Upload
ashishgiri
View
2.611
Download
4
Embed Size (px)
Citation preview
Research in BusinessResearch in Business
Why Study Research?Why Study Research?
– Research provides you with the knowledge and skills needed for the fast-paced decision-making environment
– Search for knowledge through objective and systematic method of finding solution to the problem is research
– Systematic approach concerns generalization and formulation
Why managers need Better Why managers need Better InformationInformation
Global and domestic competitions is more vigorous– Organizations are increasingly practicing data mining
and data warehousing- data warehousing market consists of tools, technologies, and methodologies that allow for the construction, usage, management, and maintenance of the hardware and software used for a data warehouse, as well as the actual data itself. To discover patterns and trends in the data
The Value of Acquiring SkillsThe Value of Acquiring Skills
To gather more information before selecting a course of action
To do a high-level research studyTo understand research designTo evaluate and resolve a current
management dilemmaTo establish a career as a research specialist
Different Styles of ResearchDifferent Styles of Research
Applied Research( action)1. Improved agricultural production 2. Treat or cure a specific disease3. Improve energy efficiency of homes ,offices or modes of
transporation4. Global overpopulation
Pure Research(generations and formulations) Quantitative and qualitative research Descriptive and analytical research Conceptual and emperical reseach
What is Good Research?What is Good Research?
Following the standards of the scientific method:– Purpose clearly defined-problem involved and decision
made– Research process detailed-permit others– Research design thoroughly planned-yeild accurate
results– High ethical standards applied –standards and norms– Limitations frankly revealed
What is Good Research?What is Good Research?
Following the standards of the scientific method (cont.)– Adequate analysis for decision-maker’s needs– Findings presented unambiguously –clear
logical,executive summary – Conclusions justified-conclusions matched with
detailed finding– Researcher’s experience reflected
Role of buisness research in Role of buisness research in decision makingdecision making
Problem identification
Problem prioritization
Problem resolution
Implementing course of action
Factors affecting business Factors affecting business research research
Time constraint
Availability of resources
Nature of information sought
Benefit versus cost
Scientific ThinkingScientific Thinking
The Essential Tenets of The Essential Tenets of ScienceScience
Direct observation of phenomenaClearly defined variables, methods, and
proceduresEmpirically testable hypothesesAbility to rule out rival hypotheses-more
facts or greater variety or scopeStatistical justification of conclusionsSelf-correcting process
Ways to CommunicateWays to Communicate
Exposition– Descriptive statements that merely state and so
not give reason
Argument – Allows us to explain, interpret, defend,
challenge, and explore meaning
Important Arguments in Important Arguments in ResearchResearch
Deduction is a form of inference that purports to conclusive
Induction draws conclusions form one or more particular facts
The Building Blocks of TheoryThe Building Blocks of Theory
ConceptsConstructsDefinitionsVariablesPropositions and HypothesesTheoriesModels
Understanding ConceptsUnderstanding Concepts
A concept is a bundle of meanings or characteristics associated with certain events
Concepts have been developed over time through shared usage
The success of research hinges on:– How clearly we conceptualize and– How well others understand the concepts we
use
What is a Construct?What is a Construct?
An image or idea specifically invented for a given research purpose
Composed of conceptsTherotical level of concepts
Types of VariablesTypes of Variables
IndependentDependentModeratingExtraneous-infinite numberInterveningDichotomous Symbol to assign numerals and value
The Role of the HypothesesThe Role of the Hypotheses
Guides the direction of studyProvides a framework for organizing the
conclusions that resultIdentifies the fact that are relevant and those
that are notSuggest which form of research design is
likely to be most appropriate
What is a Good Hypotheses?What is a Good Hypotheses?
A good hypotheses should fulfill 3 conditions:– Must be adequate for its purpose– Must be testable– Must be better than its rivals
The Value of a Theory The Value of a Theory
Narrows the range of facts we need to studySummarizes what is know about an object
of studyUsed to predict further facts that should be
found Systematically interrelated
concepts,defination that are advanced to explain and predict phenomenon
The Research ProcessThe Research Process
Steps in Research ProcessSteps in Research Process
Identifing and defining problemPlanning the research designSelecting the research method Selecting a sampling procedureData collection Evaluating the dataPresenting the research project
The Management- Research The Management- Research Question HierarchyQuestion Hierarchy
Measurement Questions
Investigative Questions
Research Questions Management
Questions Management Dilemma
Level 5
Level 4
Level 3 Level 2
Level 1
Working with Hierarchy Working with Hierarchy
Management Dilemma– The symptoms of an actual problem– Not difficult to identify a dilemma, however
choosing one to focus on may be difficult
Working with the HierarchyWorking with the Hierarchy
Management Question Categories– Choice of purpose or objective– Generation and Evaluation of solutions– Troubleshooting or control situation
Working with the HierarchyWorking with the Hierarchy
Fine tune the research question– Examine concepts and constructs– Break research question into specific second
and third level questions– Determine what evidence answers the various
questions and hypothesis– Set the scope of your study
Working with the HierarchyWorking with the Hierarchy
Investigative Questions– Questions the researcher must answer to
satisfactorily arrive at a conclusion about the research question
Working with the HierarchyWorking with the Hierarchy
Measurement Questions– The questions we actually ask or extract from
respondents
Other Process in the Other Process in the HierarchyHierarchy
Exploration– Recent developments– Predictions by informed figures about the
prospects of technology– Identification of those involved in the area– Accounts of successful ventures and failures by
others in the field
Research Process ProblemsResearch Process Problems
The Favored Technique SyndromeCompany Database Strip-MiningUnresearchable QuestionIll-Defined Management ProblemsPolitically Motivated Research
Designing the StudyDesigning the Study
– Select a research design from the large variety of methods, techniques, procedures, protocols and sampling plans
Resource Allocation and Resource Allocation and BudgetBudget
Guides to plan a budget– Project planning– Data gathering– Analysis, interpretation, and reporting
Types of budgeting– Rule of thumb– Departmental or functional area– Task
Evaluation MethodsEvaluation Methods
Ex Post Facto EvaluationPrior EvaluationOption AnalysisDecision Theory
Contents of a Research Contents of a Research ProposalProposal
A statement of the research questionA brief description of research methodologyData collectionData preparationData analysis and interpretationResearch reporting
Data CollectionData Collection
Characterized by – Abstractness– Verifiability– Elusiveness– Closeness to the phenomenon
Secondary DataPrimary Data
Final Steps in ResearchFinal Steps in Research
Data analysis – Reporting the results
Executive SummaryOverview of the research Implementation strategies for the
recommendationsTechnical appendix
The Research ProposalThe Research Proposal
Purpose of the Research Purpose of the Research Proposal Proposal
To present the question to be researched and its importance
To discuss the research efforts of others who have worked on related questions
To suggest the data necessary for solving the question
The Research SponsorThe Research Sponsor
All research has a sponsor in one form or another:
In a corporate setting, management sponsors research
In an academic environment, the student is responsible to the class instructor
What are the Benefits of the What are the Benefits of the Proposal to a Researcher?Proposal to a Researcher?
Allows the researcher to plan and review the project’s steps
Serves as a guide throughout the investigation
Forces time and budget estimates
Types of Research ProposalsTypes of Research Proposals
Internal
External
Proposal ComplexityProposal Complexity
3 Levels of Complexity– The Exploratory study is used for the most
simple proposals– The Small-scale study is more complex and
common in business– The Large-scale professional study is the most
complex, costing millions of dollars
How to Structure the How to Structure the Research Proposal?Research Proposal?
Create proposal modules
Put together various modules to tailor your proposal to the intended audience
Modules in a Research Modules in a Research ProposalProposal
Executive Summary Problem statement Research objectives Literature Reviews Importance of the Study Research Design Data Analysis Nature and Form of
Results
Qualifications of Researcher
Budget Schedule Facilities and Special
Resources Project Management Bibliography Appendixes
What to include in the What to include in the Appendixes?Appendixes?
A glossary of concepts, constructs, and definitions
Samples of the measurement instrumentOther materials that reinforce the body of
the proposal
Evaluating the Research Evaluating the Research ProposalProposal
Proposal must be neatly writtenMajor topic should be easily found and
logically organizedProposal must meet specific guidelines set
by the sponsorTechnical writing style must be clearly
understood and explained
Ethics in Business Research Ethics in Business Research
What are Research Ethics?What are Research Ethics?
Ethics are norms or standards of behavior that guide moral choices about our behavior and our relationships with others
The goal is to ensure that no one is harmed or suffers adverse consequences from research activities
Ethical Treatment of Ethical Treatment of Respondents and SubjectsRespondents and Subjects
Begin data collection by explaining to the respondent the benefits expected from the research
Explain to the respondent that their rights and well-being will be adequately protected, and say how this will be done
Be certain that interviews obtain the informed consent of the respondent
DeceptionDeception
The respondent is told only part of the truth when the truth is fully compromised– To prevent biasing the respondents before the
survey or experiment– To protect the confidentiality of a third party
Issues Related to Protecting Issues Related to Protecting RespondentsRespondents
Informed consentDebriefingConfidentialityRight to Privacy
Ethical Issues Related to the Ethical Issues Related to the ClientClient
Sponsor nondisclosurePurpose nondisclosureFindings nondisclosure
– Right to quality research
Ethical Issues Related to Ethical Issues Related to Researchers and Team Researchers and Team
MembersMembers
Safety
Ethical behavior of assistants
Protection of anonymity
Design StrategiesDesign Strategies
What is Research Design?What is Research Design?
A plan for selecting the sources and types of information used to answer research questions
A frame work for specifying the relationships among the study variables
A blueprint that outlines each procedure from the hypothesis to the analysis
Classifications of DesignsClassifications of Designs
Exploratory study is usually to develop hypotheses or questions for further research
Formal study is to test the hypotheses or answer the research question posed
Methods of Data CollectionMethods of Data Collection
Monitoring, which includes observational studies
Interrogation/ Communication mode
The Power of a ResearcherThe Power of a Researcher
In an experiment, the researcher attempts to control and/or manipulate the variables in the study
In an ex post facto design, the researcher has no control over the variables, they can only report what has happened
What type of Study to use?What type of Study to use?
Descriptive is how one variable produces changes in another
Causal tries to explain relationships among variables
The Time DimensionThe Time Dimension
Cross-sectional studies are carried out once and the represent a snapshot of one point and time
Longitudinal studies are repeated over an extended period
The Topical ScopeThe Topical Scope
Statistical studies attempt to capture a population’s characteristic’s by making inferences form a sample’s characteristics
Case studies place more emphasis on a full contextual analysis of fewer events or conditions and their interrelations
The Research EnvironmentThe Research Environment
Field Conditions
Laboratory Conditions
Simulations
A Subjects’ PerceptionsA Subjects’ Perceptions
Usefulness of a design may be reduced when people in the study perceive that research is being conducted
Subject’s perceptions influence the outcomes of the research
Why do Exploratory Studies?Why do Exploratory Studies?
Exploration is particularly useful when researchers lack a clear idea of the problems
Data Collection TechniquesData Collection Techniques
Qualitative TechniquesSecondary DataFocus GroupsTwo-stage Design
The Concept of a Causal The Concept of a Causal StudyStudy
The essential element of causation is that A “produces” B or A “forces” B to occur
Relationships that Occur with Relationships that Occur with a Causal Studya Causal Study
Symmetrical
Reciprocal
Asymmetrical
Types of Asymmetrical Types of Asymmetrical RelationshipsRelationships
Stimulus-ResponseProperty-DispositionDisposition-BehaviorProperty-Behavior
Achieving the Ideal Achieving the Ideal Experimental DesignExperimental Design
Random Assignment
Matching
– Manipulation and control of variables
MeasurementMeasurement
MeasurementMeasurement
Selecting observable empirical events
Using numbers or symbols to represent aspects of the events
Applying a mapping rule to connect the observation to the symbol
What is Measured?What is Measured?
Objects-things of ordinary experience and that are not that concrete
Properties-characteristics of objects
Characteristics of DataCharacteristics of Data
Order
Interval between numbers
Origin of number series
Data Types Data Types
Order Interval OriginNominal none - none - none
Ordinal yes - unequal - none
Interval yes - equal or unequal -none
Ratio yes - equal - zero
Sources of Measurement Sources of Measurement DifferencesDifferences
RespondentSituational factorsMeasurer or researcherInstrument
ValidityValidity
Content Validity
Criterion-Related Validity– Concurrent– Predictive
Construct Validity
ReliabilityReliability
Stability– Test-retest
Equivalence– Parallel forms
Internal Consistency– Split-half– KR20– Cronbach’s alpha
PracticalityPracticality
Economy
Convenience
Interpretability
Chapter 8Chapter 8Scaling Design Scaling Design
What is Scaling?What is Scaling?
Assigning numbers to indicants of the properties of objects
Types of Response ScalesTypes of Response Scales
Rating Scales
Ranking Scales
Types of Rating ScalesTypes of Rating Scales
Simple categoryMultiple choice,
multiple responseLikert scaleSemantic
differential
Numerical Multiple fixed
ratingFixed sumStapelGraphic rating
Rating Scales Problems to Rating Scales Problems to AvoidAvoid
LeniencyNegative LeniencyCentral Tendency Halo Effect
Types of Ranking ScalesTypes of Ranking Scales
Paired-comparison
Forced Ranking
Comparative
Dimensions of a ScaleDimensions of a Scale
Unidimensional
Multidimensional
Scale Design TechniquesScale Design Techniques
Arbitrary
Consensus
Item Analysis– Cumulative
Factor
Sampling DesignSampling Design
Selection of ElementsSelection of Elements
Sampling
Population Element
Population
Census
What is a Good Sample?What is a Good Sample?
Accurate
Precision of estimate
Types of Sampling DesignsTypes of Sampling Designs
Probability
Nonprobability
Steps in Sampling DesignSteps in Sampling Design
What is the relevant population?What are the parameters of interest?What is the sampling frame?What is the type of sample?What size sample is needed?How much will it cost?
Concepts to help understand Concepts to help understand Probability SamplingProbability Sampling
Standard error of the mean
Confidence interval
Central limit theorem
Probability Sampling DesignsProbability Sampling Designs
Simple RandomSystemicStratified
– Proportionate
ClusterDouble
Designing Cluster SamplesDesigning Cluster Samples
How homogeneous are the clusters?Shall we seek equal or unequal clusters?How large a cluster shall we take?Shall we use a single-stage or multistage
cluster?How large a sample is needed?
Nonprobability SamplingNonprobability Sampling
Reasons to use Nonprobability Sampling instead of Probability Sampling– The nonprobability procedure satisfactorily meets
the sampling objectives– Lower cost– Limited Time– Not as much human error as selecting a completely
random sample– Total list population not available
Nonprobability Sampling Nonprobability Sampling DesignsDesigns
Convenience SamplingPurposive Sampling
– Judgement Sampling– Quota Sampling
Snowball Sampling
Secondary Data SourcesSecondary Data Sources
Information is Classifies by Information is Classifies by Two Sources:Two Sources:
Primary Data
Secondary Data
Uses of Secondary DataUses of Secondary Data
Provides specific reference or citation on some point
Helps decide what further research needs to be done
Justifies bypassing the costs and benefits of doing primary research
May be used as the sole basis for a research study
Classifying Secondary DataClassifying Secondary Data
By SourceBy CategoryBy MediumBy Database format
Classifying Secondary Data Classifying Secondary Data by Sourceby Source
Internal
External
Classifying Secondary Data Classifying Secondary Data by Categoryby Category
DatabasePeriodicalsGovernment DocumentsSpecial Collections
Classifying Secondary DataClassifying Secondary Databy Medium by Medium
Hard copy
– Local-area on-line
– Internet
The Library’s Role in The Library’s Role in ResearchResearch
Resources may be acquired through interlibrary loans (ILL)
Certain Databases are available on a local-area network (LAN)
Access to the internet an commercial CD/ DVD-ROM
Strategy for Searching for Strategy for Searching for Secondary DataSecondary Data
Select and analyze a topic Explore the topic and state a hypothesisGet an overview and retrospective
informationGet more current and specific informationGet more in-depth informationEvaluate and close the library research
Using Search Engines and Using Search Engines and IndexesIndexes
The search engine consists of two elements:– Robot/Crawler– Indexer
How to Keep Track of How to Keep Track of Research?Research?
Be selective in what you record
Decide how to record what you will extract from the published material
Develop an orderly recording system
Survey Methods: Survey Methods: Communicating with Communicating with
RespondentsRespondents
Communication Approach Communication Approach Impacts the Research Impacts the Research
ProcessProcessCreation and selection of measurement questionsSampling issues, drive contact and callback
proceduresInstrument design, which incorporates attempts to
reduce error and create respondent-screening procedures
Data collection procedures and possible interviewer training
Personal InterviewPersonal Interview
Requirements for success– Availability of the needed information from the
respondent– An understanding by the respondent of his or
her role– Adequate motivation by the respondent to
cooperate
Personal InterviewPersonal Interview
To Increase Respondent’s Receptiveness they must– believe the experience will be pleasant and
satisfying– think answering the survey is an important and
worthwhile use of their time– have any mental reservations satisfied
The InterviewThe Interview
Introduction– Establish a good relationship
Gather the data – Probing
Record the Interview
Probing StylesProbing Styles
A brief assertion of understanding and interest
An expectant pauseRepeating the questionRepeating the respondent’s replyA neutral question or comment Question clarification
Interview ProblemsInterview Problems
Non-response error
Response error
Interviewer error– Cost
Telephone InterviewTelephone Interview
Types– Computer-assisted telephone interviewing– Computer-administered telephone survey
Problems– Non-contact rate– Refusal rate
Self-Administered Self-Administered
Types– Intercept study– Mail survey
Disadvantages– Large non-response error– Cannot obtain detailed or large amounts of
information
Concurrent Techniques to Concurrent Techniques to Improve Mail ResponseImprove Mail Response
Reduce Length Survey Sponsorship Return Envelopes Postage Personalization
Anonymity Size, color, and
reproduction Money Incentives Deadline Dates Cover Letters
Outsourcing Survey ServicesOutsourcing Survey Services
Research Firms Provide– Centralized-location interviewing– Focus group facilities– Trained staff with experience– Data-processing and statistical analysis
capabilities– Access to point of scale data
Panels
Instruments For Respondent Instruments For Respondent CommunicationCommunication
3 Phases of the Instrument 3 Phases of the Instrument Design ProcessDesign Process
Developing the instrument design process
Constructing and refining the measurement questions
Drafting and refining the instrument
Developing the Instrument Developing the Instrument Design StrategyDesign Strategy
You must go through four question levels:– The management question– Research question– Investigative questions– Measurement questions
Strategic Concerns of Strategic Concerns of Instrument DesignInstrument Design
What type of data is needed to answer the management question
What communication approach will be used – Should the question be structured, unstructured,
or some combination– Should the question be disguised or
undisguised
Ways to Interact with the Ways to Interact with the RespondentRespondent
Personal Interview TelephoneMail Computer
What are the Three Types of What are the Three Types of Measurement Questions?Measurement Questions?
TargetClassificationAdministrative
4 Questions for Selecting 4 Questions for Selecting Appropriate Question ContentAppropriate Question Content
Should this question be asked?Is the question of proper scope and
coverage?Can the respondent adequately answer this
question, as asked?Will the respondent willingly answer this
question, as asked?
How to test a Respondent’s How to test a Respondent’s KnowledgeKnowledge
Filter Questions
Screen Questions
Question Wording CriteriaQuestion Wording Criteria
Is the question stated in terms of a shared vocabulary?
Does the question contain vocabulary with a single meaning?
Does the question contain unsupported assumptions?
Is the question correctly personalized?Are adequate alternatives presented within the
question?
What Dictates Your Response What Dictates Your Response Strategy?Strategy?
Characteristics of respondentsNature of the topic being studiedType of data neededYour analysis plan
Types of Response QuestionsTypes of Response Questions
Free-responseDichotomousMultiple choiceRatingRanking
Guidelines to Refining the Guidelines to Refining the InstrumentInstrument
Awaken the respondent’s interests
Use buffer questions as a guide to request sensitive information
Use the funnel approach to move to more specific questions
Final Step Toward Improving Final Step Toward Improving Survey ResultsSurvey Results
Pre-testing is an established practice for discovering errors and useful for training the research team
Observational StudiesObservational Studies
ObservationObservation
Non-behavioral observation– Record analysis– Physical condition analysis– Physical process analysis
Behavioral observation– nonverbal analysis– Linguistic analysis– Extra-linguistic analysis– Spatial analysis
Advantages of the Advantages of the Observational MethodObservational Method
Only method available to collect certain types of data
Collect the original data at the time it occursSecure information that participants would
ignore because it’s so common it is not seen as relevant
Advantages of the Advantages of the Observational Method (cont..)Observational Method (cont..)
Capture the whole event as it occurs in its natural environment
Subjects seem to accept an observational intrusion better than they respond to questioning
Limitations of the Limitations of the Observational Method Observational Method
Observer or recording equipment must be at the scene of the event when it takes place
Slow processExpensive processMost reliable results are restricted to
information that can be learned by overt action or surface indicators
Limitations of the Limitations of the Observational Method (cont..)Observational Method (cont..)
Research environment is more likely suited to subjective assessment and recording of data than to quantification of events
Limited as a way to learn about the pastCannot observe rationale for actions, only
actions themselves
Relationship between Relationship between observer and subjectobserver and subject
Direct or indirect observationObserver’s presence known or unknown to
the subjectObserver’s involvement level with the
respondent
ObservationObservation
DirectIndirectParticipantSimpleSystematic
Guidelines for selecting Guidelines for selecting observersobservers
Ability to concentrate in a setting full of distractions
Ability to remember details of an experience
Ability to be unobtrusive in the observational situation
Data collectionData collection
Who?What?
– Event Sampling– Time Sampling
When?How?
ExperimentationExperimentation
Types of variables in Types of variables in ExperimentsExperiments
Independent Variables
Dependent Variables
What are the Advantages of What are the Advantages of an Experiment?an Experiment?
Researcher’s ability to manipulate the independent variable
Contamination from extraneous variables can be controlled more efficiently
Convenience and costReplication
What are the Disadvantages?What are the Disadvantages?
Artificiality of the laboratoryGeneralization from non-probability
samplesLarger budgets needed Restricted to problems of the present or
immediate futureEthical limits to manipulation of people
How to Conduct an How to Conduct an Experiment?Experiment?
Select relevant variablesSpecify the treatment levelsControl the experimental environmentChoose the experimental design Select and assign the subjectsPilot-test, revise, and testAnalyze the data
Ways to Assign Subjects?Ways to Assign Subjects?
Random Assignment
Matching Assignment
Quota Matrix
Does a Measure Accomplish Does a Measure Accomplish What it Claims?What it Claims?
Internal validity
External validity
Variations in Experimental Variations in Experimental DesignsDesigns
Pre-experimental designs
True experimental designs
Field experiments
Types of Pre-experimental Types of Pre-experimental Designs?Designs?
One-shot case study
One-group pretest-posttest design
Static group comparison
Types of True Experimental Types of True Experimental DesignsDesigns
Pretest-posttest control group design
Posttest only control group design
Operational Extensions of Operational Extensions of True DesignsTrue Designs
Completely randomized designsRandomized block designLatin squareFactorial designCovariance analysis
What are Field Experiments:What are Field Experiments:Quasi or Semi?Quasi or Semi?
Non equivalent control group design
Separate sample pretest-posttest design
Group time series design
Data preparation and Data preparation and Preliminary AnalysisPreliminary Analysis
EditingEditing
Detects errors and omissions, corrects them when possible, and certifies that minimum data quality standards are achieved
Editing (cont..)Editing (cont..)
Guarantees that data are– accurate– consistent with other information– uniformly entered– complete– arranged to simplify coding and tabulation
CodingCoding
Rules that guide the establishment of category sets– Appropriate to the research problem and
purpose– Exhaustive– Mutually exclusive – Derived from one classification principal
Content Analysis Content Analysis
Follows a systematic process with the selection of a unitization scheme– Syntactical unit– Referential unit – Propositional unit– Thematic unit
Data Entry OptionsData Entry Options
Optical ScanningSpreadsheetsData warehouse
– Transformation and cleaning– End-user access tools
Data marts
Descriptive StatisticsDescriptive Statistics
Distribution Standard normal
distribution Central tendency
– Mean– Median– Mode
Variability
Variance Standard deviation Range Interquartile range Skewness Kurtosis
Techniques to Display and Techniques to Display and Examine DistributionsExamine Distributions
Frequency tableHistograms
– Display all intervals in a distribution, even without observed values
– Examine the shape of the distribution for Skewness, kurtosis, and the modal pattern
Stem and leaf display
Techniques (cont.)Techniques (cont.)
Box and whisker-plot– Rectangular plot tat encompasses 50% of the
data values– A center line marking the median and going go
through the width of the box– The edges of the box (hinges)– Whiskers that extend from the right and left
hinges to the largest and smallest values
Techniques (cont.)Techniques (cont.)
Transformation– To improve interpretation and compatibility
with other data sets– To enhance symmetry and stabilize spread– To improve linear relationships between and
among variables
Data Mining TechniquesData Mining Techniques
Data visualization– Dimensions– Measurements– Hierarchies
ClusteringNeural networksTree ModelsClassification
Data Mining Techniques Data Mining Techniques (cont.)(cont.)
Market-Basket AnalysisSequence Based AnalysisFuzzy LogicGenetic Algorithms Fractal-base Transformation
Data Mining ProcessData Mining Process
Sample ExploreModify ModelAssess
Hypothesis TestingHypothesis Testing
Two Approaches to Two Approaches to Hypothesis TestingHypothesis Testing
Classical Statistics
Bayesian Statistics
Types of HypothesesTypes of Hypotheses
Null
Alternative
The Logic of HypothesisThe Logic of HypothesisTestingTesting
Two tailed test
One tailed test
Decision Errors in TestingDecision Errors in Testing
Type I error
Type II error
Testing for Statistical Testing for Statistical Significance Significance
State the null hypothesisChoose the statistical testSelect the desired level of significance Compute the calculated difference valueObtain the critical valueMake the decision
What are Significant Tests?What are Significant Tests?
Parametric tests
Non-parametric tests
How to Test the Null How to Test the Null HypothesisHypothesis
Analysis of variance (ANOVA)
How to select a testHow to select a test
Does the test involve one sample, two samples, or k samples?
If two samples or k samples are involved, are the individual cases independent or related?
Is the measurement scale nominal, ordinal, interval, or ratio?
When to use the K Related When to use the K Related Sample TestsSample Tests
The grouping factor has more than two levels
Observations or subjects are matched or the same subject is measured more than once
The data are at least interval
Measures of AssociationMeasures of Association
Bivariate Correlation vs.. Non-Bivariate Correlation vs.. Non-parametric Measures of parametric Measures of
Association Association Parametric correlation requires two
continuous variables measured on an interval or ratio scale
The coefficient does not distinguish between independent and dependent variables
Bivariate Correlation Analysis Bivariate Correlation Analysis
Pearson correlation coefficient – r symbolized the coefficient’s estimate of linear
association based on sampling data– Correlation Coefficients reveal the magnitude and
direction of relationships – Coefficient’s sign (+ or -) signifies the direction of the
relationship
Assumptions of r– Linearity– Bivariate normal distribution
Bivariate Correlation AnalysisBivariate Correlation Analysis
Scatterplots– Provide a means for visual inspection of data– Both direction and shape of a relationship are
conveyed
Interpretation of CoefficientsInterpretation of Coefficients
Coefficient of determinationCorrelation matrix
– used to display coefficients for more than two variables
Correlation coefficient does not imply causation
Interpretation of CoefficientsInterpretation of Coefficients
Suggests alternate explanations for correlation results– X causes Y, or Y causes X, or X&Y are
activated by one or more other variables, or X&Y influence each other reciprocally
Practical SignificanceStatistical Significance Artifact correlations
Bivariate Linear RegressionBivariate Linear Regression
Used to make simple and multiple predictions
Regression coefficients– Slope– Intercept
Error termMethod of least squares
Interpreting Linear RegressionInterpreting Linear Regression
Residuals
Prediction and confidence bands
Interpreting Linear RegressionInterpreting Linear Regression
Goodness of fit– Zero slopes come from
Y completely unrelated to X and no systematic pattern is evident
Constant values of Y for every value of X data are related, but represented by a nonlinear function
– Tests t test F test Coefficient of Determination
Non-parametric Measures of Non-parametric Measures of AssociationAssociation
Measures for nominal data – When there is no relationship at all, coefficient
should be 0– When there is a complete dependency, the
coefficient should display unity or 1
Non-parametric Measures of Non-parametric Measures of AssociationAssociation
Chi-square based measure– Phi– Cramer’s V– Contingency coefficient of C
Proportional reduction in error (PRE)– Lambda– Tau
Characteristics of Ordinal Characteristics of Ordinal DataData
Concordant- subject ranks higher on one variable also ranks higher on the other variable
Discordant- subject ranks higher on one variable is ranked lower on the other variable
Measures for Ordinal DataMeasures for Ordinal Data
GammaSomer’s dSpearman’s rhoKendall’s tau bKendall’s tau c
– No assumption of bivariate normal distribution – Values range from +1.0 to -1.0
Multivariate Analysis:Multivariate Analysis:An Overview An Overview
Selecting a Multivariate Selecting a Multivariate TechniqueTechnique
Dependency
Interdependency
What are Dependency What are Dependency Techniques?Techniques?
Multiple regressionDiscriminant analysisMultivariate analysis if variance, or
MANOVALinear structural relationships, or LISREL Conjoint analysis
What are Interdependency What are Interdependency Techniques?Techniques?
Factor analysis
Cluster analysis
Multidimensional scaling (MDS)
Use Multiple Regression as a Use Multiple Regression as a Descriptive ToolDescriptive Tool
Predict values for a criterion variable by developing a self-weighting estimating equation
Control for confounding variables to better evaluate the contribution other variables
Test and explain causal theories
Uses for Discriminant AnalysisUses for Discriminant Analysis
Classify persons or objects into various groups
Analyze known groups to determine the relative influence of specific factors
Why Use MANOVA?Why Use MANOVA?
In business research, MANOVA can be used to test differences among samples of employees, customers, manufactured items, and production parts.
The Two Models of LISRELThe Two Models of LISREL
Measurement
Structural equation
Applications for Conjoint Applications for Conjoint AnalysisAnalysis
Market Research
Product development
What is Factor Analysis?What is Factor Analysis?
Computational techniques that reduce variables to a manageable number
Measurement statistics
Five Basic Steps to the Five Basic Steps to the Application of Cluster StudiesApplication of Cluster Studies
Selection of the sample to be clustered Definition of the variables on which to
measure the objects, events, or peoplecomputation of similarities among the entities
through correlation, Euclidean distances, and other techniques
Selection of mutually exclusive clustersCluster comparison and validation
What does Multidimensional What does Multidimensional Scaling Do?Scaling Do?
Creates a special description of a respondent’s perception about a product, service, or other object of interest
Written and Oral ReportsWritten and Oral Reports
Written Research ReportWritten Research Report
Short report– Tell the reader why you are writing– If in response, remind reader the exact point, answer
it, and follow with details – Write in expository style with brevity and directness– Write report today and leave it for tomorrow to
review before sending it– Attach detailed material as appendices when needed
Written Research ReportWritten Research Report
Long report– Technical report– Management report
Research Report ComponentsResearch Report Components Prefatory Items
– Letter of transmittal– Title page– Authorization letter– Executive summary– Table of contents
Introduction– Problem Statement– Research objectives– Background
Methodology – Sampling design– Research design– Data collection– Data analysis– Limitations
Conclusions– Summary and conclusions– Recommendations– Appendices– Bibliography
Written Report ConsiderationsWritten Report Considerations
Order – Sentence outline– Topic outline
Readability indicesPaceTone
Presentation of StatisticsPresentation of Statistics
Text paragraphSemi-tabular formTablesGraphics
GraphicsGraphics
Line graphs Area charts Pie charts Bar charts Pictograph 3-D graphics
Control charts– Outliners- observations
that fall outside the control lines
– Runs- data points in a series above or below the central line
Pareto diagram
Oral PresentationsOral PresentationsPreparation
– Length– Content– Opening– Findings and conclusions– Recommendations
OutlineDelivery
– Vocal Characteristics– Physical Characteristics
AudiovisualsAudiovisuals
Chalkboard and whiteboardsHandout materialFlip chartsOverhead transparenciesSlidesComputer drawn visualsComputer animation