2. Summary
The Case contemplateisaboutIndianAirconditionerIndustry.Someshortof the Indianbuyerdurable
industry. The aggregate market estimate is of around Rs.250,000 million. The Home Appliance
Industrymeasure isaroundRs87,500 million.Among them,Refrigeratorcontributesthe biggestoffer
of Rs 38000 million followed by room forced air systems at around Rs 27,500 million. Also, clothes
washer at Rs 14,000 million. Among them, Air conditioner Industry appreciates the most elevated
developmentof 20% everyyear.Because of a decrease inextractobligationandadecrease inimport
obligations limited the hole among sorted out and sloppy segment. The Share of the Unorganized
market was 70% of every 1980 has dropped down and is currently 25%. Expanding in the extra cash
and change inthe way of life are the main considerationsthatbackingsthe interestof the forced air
system industry.
What will be your sampling frame, appropriate sampling techniques, sample
size, andsampling process?
A perfectinspectingedge will have the accompanyingcharacteristics:
1. all unitshave a sensible,numerical identifier
2. all unitscan be found – theircontactdata, delineateorotherpertinentdataisavailable
3. the casing issortedoutin a coherent,deliberate form
4. the casing hasextradata about the unitsthat permitthe utilizationof furtherdeveloped
inspectingoutlines
5. each componentof the numberof inhabitantsinintrigueisavailable inthe casing
6. each componentof the populace isavailable justonce inthe casing
7. no componentsfromoutside the numberof inhabitantsinintrigue are available inthe edge
8. the informationis up-to-date
If we talkaboutthe samplingframe,thenforthe ACframe shouldall the followingqualities:
i) The frame shouldcontainvariable like the qualityof all the powerconsumptionof units,
i.e., it should have a logical number identifier.
ii) The location of the frame should be having the relevant information of the people from
where the surveyis takenwhichmeansthe sectorand the area for the surveyshouldbe
defined.
iii) The data for the same should be of the recent one.
iv) The elements or the variable which are present for the frame would be of the logical
reasoning and should be present only once.
The main product features which are expected by customers.
Thermostat
Adjustable Fan Speed
Remote
Programmable Timer
Air Direction Control
3. For the most of the cases,in the AC production, we identifyandmeasure everysingle itemknownas
direct element sampling.
Thismethodwouldallowustomeasureevery singlecomponentwhichisinvolvedinthe measurement
of the qualities of the ACs that are installed into the sector whether it is the organized one or the
unorganized one. The most effective way would be an opinion poll, which would be taking into
considerationof the frameswhichcaninclude employmentrecords,the salaryof the employees,the
population of the same sector, the awareness about the same, the availability of the product, the
other like guarantee & warranty of the product users can also be included.
In the most straightforward cases, such as when dealing with a batch of material from a production
run,or usinga census,itispossibletoidentifyandmeasure everysingleiteminthe populationandto
include any one of them in our sample; thisis known as direct element sampling .However, in many
other cases this is not possible; either because it is cost-prohibitive or impossible.
The surveyshouldbe done thatthe coveredarea containseverydimensionof the societysothatthe
statistics results are according to the requirement.
Before we cancalculate asample size,we needtodetermine afew thingsaboutthe targetpopulation
and the sample we need:
1. Population Size
2. Margin of Error (Confidence Interval)
3. Confidence Level
Sampling Frame
(Done at Organized Sector)
This includes also the metropolitan cities
Direct Element Sampling
Sample Size
i) Estimation of Proportion
ii) Estimation of Mean
RequiredSample SizethroughTables
Sampling Process: 7 Steps
Sampling Technique
Probability Sampling
Stratified Sampling
Error
Sampling Error
Non-Sampling Error
4. 4. Standard of Deviation
Necessary Sample Size = (Z-score)2 * StdDev*(1-StdDev) / (margin of error)2
Here is an example ofhow the math works assuming you chose a 95% confidence level, .6 standard deviation, and a margin oferror
(confidence interval) of+/-4%.
((1.96)2 x .6(.6)) / (.04)2
(3.8416 x .36) / .0016
1.3829 / .0016
864.36
865 respondents are needed
A sample size of the data could be determinedbyusingthe way where the confidence interval plays
a major role, i.e., the larger the required confidence level, the larger the sample size.
The largersample size for the ACs companieswouldgenerallyleadtothe increase inprecisionwhen
the estimation of the unknown parameters is done.
For the calculationof the sample size,researcherinthe ACcompaniesshouldtake intoconsideration
where the total size of the populationpresentinthe organizedaswellasthe unorganizedsectormust
be kept in mind. Secondly, the margin of error, the reliability on the fact that data which is been
collectedbyme doesnot give me the garbage informationbutisaccurate; and the confidence level,
which is the probability of margin of error should be taken into consideration.
Estimationof a proportion
Usingthis approximation,itcanbe shownthat around95% of thisdistribution'sprobabilitylies
within2 standarddeviationsof the mean.Usingthe Waldmethodforthe binomial distribution,an
interval of the form
Estimationof a mean
Thisexpressiondescribesquantitativelyhow the estimatebecomesmore preciseasthe sample size
increases.Usingthe central limittheorem tojustifyapproximatingthe samplemeanwithanormal
distributionyieldsanapproximate95%confidence interval of the form
The requiredsamplesizeforthesame couldbedeterminedfromthepre-determinedtablesforcertain
values. These tables would contain the values as the desired power of the trial.
5. The table showncan be usedin a two-sample t-testtoestimate the sample sizesof an experimental
groupand a control group that are of equal size,thatis, the total numberof individualsinthe trial is
twice thatof the numbergiven,andthe desired significance level is 0.05. The parameters used are:
The desired statistical power of the trial
Cohen'sd (=effectsize),whichisthe expecteddifferencebetween the meansof the targetvalues
between the experimental group and the control group, divided by the expected standard
deviation.
Dangers of Small Sample Size
Large sample sizesare neededforastatistictobe accurate and reliable,especiallyif itsfindingsare to
be extrapolated to a larger population or group of data. Say you were conducting a survey about
exercise andinterviewedfivepeople,twoof whomsaidtheyrunamarathonannually.If youtake this
survey to represent the population of the country as a whole, then according to your research, 40
percentof peoplerunatleastone marathonannually-- anunexpectedlyhighpercentage. The smaller
your sample size, the more likely outliers -- unusual pieces of data -- are to skew your findings.
For the sampling process, there are 7 steps that can be done:
i) Definingthe Population(Target):Thiswouldbe the mostimportantstepof the sampling
processforthe analystwhowantstosee the reactionof the population.The analystcould
take out the populationwhichisof the middle classandabove thus includingthe person
whocan affordto buythe AC andcan maintainitwhenthe replacementof the productis
needed.
ii) Specifying the Sampling Frame: The Sampling Frame would contain the variables which
are used to define the sample thus, thus specifying it would mean making use of the
techniquesandprocedureswhich givethe bestoutputtothe AC Company.Thiscouldbe
anything like the durability of product or variables used in designing of the ACs.
iii) Specifying the Sampling Unit: The specific specifying of the unit implies how much
importance we are giving to each variable. As the income of the person is an important
6. variable andshouldnotbe ignoredwhiletakingthe samplingunit.Thusmakinguseof this
will make the researcher clear about the situation.
iv) SelectionofSamplingMethod: The samplesize playsacrucial roleinthesamplingprocess.
There are variouswaysof classifyingthe techniquesusedindeterminingthe sample size.
A couple those hold primary importance and are worth mentioning are whether the
technique deals with fixed or sequential sampling and whether its logic is based on
traditional or Bayesian methods.
v) Determinationof Sample Size: Inthisstep,the specificationsanddecisionsregardingthe
implementation of the research process are outlined. Suppose, blocks in a city are the
sampling units and the households are the sampling elements. This step outlines the
modus operandi of the sampling plan in identifying houses based on specified
characteristics.
vi) Specifyingthe Sampling Plan: This is the final step in the sampling process, where the
actual selection of the sample elements is carried out. At this stage, it is necessary that
the interviewerssticktothe rulesoutlinedforthe smoothimplementationof the business
research. This step involves implementing the sampling planto select the sampling plan
to select a sample required for the survey.
Stratified Sampling:
A stratified sampling approach is most effective when three conditions are met
1. Variability within strata are minimized
2. Variability between strata are maximized
3. The variables upon which the population is stratified are strongly correlated with the
desired dependent variable.
Advantages over other sampling methods
1. Focuses on important subpopulations and ignores irrelevant ones.
2. Allows use of different sampling techniques for different subpopulations.
3. Improves the accuracy/efficiency of estimation.
4. Permits greater balancing of statistical power of tests of differences between strata by
sampling equal numbers from strata varying widely in size.
This type of sampling would be mostly used due to the following advantages:
i) The companywouldgetthe measurementwithoutgettingleasterror,thusitmeansthat
the error generated in the sampling process by the analyst would be less and would be
profitable to the AC Company.
ii) The measurementwouldbecomemore manageableandthe company hasthe knowledge
about the same for the how things can be arranged.
In this,asan analyst,I can formstrata to the variousgroupslike the personpresentinthe industryor
the income basis. This strata formed will have every variable that I am looking for. Then from the
strata, the survey should be done so that we can get the thought process of the whole strata easily.
Thenforgettingthe resulttobe inrange if the resultisnotaccurate,we couldtake the meanof every
7. sample whichwouldbe allowingusto the normallydistributedcurve,thusmakingour predictionor
the research work for the company almost appropriate.
Stratified Sampling
On the basis on durability of
Product
Industrial People
People of Different
Environment
Sample
Unorganized Sector People
Company-Wise bifurcation
on the basis of usage of
product
Middle Class
&
High Income Family
8. Will you be using a probability sampling technique or non-
probability sampling technique and why?
Meaning of Probability Sampling
Ininsights,likelihoodexaminingalludestothe inspectingstrategy inwhicheveryoneof the individuals
fromthe populace hasa pre-determinedandanequivalentopportunitytobe apiece of the example.
This system depends on the randomization guideline, wherein the strategy is so composed, which
ensuresthateverylast individual of the populace hasanequivalentchoice chance.Thisdecreasesthe
likelihood of predisposition.
Factual deductions can be made by the scientistsutilizing this procedure,i.e. the outcome acquired
can be summed up from the overviewed test to the objective populace
Meaning of Non-Probability Sampling
At the pointwheninan examiningstrategy,everyone of the people of the universe are notgivenan
equivalent chance of turning into a piece of the example, the technique is said to be Non-likelihood
testing. Under this method all things considered, there is no likelihoodconnected to the unit of the
populace and the choice depends on the emotional judgment of the scientist. In this way, the ends
drawn by the sampler can't be gathered from the example to the entire populace.
Probability Sampling: This type of sampling is profitable to the analyst for the AC business in the
following ways:
i) Here the technique wouldallow pickingupthe personwhoisinthe equallylikelychange
whether he is from the organized sector or from the north side, east side or from
anywhere.Thiswouldmake the possibilityof pickingupthe personsame foreach of the
individual.Thisscheme couldbe appliedalsoif the surveyis done on the sampleswhich
are using the AC this means that the selection of the person using the Voltas AC or the
HaierAC is same in the probabilitysamplingtechnique.Thiswouldreduce the possibility
of getting less biased for some variables of the frame used like the one which directly
effects on the mindset of the population for the given sample.
ii) Thisalso allowsusto make the predictionsaboutthe usage of the ACs easiersothat this
would allow increasing the profit of the company, thus from any environment, it is
whether be the north or the south.
iii) This would have a higher extent that analyst sample taken gives a more accurate result
for the population when the survey is done. This statement means that the inference
about the populationiseasierinthe one whichwouldhelpthe new startingcompanyto
generate its profit to the maximum.
The technique wouldallowthe randomizationof the personselectedforthe survey.Thismeansthat
not any general pattern would be formed for the selection criteria of the person selected by the
analyst for the marking of his answers to get the appropriate results.
9. What will be your plans tocontrol sampling andnon-sampling errors toobtain
an accurate result?
Sampling error occurred in the case are as follows:
i) The marketsurveys done by the companiesof AC that how much is usedby the home &
by the industries could be wrong sometimes.
ii) The mainerrorcouldbe inthe marketsegmentationthathasbeendone inthe case study
for the AC could be old or the data used could be secondary.
iii) The recordingof the variouscompaniesprofitslike of the Voltasand Blue Star is done in
the years which are a decade old thus leading to the data collection which is of no use.
iv) The preference about the market segmentation could not be only based on only these
variables given, some other variables like the electricity consumption by an AC or the
durability of the product can also come in handy as variables.
v) The error couldalsoarise inthe type of statistical analysisusedinsayingthatonly75%of
organized sector used windows AC which means there are many statistical tools also,
these toolsallowus to predict aboutthe future,but as an analystI wouldbe sure about
the statistical technique that I am using for the same purpose.
Some of the methods to control the non-sampling errors are:
i) The data collectedcouldbe fromthewrongenvironmentwhichmeanspeople whoare of
the average earnings and could not afford to use ACs at home.
ii) The questionnaire prepared could not be appropriate as the person asking the question
would not have selected the range scale for the measurement of the questions which
provide us with the forced options.
iii) The data recorderwouldbe notworkingwithhisfullefficiencyashe hastogoto boththe
organized and unorganized sector which means traveling is involved at a huge amount
which would have made the person to be more of a lazy kind and have made up of his
own answers.
Some of the more errors which could be found are:
i) ResearcherError:The errorcouldbe due tothe samplingmethodwhichisdonefortaking
out the sample. The error would also include the measurement error or the population
definition error. One of the major non-sampling errors which are included in the system
would be the respondent error from which the questionnaire has been asked.
ii) RespondentError:Thiserror will dependonsomany factorslike the unwillingnessof the
personto answerthe question.The respondentwhichisbeenasked the question should
10. be highlyinterestedingivinghisviews,tocome upwiththe solutionthe dataasked maybe
not monotonous making it boring for the respondent.
iii) Interviewer Error: The error from the interviewer side can be minimized by the making
the personavailablewithallthe informationonwhichthe ACsare workingorisbeensold
inthe market.So that duringthe askingof questionshe doesn’thave aclue whatisbeing
asked.
iv) Non-Response Error: this error could be removed by making use of the criteria where
everyrespondentinthe sample isimportant,thismeansthattakingcare of the situation
with the help of the schemes like giving themadvantagesaccordinglyso that they want
to respond to the questions asked.