Sampling Fundamentals

Marketing ResearchAaker, Kumar, Leone and Day Twelfth EditionInstructor’s Presentation Slides1Chapter Fourteen2Sampling FundamentalsMarketing Research 12th Edition 2Sampling FundamentalsWhen is census appropriate?Population size is quite smallInformation is needed from every individual in the populationCost of making an incorrect decision is high Sampling errors are high3Marketing Research 12th Edition 3Sampling Fundamentals (Contd.)When is sample appropriate? Population size is largeBoth cost and time associated with obtaining information from the population is highQuick decision is neededTo increase response quality since more time can be spent on each interviewPopulation being dealt with is homogeneousIf census is impossible4Marketing Research 12th Edition 4Error in SamplingTotal Error Difference between the true value and the observed value of a variableSampling Error Error is due to samplingNon-sampling Error Error is observed in both census and sample5Marketing Research 12th Edition 5Error in Sampling (contd.)Sources of non-sampling errorMeasurement ErrorData Recording ErrorData Analysis ErrorNon-response Error6Marketing Research 12th Edition 6Sampling ProcessDetermining Target PopulationWell thought out research objectives Consider all alternativesKnow your marketConsider the appropriate sampling unitSpecify clearly what is excludedShould be reproducibleConsider convenience7Marketing Research 12th Edition 7Sampling Process (Contd.)Determining Sampling FrameList of population members used to obtain a sampleIssues:Obtaining appropriate listsDealing with population sampling frame differencesSuperset problemIntersection problemSelecting a Sampling ProcedureChoose between Bayesian and Traditional sampling procedureDecide whether to sample with or without replacement8Marketing Research 12th Edition 89The Sampling ProcessMarketing Research 12th Edition 9Sampling Techniques10Marketing Research 12th Edition Probability SamplingAll population members have a known probability of being in the sampleSimple Random SamplingEach population member and each possible sample has equal probability of being selectedStratified SamplingThe chosen sample is forced to contain units from each of the segments or strata of the population10Types of Stratified SamplingProportionate Stratified Sampling Number of objects/sampling units chosen from each group is proportional to number in populationCan be classified as directly proportional or indirectly proportional stratified samplingDisproportionate Stratified SamplingSample size in each group is not proportional to the respective group sizesUsed when multiple groups are compared and respective group sizes are small11Marketing Research 12th Edition 1112Directly Proportional Stratified SamplingMarketing Research 12th Edition 1213Assume that among the 600 consumers in the population, 200 are heavy drinkers and 400 are light drinkers. If a research values the opinion of the heavy drinkers more than that of the light drinkers, more people will have to be sampled from the heavy drinkers group. If a sample size of 60 is desired, a 10 percent inversely proportional stratified sampling is employed.The selection probabilities are computed as follows:Inversely Proportional Stratified SamplingDenominatorHeavy Drinkers proportional and sample sizeLight drinkers proportional and sample size600/200 + 600/400 = 3 + 1.5 = 4.53/ 4.5 = 0.667; 0.667 * 60 = 401.5 / 4.5 = 0.333; 0.333 * 60 = 20Marketing Research 12th Edition 13Cluster SamplingInvolves dividing population into subgroups Random sample of subgroups/clusters is selected and all members of subgroups are interviewed Very cost effectiveUseful when subgroups can be identified that are representative of entire population14Marketing Research 12th Edition 14Comparison of Stratified & Cluster Sampling Processes15Homogeneity within groupHeterogeneity between groupsAll groups are includedSampling efficiency improved by increasing accuracy at a faster rate than costHomogeneity between groupsHeterogeneity within groupsRandom selection of groupsSampling efficiency improved by decreasing cost at a faster rate than accuracy.Stratified samplingCluster samplingMarketing Research 12th Edition Systematic SamplingInvolves systematically spreading the sample through the list of population membersCommonly used in telephone surveysSampling efficiency depends on ordering of the list in the sampling frame16Marketing Research 12th Edition 16Non Probability SamplingCosts and trouble of developing sampling frame are eliminatedResults can contain hidden biases and uncertainties17Used in: The exploratory stages of a research project Pre-testing a questionnaire Dealing with a homogeneous population When a researcher lacks statistical knowledge When operational ease is requiredMarketing Research 12th Edition 17Types of Non Probability SamplingJudgmental"Expert" uses judgment to identify representative samplesSnowballForm of judgmental samplingAppropriate when reaching small, specialized populationsEach respondent, after being interviewed, is asked to identify one or more others in the fieldConvenience Used to obtain information quickly and inexpensively Quota Minimum number from each specified subgroup in the populationOften based on demographic data18Marketing Research 12th Edition 18Quota Sampling - Example19Marketing Research 12th Edition 19Non-Response ProblemsRespondents may:Refuse to respondLack the ability to respondBe inaccessibleSample size has to be large enough to allow for non responseThose who respond may differ from non respondents in a meaningful way, creating biasesSeriousness of non-response bias depends on extent of non response20Marketing Research 12th Edition 20Solutions to Non-response ProblemImprove research design to reduce the number of non-responsesRepeat the contact one or more times (call back) to try to reduce non-responsesAttempt to estimate the non-response bias21Marketing Research 12th Edition 21Shopping Center Sampling20% of all questionnaires completed or interviews granted are store-intercept interviewsBias is introduced by methods used to select22Source of BiasSelection of shopping centerPoint of shopping center from which respondents are drawnTime of dayMore frequent shoppers will be more likely to be selectedMarketing Research 12th Edition 22Shopping Center Sampling (Contd.)Solutions to BiasShopping Center BiasUse several shopping centers in different neighborhoodsUse several diverse citiesSample Locations Within a CenterStratify by entrance locationTake separate sample from each entranceTo obtain overall average, strata averages should be combined by weighing them to reflect traffic that is associated with each entrance23Marketing Research 12th Edition 23Solutions to Bias (contd.)Time SamplingStratify by time segmentsInterview during each segmentFinal counts should be weighed according to traffic countsSampling People versus Shopping Visits – Options:Ask respondents how many times they visited the shopping center during a specified time period, such as the last four weeks and weight results according to frequencyUse quotas, which serve to reduce the biases to levels that may be acceptableControl for sex, age, employment status etc.The number sampled should be proportional to the number of the quota in the population24Marketing Research 12th Edition 24Different Levels of Sampling Frames25Marketing Research 12th Edition 2526End of Chapter FourteenMarketing Research 12th Edition