section iii: sample design reserch

Upload: ravindra-goyal

Post on 14-Apr-2018

215 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/30/2019 Section III: sample design reserch

    1/35

    SECTION III: SAMPLE DESIGN

    COVERAGE IN TOPIC:

    1. Concept of Sampling & Meaning

    2. Population and sampling

    3. Sampling frame

    4. Sampling Unit (the respondent we want

    to meet may be a family or an Individual. 5. Methods of Sampling Two Types:

    Probability Sampling and Non-Probability.

  • 7/30/2019 Section III: sample design reserch

    2/35

    1. SAMPLING CONCEPT:

    A Sample out of population is a predefined set ofpotential respondents in a geographical area.

    The most common sampling element in MarketingResearch is Human Respondent who could be :

    -Consumer,

    -A potential Consumer,

    -A Dealer or Retailer

    -A person exposed to an advertisement.

  • 7/30/2019 Section III: sample design reserch

    3/35

    Other sampling may be from :

    -Companies, or

    - Families, or

    - Households, or

    - Retail-Stores and so on.

    2. POPULATION :

    Population means not the entire population ofa given geographical area. But it is a set ofpotential respondents in a geographical area.

  • 7/30/2019 Section III: sample design reserch

    4/35

    For Example :

    (a) All mothers who buy branded BABY-

    FOOD in a given area ( say in a district or

    a city area).

    (b) All teenagers who watch MTV in the

    country. (c) All adult males who use the Shaving

    Cream PALMOLIVE.

  • 7/30/2019 Section III: sample design reserch

    5/35

    (d) All College going students who use

    LAPTOPS.

    3. SAMPLING FRAME:

    Sampling frame means a subset of the definedtarget population, from which we can select asample for our research.

    Example: Use of Telephone Directory of Jaipur to be

    used as Adult Residents of Jaipur

  • 7/30/2019 Section III: sample design reserch

    6/35

    A sampling frame is usually a practical listingof the population, or a definition of thepersons or areas which can be used for thesampling exercise. In Jaipur--

    Persons can be Doctors, Engineers,

    Advocates, Professors,

    Govt. Employees etc.

    Areas may be -- C-Scheme, Bani-Park, Raja- Park, Bapu Nagar etc.

  • 7/30/2019 Section III: sample design reserch

    7/35

    4. SAMPLING UNIT:

    In Marketing Research, there is a multi-stageselection of sample unit.

    Ist Stage Select Area - Bani Park

    IInd Stage Specific block- Near Collectorate

    IIIrd Stage may select -Names of Apartm.

    Apartments- Street/Lane H.No. IVth Stage Reach to Indvi. Respondent

    (Other Example: SHG-Bank Linkage)

  • 7/30/2019 Section III: sample design reserch

    8/35

    SAMPLE SIZE CALCULATION :

    Question : What should be the size of

    sample in Research Studies ?

    Ans.: Sample size determination is based

    on or will depend on following:

    (a) a blend of using different formulae, (say

    5% of population)

    (b) experience of similar studies, (how othershave taken sample)

  • 7/30/2019 Section III: sample design reserch

    9/35

    (c) Time & Budget constraints,

    (d) Few other elements

    Number of segments of the Target Population or Number of centers of

    study.

    (e) Analysis Requirement.

  • 7/30/2019 Section III: sample design reserch

    10/35

    5. METHODS OF SAMPLING:

    There are two Methods of sampling :-(a) Probability Sampling

    (b) Non-Probability Sampling.

    (a) Probability Sampling:- In probablesampling technique each sampling unit

    (household or individual) has a knownprobability of being included in the sample.

  • 7/30/2019 Section III: sample design reserch

    11/35

    Probability of inclusion in sample sometimes it isequal and sometime it is unequal. Probablesamples are unbiased.

    The major types of probability samplingtechniques are as under :-

    1. Simple Random Sampling.

    2. Stratified random sampling

    3. Cluster sampling

    4. Systematic sampling

    5. Multistage or Combination sampling.

  • 7/30/2019 Section III: sample design reserch

    12/35

    (b) Non-probability sampling Techniques:

    (1) Quota Sampling ( a fixed number)

    (2) Judgment sampling

    (3) Convenience Sampling

    (4) Snowball Sampling.

    Now, we will discuss these sampling techniques

    in detail one by one as under:

  • 7/30/2019 Section III: sample design reserch

    13/35

    1. SIMPLE RANDOM SAMPLING:

    When we select randomly the requirednumber of sample of unit out of a totalnumber of units, it is termed as simplerandom sample. (any 5 students in class)

    For example : the average income level of100 employees of a Company can be knownby writing number 1 to 100 on a slip of paperand their income is written on it by

    them.Then, if we need a quick estimate

  • 7/30/2019 Section III: sample design reserch

    14/35

    then we have to decide with a sample ofany 5/10 slips of paper out of 100 (aftermixing them by shuffling) and use theseemployees as our sample for knowingtheir average income. It is known assimple random sampling.

    For the same information if we

    interview all the 100 employees, andcollect the information then it will take

  • 7/30/2019 Section III: sample design reserch

    15/35

    more time.

    Simple Random sample is more usefulwhere exact number of respondents are not

    known and we require certain sample. 2. STRATIFIED RANDOM SAMPLING:

    In this technique, the total targetpopulation is divided into Strata or Segments

    on the basis of some important variables(Variables may be Age, Income, educationlevel Science, Commerce, Arts, Engineeringetc.).

  • 7/30/2019 Section III: sample design reserch

    16/35

    For Example: a consumer populationmay be divided into Age brackets ofbelow 25 years, 25 to 40 years and above

    40 years. Then, a sample is taken fromeach of the segment defined (3segments). Then as per need sample canbe taken from each segment. This type of

    sampling is known as Stratified RandomSampling.

  • 7/30/2019 Section III: sample design reserch

    17/35

    Example:2: Total population of a Blockor Area is 50,000. Block or Area isdivided into 10 segments. 5% sample

    is decided to be taken by theResearcher. Total sample would be2,500. Thus, 250 sample from eachsegment (10) will be taken.

    We can take proportionate samplealso based on population of the Blockor Area.

  • 7/30/2019 Section III: sample design reserch

    18/35

    Most of the population do not fallinto extreme zones, and generallyStratified Sampling is the most

    efficient method of probabilisticsampling, if it is feasible.

    3. CLUSTER SAMPLING/AREASAMPLING:

    In Cluster Sampling a group ofobjects/Units for sampling is selected.

  • 7/30/2019 Section III: sample design reserch

    19/35

    A Cluster is a group of sampling units orelements, which can be identified, listed, and asample of which can be chosen.

    (Example: Selection of MBA students from

    different MBA/PGDM Institutions in Jaipur) Cluster represents

    -(i) Geographical Areas,( say avillage/Block/Town/District etc.)

    -(ii) Membership of some Group: CA, CS Members of Lions Club etc.

    -(iii) Members of Cultural Club.

  • 7/30/2019 Section III: sample design reserch

    20/35

    (iv) Members of Sports Club.

    (v) Members of some Social Organization

    e.g. Agrawal Samaj/Jain Samaj etc.

    (vi) MP/MLA etc.

    Marketing Researchers use cluster ofHouseholds, located in a city or block orcolony or any specified area.

    When the Clusters are selected on the basisof Geographical area, it is also called AreaSampling. You can define the area.

  • 7/30/2019 Section III: sample design reserch

    21/35

    In Cluster Sampling the Researcher should:

    (i)Prepare a list of all available clusters

    (ii) All Clusters should be numbered as

    1,2,3,4,5,6,7,8,9,10,11,12, and so on. (iii) A sample of clusters (number to be decided by

    Researcher) should be randomly drawn say-2,4,6,8,10,12 etc. or 1,3,5,7,9,11 etc.

  • 7/30/2019 Section III: sample design reserch

    22/35

    (iv) All sampling units/elements such asHousehold in the selected clusters should bechosen to be a part of the sample.

    Advantage of Cluster Sampling: It is usually low cost oriented;

    It is convenient to Researcher;

    Disadvantage:- Members of cluster tend to besimilar same socio-economic background,similar tastes, and buying behaviour.

  • 7/30/2019 Section III: sample design reserch

    23/35

    4. SYSTEMATIC SAMPLING:

    Systematic Sampling is just likesimple random sampling. Here wedecide sample size. We divide the total

    population into parts. Example of IRM:

    IRM, Jaipur have 360 students in the

    Institute. For research purpose, weneed a sample of 20 students out of360.

  • 7/30/2019 Section III: sample design reserch

    24/35

    The sampling fraction is 360/20 whichmeans 1 out of every18 students will beselected on an average basis. We divide

    the list of 360 into 18 parts. Out of thefirst 18 students we choose any one atrandom.

    Let us say, we choose student Number 9

    (all students are listed from No.1 to 360).

  • 7/30/2019 Section III: sample design reserch

    25/35

    We choose student numbers 9 + 18 then

    9 + 18 + 18 and so on in a systematicsampling plan.

    Thereafter the selected students will benumbered as9,27,45,63,81,99,117,135,153,171,189,

    207, 225,243, 261, 279, 297, 315 ,333,and 351 (total 20 students).

  • 7/30/2019 Section III: sample design reserch

    26/35

    All these 20 students will be sample forthe study which is known as systematicsampling.

    5. MULTI-STAGE OR COMBINATIONSAMPLING:

    (a) In this type of Sampling combinationof any two or more than two samplingmethods is used, e.g. Combination ofCluster sampling and Stratified random

  • 7/30/2019 Section III: sample design reserch

    27/35

    (segment) sampling.

    (b) We have combined TWO or more differentmethods of probability sampling.

    (c) This type of sampling is used when it isdone at National Level Research.

    In such a research, we may divide India into 5Metro Clusters, then 20 Class A towns/Cities,

    and then 200 Class B towns/Cities.

  • 7/30/2019 Section III: sample design reserch

    28/35

    At Ist Stage:

    5 Metro Clusters

    20 Class A Towns/Cities

    200 Class B Towns/Cities

    At 2nd Stage: (Out of Ist stage)

    1 Metro Cluster

    3 Class A Towns/Cities

    10 Class B Towns/Cities

  • 7/30/2019 Section III: sample design reserch

    29/35

    In 2nd Stage we may have sample planas Stratified/Segment Sampling.

    In 3rd Stage we may decide about

    House-hold Income or Age ofRespondents where the cluster samplingcan be used by the researcher.

    Thus, at different stages, we may usedifferent sampling methods to completethe Research work.

  • 7/30/2019 Section III: sample design reserch

    30/35

    NON-PROBABILITY SAMPLINGTECHNIQUES:

    Question : When we use Non-probability

    Sampling Technique/Method?

    Answer: When it is not feasible to useprobability based methods, researcher may useNon-probability methods.

    For approximate sampling sometimesResearcher use the Non-probability Samplingmethods, and these are as under :-

  • 7/30/2019 Section III: sample design reserch

    31/35

    1. Quota Sampling :- Quotasampling is just like Segmentsampling. As per the judgment and

    Researcher the sample is taken inquota sampling method.

    2. Judgment Sampling:- The

    Researcher choose the sample as perhis choice and judgment. Thus,biasness of Researcher is there.

  • 7/30/2019 Section III: sample design reserch

    32/35

    3. Convenience Sampling:- Thistype of sampling is taken by Researcheras per his convenience. Example: TV

    Reporters take sample interview in themarket which is convenient to him.Whosoever meets the TV reporter, hetakes interview and record the matter.

    Proper identification or selection is notdone in this method.

  • 7/30/2019 Section III: sample design reserch

    33/35

    4. Snowball Sampling :- One respondent isselected to generate names of others is calledsnowballing. In this type of sampling therespondents are having net-working and they all

    know each other. This type of sampling is donewhere population is small and are selective.

    For Example : 1. Golf-players.

    2. Pilots of Airlines

  • 7/30/2019 Section III: sample design reserch

    34/35

    3. Owners of Honda City Car.

    4. Owners of Mercedeze Benz Car.

    ERRORS IN MARKETING RESEARCH:

    Two Types of Errors may be there:

    (i) Sampling Error: Such error is done bythe Researcher. It is controllable.

    (ii) Non-sampling Error:- Error done byInterviewer, or data entry operator or theResearcher himself.

  • 7/30/2019 Section III: sample design reserch

    35/35

    Examples of such Errors:

    (1) Interchange of column Yes or No.

    (2) Not filling data in field. (3) Cheating by Interviewer (data filled

    without taking interview of the

    respondents. =======