Fundamentals of Data Analysis

Marketing ResearchAaker, Kumar, Leone and Day Twelfth EditionInstructor’s Presentation Slides1Chapter Sixteen2Fundamentals of Data AnalysisMarketing Research 12th Edition Data AnalysisA set of methods and techniques used to obtain information and insights from dataHelps avoid erroneous judgments and conclusionsCan constructively influence the research objectives and the research design3Major Data Preparation techniques: Data editing Coding Statistically adjusting the dataMarketing Research 12th Edition Data EditingIdentifies omissions, ambiguities, and errors in responsesConducted in the field by interviewer and field supervisor and by the analyst prior to data analysisProblems identified with data editing:Interviewer ErrorOmissionsAmbiguityInconsistenciesLack of CooperationIneligible Respondent4Marketing Research 12th Edition CodingCoding closed-ended questions involves specifying how the responses are to be enteredOpen-ended questions are difficult to code Lengthy list of possible responses is generated5Marketing Research 12th Edition Statistically Adjusting the DataWeightingEach response is assigned a number according to a pre-specified ruleMakes sample data more representative of target population on specific characteristicsModifies number of cases in the sample that possess certain characteristicsAdjusts the sample so that greater importance is attached to respondents with certain characteristics6Marketing Research 12th Edition Statistically Adjusting the Data (Contd.) Variable RespecificationExisting data are modified to create new variablesLarge number of variables are collapsed into fewer variables Creates variables that are consistent with study objectivesDummy variables are used Binary, dichotomous, instrumental, quantitative variables)Use (d-1) dummy variables to specify (d) levels of qualitative variable7Marketing Research 12th Edition Statistically Adjusting the Data (Contd.)Scale TransformationScale values are manipulated to ensure comparability with other scalesStandardization allows the researcher to compare variables that have been measured using different types of scalesVariables are forced to have a mean of zero and a standard deviation of oneCan be done only on interval or ratio-scaled dataStandardized score, 8Marketing Research 12th Edition Simple TabulationConsists of counting the number of cases that fall into various categories9Uses: Determine empirical distribution (frequency distribution) of the variable in questionCalculate summary statistics, particularly the mean or percentagesAid in "data cleaning" aspectsMarketing Research 12th Edition Frequency DistributionReports the number of responses that each question receivedOrganizes data into classes or groups of valuesShows number of observations that fall into each classCan be illustrated simply as a number or as a percentage or histogramResponse categories may be combined for many questionsShould result in categories with worthwhile number of respondents10Marketing Research 12th Edition Frequency Distribution11Marketing Research 12th Edition Descriptive StatisticsStatistics normally associated with a frequency distribution to help summarize information in the frequency tableIncludes:Measures of central tendency mean, median and modeMeasures of dispersion (range, standard deviation, and coefficient of variation)Measures of shape (skewness and kurtosis)12Marketing Research 12th Edition Cross TabulationsStatistical analysis technique to study the relationships among and between variablesSample is divided to learn how the dependent variable varies from subgroup to subgroupFrequency distribution for each subgroup is compared to the frequency distribution for the total sampleThe two variables that are analyzed must be nominally scaled13Marketing Research 12th Edition Factors Influencing the Choice of Statistical TechniqueTypes of DataClassification of data involves nominal, ordinal, interval and ratio scales of measurementNominal scaling is restricted in that mode is the only meaningful measure of central tendencyBoth median and mode can be used for ordinal scaleNon-parametric tests can only be run on ordinal dataMean, median and mode can all be used to measure central tendency for interval and ratio scaled data14Marketing Research 12th Edition Factors Influencing the Choice of Statistical Technique (Contd.)Research DesignDepends on:Whether dependent or independent samples are used Number of observations per objectNumber of groups being analyzedNumber of variablesControl exercised over variable of interest15Marketing Research 12th Edition Factors Influencing the Choice of Statistical Technique (Contd.)Assumptions Underlying the Test Statistic Two-sample t-test :The samples are independent.The characteristics of interest in each population have normal distribution.The two populations have equal variances.16Marketing Research 12th Edition Overview of Statistical TechniquesUnivariate TechniquesAppropriate when there is a single measurement of each of the 'n' sample objects or there are several measurements of each of the `n' observations but each variable is analyzed in isolationNonmetric data - measured on nominal or ordinal scaleMetric data - measured on interval or ratio scaleDetermine whether single or multiple samples are involvedFor multiple samples, choice of statistical test depends on whether the samples are independent or dependent17Marketing Research 12th Edition Classification of Univariate Statistical Techniques18Marketing Research 12th Edition Overview of Statistical Techniques (Contd.)Multivariate TechniquesA collection of procedures for analyzing association between two or more sets of measurements that have been made on each object in one or more samples of objects19Uses:To group variables or people or objectsTo improve the ability to predict variables (such as usage)To understand relationships between variables (such as advertising and sales)Marketing Research 12th Edition Classification of Multivariate Statistical Techniques20Marketing Research 12th Edition Classification of Multivariate Techniques (Contd.)Dependence TechniquesOne or more variables can be identified as dependent variables and the remaining as independent variablesChoice of dependence technique depends on the number of dependent variables involved in analysisInterdependence TechniquesWhole set of interdependent relationships is examinedFurther classified as having focus on variable or objects21Marketing Research 12th Edition 22End of Chapter SixteenMarketing Research 12th Edition