sources of errors in a sample survey
5. If the sample is not truly representative and wrong sampling
method is selected, then sample will fail to give true characteristics
of the population.
38. SAMPLING AND NON-SAMPLING ERRORS
ā¢ The errors that may occur due to the collection, classification,
processing and analysis of data may be broadly classified into two:
1. Sampling errors
2. Non sampling errors
39. SAMPLING ERRORS
ā¢ Sampling errors occurs only in sampling
ā¢ Statistic ā Any value computed using sample observations
ā¢ Parameter ā Any value computed from a population
ā¢ Sampling error can be defined as error between statistic and
parameter.
ā¢ Sampling error can be reduced to a great extent by increasing the
sample size and using appropriate sampling method.
40. SAMPLING ERRORS
ā¢ Sampling error occur due to the following reasons.
1. Faulty selection of sample
2. Substitution
3. Faulty demarcation of sampling units
4. Improper choice of the statistics as estimators
41. NON SAMPLING ERRORS
ā¢ Non sampling errors arise at the stages of data collection,
classification, analysis and interpretation of the data
ā¢ Some sampling errors arise from following factors:
1. Wrong definition
2. Response errors
3. Non-response bias
4. Errors in coverage
5. Compiling errors
6. Publication errors
42. PREPARATION OF QUESTIONNAIRE
ā¢ Primary data is collected using questionnaire or schedule.
ā¢ The questionnaire is ordinarily filled up by the informant while
schedule is filled by the trained enumerator who questions the
informants.
ā¢ Following are the tips while drafting a good questionnaire or
schedule:
43. PREPARATION OF QUESTIONNAIRE
1. Few questions
2. Simple and easy to understand
3. Not ambiguous
4. Max. Yes or No should be given
5. Need to be answered without
bias
6. Not be unnecessarily
inquisitorial
7. Include specific information
questions
8. Avoid open questions
9. Instruct the informants
10. Objective answer questions
11. Questionnaire should Look
attractive
12. Pre test questionnaire
13. Donāt ask question with
calculations