Business Administration Department

BA341 Marketing Research Fall 2007

Dr. Geoffrey P. Lantos

CLASSROOM LECTURE/DISCUSSION OUTLINE

OVERVIEW AND PERSPECTIVE

Introduction to BA 341: Course Structure and Requirements

I. Syllabus and related handouts: Exam Template and Audiovisual Presentations

Audio: 2003 Mercuries #3 Motel 6 “BusinessTalk”

II. Marketing Research Term Paper Project

III. Resources: Using the Business Library, Proofreader's Marks, Classroom Lecture/Discussion Outline, Transparency Handout Packet, 18 Ideas For Becoming A Master Student, Student Information Inventories

IV. Assignments: Optional Marketing Research Statistical Data Analysis Problems, Business

Week/Harris Poll Case

V. Hand In: student information inventories

The Role of Marketing Research and Information Systems

(Related Reading: Chapters 1 and 2 and Appendix 2A)

I. What is Marketing Research (MR)?

· MR and the Strategic Marketing Process

· MR Defined - the function that links the consumer, customer, and public to the marketer through information used to identify and define marketing opportunities and problems; generate, refine and evaluate marketing actions; monitor marketing performance; and improve understanding of the marketing process.

- the systematic and objective process of generating information to aid in

making marketing decisions.

· Characteristics of MR – (1) Scientific

(2) Discontinuous/project basis

vs. Marketing intelligence

· Purpose of MR –

· Forms of information – What about intuition?

· When can we skip MR? Consider:

1. Time Constraints

2. Availability of data

3. Nature of Decision: high vs. low risk; strategic vs. tactical; expensive vs. inexpensive

4. Cost vs. Value/Benefit of Info

II. History of MR

III. Users and Doers of MR/Careers in MR

IV. Basic Research vs. Applied Research

· Basic research (pure research, fundamental research)

· Applied research (decisional research)

V. The Scientific Method (SM) in MR

· How do we derive knowledge?

· SM Defined – A systematic process used to analyze empirical evidence in an objective and accurate manner so as to confirm or disconfirm prior conceptions (hypotheses) in an ongoing, open-minded fashion.

· Three areas where MR fall short of successfully using the SM:

(1) Objectivity

(2) Accuracy:

· Two general traits of SM's results

1. Validity –

2. Reliability –

(3) Continuing and exhaustive nature of the investigation – replication tradition

VI. MR and the MKT'G concept

· MKT'G concept defined -

· Threefold conceptualization

· Video: 1995 Clio Awards: Little Caesar's "More"

MR and Total Quality Management (TQM)

· Customer relationship management (CRM)/relationship marketing

VII. Information (Info) and Decision Making

^{Data vs. info vs. knowledge}

Decision Making Planning Info Data Collection

· Two general purposes of info:

1. Uncertainty reduction

2. Opportunity identification

·Characteristics of valuable information

·Information and the decision-making process: scope of MR activities

VIII Value and Limitations of MR

IX. MKT'G Info Systems (MKIS)

· MIS defined -

· Components:

1. Internal records and reports system

2. MKT'G intelligence system

3. MR system

4. Decision Support System (MDSS)

a. Database system

b. Software

X. Video Applications

· Video: Hard Candy (10 min.)

· Video: Toronto Blue Jays (9 min.)

An Overview of the MR Process

(Related Reading: Chapter 3)

I. Introduction: Decision Making and Marketing Research

· MR Process

· Management Decision-Making Process

· Management Problem (Business Issue) vs. MR Problem

· Forward and Backward Linkages in the MR Process

II. Decision Theory

· Decision making - the process used to resolve a problem or choose from among alternative

opportunities.

· Objective of decision making - find the alternative course(s) of action that maximizes

managerial objectives in an uncertain environment.

· Structuring the decision situation:

· controllable (should we?) variables (decision variables, alternative courses of action) (A₁, A₂,... A_n) Planning res.

· uncontrollable (what if?) variables (environmental variables, situational variables, states of nature) (S₁, S₂,... S_n) Situation analysis res.

· outcomes (what happens?) variables (payoffs, returns, results, responses, performance, objectives) (O₁, O₂,...,O_n) Performance-

monitoring/evaluative res.

· Payoff matrix: S₁ S₂ S₃

A₁ O₁₁ O₁₂ O₁₃

A₂ O₂₁ O₂₂ O₂₃

· Continuum of decision making (states of doubt):

· ignorance/ambiguity uncertainty certainty

III. Stages in the MR process

0. Should research be done?

1. Problem definition and research (res.) objectives

2. Plan res. design (res. methodology)

· Classification of res. designs by function (vs. by technique/methodology—explore, describe, explain, predict, or evaluate/control:

a. Exploratory res. (preliminary situation analysis, developmental res., nonconclusive res).

[Analysis Res.]

1. Secondary data

2. Experience survey (expert opinion)

3. Pilot studies: in-depth interviews, projective techniques, focus groups, and observational

res.

4. Case studies

· Qualitative res. vs. quantitative res.

b. Conclusive res. (confirmatory res., selective res., problem-solving res.)

(1) Descriptive res.: describe [Planning Res.]

(2) Causal res.: explain and prescribe [Planning Res.]

(3) Predictive res.: predict [Planning Res.]

(4) Evaluative res. (performance-monitoring res., accountability res.): control [Control

Res.]

3. Select (primary) data collection technique

· Types of data collection:

(1) Secondary data

(2) Primary data res. techniques

a. Observation

b. Survey

c. Experimentation

4. Sampling plan

· Sampling decisions:

(1) Target population

(2) Sampling unit

(3) Sample size

(4) Sampling procedure/method

5. Develop res. instrument (questionnaire design)

· Decisions:

(1) Question structure

(2) Question sequence

(3) Questionnaire design (layout)

^. Prepare res. proposal

6. Data collection (fieldwork)

7. Data preparation, processing, and analysis

(1) Data preparation: editing and coding

(2) Data processing

(3) Data analysis

8. Report preparation and presentation

9. Follow-up: Evaluation and control of the MR effort

(1) Project control

(2) Total MR effort control

IV. The Res. Program Strategy (vs. res. project strategy)

V. Case Application

· Video Case 3.1: Black Forest Motors/Mercedes Benz (8 min.)

Organizing and Managing the MR Process

(Related Reading: Chapter 4)

I. Degree of MR Sophistication – the marketing manager’s extent of knowledge about, attitude

toward, and usage of MR.

1. Stage of ignorance/intuitive decision making

2. Stage of development

a. Stage of blind faith

b. Stage of disillusionment

3. Stage of research sophistication

II. Internal Organization for MR

A. Intro.

B. Structure and broad capabilities of MR dept.

· 3 levels of management:

Director

Managers (“Senior Analyst”)

Analysts

C. Responsibilities of the director of MR

D. Position of MR in the organizational structure

E. MR Job titles and responsibilities

F. MR Careers

III. Internal Organizational Conflicts and Abuse of MR

A. The marketing Mgr.’s and the MR dept.’s relative roles in the res. process

B. Conflict issues and resolving the conflict between MGT. and MR

1. Research that implies criticism of a manager

2. Money constraints

3. Time constraints

4. Nature of use of information for decision making

a. Intuitive vs. fact-based decision making

b. MR as decision maker vs. recommender

c. Bottom-line executive summaries vs. detail-oriented technical jargon

5. Pseudo-research and organizational politics

a. Rubber stamping/validation of intuition or of prior conceptions: preordained decisions

b. Marketing research as job insurance

IV. External Res. Suppliers

A. Overview

B. Types of external res. suppliers

1. Full-service suppliers

a. Syndicated services

b. Standardized services

c. Customized services

2. Limited-service suppliers

a. Field services

b. Coding and data entry services

c. Tab houses (data analysis services)

d. Analytical services

C. Making the "make or buy" decision

D. Considerations in hiring outside suppliers

IV. Applications

^. Case 4.1: Global Eating

Ethical Issues in MR

(Related Reading: Chapter 4, pp. 86-94)

I. Overview of Ethics

^. Ethics

^. Business ethics

^. Marketing ethics

^. MR ethics

^. Societal norms

^. Domain of: 1. philosophy

a. relativism (situation ethics)

b. absolutism ( moral idealism)

2. religion

^. Why care?

II. Overview of MR ethics

III. The Research Triad

IV. Treatment & Protection of Respondents/Subjects

A. Use of MR as a guise to sell products

^. Sugging

B. Invasion of Privacy of Resps.

^. Right to privacy

C. Abuse of resps.

V. Treatment and Protection of Buyers/Clients

A. Protection against Abuse of Position

B. Protection against Unnecessary Research

C. Protection against Unqualified Researchers

D. Protection of Client Confidentiality

E. Protection against Misleading Presentations of Data

1.Overly technical jargon

2. Failure to round numbers properly

3. Unnecessary use of complex analytical procedures

4. Incomplete reporting

5. Scouring data for answers wanted by the client

VI. Treatment/Protection of Researchers and of the Research Firm

A. Improper Solicitation of Proposals

B. Disclosure of Proprietary Information or Techniques

C. Misrepresentation of Findings of the MR Firm by the Client

D. Excessive Requests

E. Reneging on Promises

F. Availability of Funds

VII. Treatment/Protection of the Public

A. Incomplete Reporting

B. Misleading Reporting

C. Nonobjective Research

VIII. Protection of Competitors

^. Corporate Espionage

IX. Protection of the Research Profession

A. Use of Accepted MR Procedures

B. Inappropriate Use of MR Procedures

^. Politics

^. Social marketing

^. Judicial proceedings - advocacy research

PLANNING FOR THE MR PROJECT AND IDENTIFYING POTENTIAL DATA SOURCES AND RESEARCH DESIGNS

Problem Definition, Research Objectives, and the Research Proposal

(Related Reading: Chapter 5)

I. Intro - Overview

· Management Problems (Decision Statements, Business Issues)

· Decisions

II. The Process of Problem Definition (Business Issue Definition)

A. Determine the decision maker's (management’s) objectives

B. Define the problem: “What should we (mgt.) do to achieve the managerial objectives?”

· Problem definition (decision problem, management problem, business issue)

1. Conduct background study (situation [al] analysis)

· Management interview

· Iceberg principle

· Exploratory res.

2. Distinguish problems from symptoms

3. Identify the relevant unit(s) of analysis (sampling units and/or population elements)

4. Identify the relevant variables/constructs: A, S, O

· Dependent (criterion, predicted, outcome, response) variables vs. independent (predictor, causal, explanatory) variables

· Categorical (discrete, classificatory) variables vs. continuous variables

C. State the Res. Questions and Hypotheses

· Res. question -

· Hypothesis -

C. State the Res. Objectives:

- Criteria:

1. purpose of the MR study (to measure a variable or else the relationship between variables)

2. measurable (quantitative or qualitative)

3. measurement methods/techniques

4. managerial action standard

III. Res. Proposal

· Dummy Tables

IV. Applications

^.Case 5.2: Cane’s Goes International

Exploratory Research and Qualitative Research Methods

(Related Reading: Chapter 6)

I. Definition and Nature of Exploratory Res. (ER) (Developmental Res., Nonconclusive Res.)

· Qualitative res. vs. quantitative res.

II. Uses of ER

III. Qualitative Research Orientations

A. Phenomenology

B. Ethnography

Video: Frontline: “The Merchants of Cool” MTV’s Ethnography (3 min.)

C. Grounded Theory

D. Case Studies

IV. Techniques/Methodologies for ER: Design of Exploratory Studies

(1) Study of secondary data

(2) Experience surveys (expert opinion): uses key informants

· Professional consumer detectives

(3) Case study method

(4) Pilot studies

· Types of questioning techniques

Structured/Standardized Unstructured/Non-standardized

Direct Sample Survey Focus Group, Depth Interview

Indirect/Disguised Projective Techniques

· Motivational research techniques (a. and b. below)

a. In-depth interviews (extended interviews, one-on-ones): probe

· ^. Video: In-depth interview (approx. 10 min.)

b. Projective techniques: disguise

^. Association techniques:

1. Free association (e.g., word association)

^. Completion techniques:

a. Sentence completion

b. Story completion

^. Expressive techniques

2. Role playing

^. Constructive techniques

3. Third person technique

4. TAT

5. Cartoon tests

^. Rozenzweig picture frustration test

6. Miscellaneous methods

· Advantages and disadvantages of motivational res.

c. Focus group interviews: probe

· Methodology

· Video: Trading Cards Focus Group Interview (10 min.)

· Video:^. Simpson's Focus Group (3 min.)

· Characteristics of a good moderator

· Advantages and disadvantages of focus groups

· Trends in focus groups: videoconference focus groups and on-line focus groups

d. Conversations

e. Semi-structured interviews

f. Observational res.

g. Collages

h. Customer visits

V. Conclusion/Cautions on ER

^.^. Video: Hypnosis Focus Groups (up to 20 min.)

Search of Secondary Data

(Related Reading: Chapter 7; Skim Appendix 7A and handout "Using the Business Library")

I. Primary vs. Secondary Data

IV. Advantages and Disadvantages of Secondary Data Research/Evaluating

Secondary Data

III. Common Research Objectives for Secondary Data Research Designs

IV. Sources of Secondary Data:

A. Internal sources

1. Accounting records

2. Sales force reports

3. Corporate libraries

4. Misc.

B. External sources

1. Distributors

a. Libraries and publicly circulated reports

b. Vendors: on-line data services

The Internet and the World Wide Web

(1) Finding secondary data on the Internet

(2) Finding federal government data on the Internet

(3) Internet discussion groups and special interest groups

2. Producers

a. Government sources

b. Commercial sources

1. Syndicated info services

2. Customized res. services

c. Media sources

Trade associations

C. Databases

D Website Databases—A Marketer’s Dream

DATA COLLECTION: PRIMARY SURVEY DATA

Survey Research

(Related Reading: Chapters 8 and 9)

I. Overview of Survey Res. (SR) - the systematic gathering of information from a large,

representative sample of respondents using a structured questionnaire.

· The Nature of SR

· Quantitative res.

· Classification of descriptive survey studies:

a. by medium – method of contact and administration

b. by time frame – one shot vs. over time

1. Cross-sectional studies

2. Longitudinal studies

a. Cohort studies

b. Tracking studies

· Forward linkage: Survey objectives type of info gathered

II. Advantages and Disadvantages of SR

V. Components of Total Survey Error – the variation between the true value of the variable being

measured and its measured value.

A. Total Survey Error = (Random) Sampling Error + Systematic Error

(1) Random (variable) sampling error (RSE) (Nonsystematic error, Random errors, Statistical

sampling errors, Margin of error): threatens reliability

(2) Systematic error (Nonsampling error, Nonvariable error, Sample bias): threatens validity

· Errors vs. Biases

B. Causes of Systematic (Nonsampling) Error

(1) Respondent error

a. Nonresponse error/bias

· Response rate = # completed interviews (net sample)

# of eligible respondents approached

· Sources of nonresponse bias:

(a) nonresponse to entire questionnaire

1. Noncontact error (Not-at-home) = f (availability)

Contact rate (k) = # of respondents contacted

# of eligible respondents approached

(original planned sample size)

2. Refusals (Non-cooperation: Turndowns, Wave-offs) = f (interest,

availability of time, concern for privacy)

- self-selection bias

3. Misc.: inability/incapacity, not found/attempted

(b) fallout (drop-off) - breakoffs during the interview

· Dealing with nonresponse/improving response rates

b. Response bias/error (reactive errors)

· Causes:

1. Deliberate falsification: unwilling (lying)

2. Nondeliberate falsification (unconscious misrepresentation):

(a) unable (misunderstanding, ignorance, memory error)

(b) unduly influenced by the questioning process (i.e., the interviewer and/or the questions asked)

· Categories of response bias and ways to reduce each:

1. Acquiescence bias: yea-saying

nay-saying

2. Extremity bias

3. Interviewer bias

4. Auspices bias

5. Social desirability bias

(2) Administrative error

·Causes:

1. poor research design (poor planning)

2. poor implementation (field errors, analysis errors)

a. Sample selection error

1. population specification error

2. sampling frame error

(a) noncoverage error

(b) overcoverage error

b Interviewer error (recording error)

c. Interviewer cheating

d. Data processing error (blunder)

e. Analyst bias

C. Dealing with Systematic Error

IV. Types of SR Methods

A. Classification by Structure and Disguise

^. Structured questions (formal questions)

^. Un(non)structured questions (informal questions)

^. Disguised questions (indirect questions)

^. Un(non)disguised questions (indirect questions)

^. Four categories of interview:

1. Structured - nondisguised (direct) questions

2. Nonstructured - nondisguised questions

3. Structured - disguised (indirect) questions

4. Nonstructured - disguised questions

B. Classification on Temporal Basis

1. Cross-sectional studies

2. Longitudinal studies

a. Cohort studies

b. Tracking studies

c. Consumer panels

C. Classification by Method of Contact/Communication (Interview Medium)

1. Interviewer-administered surveys (human interactive media)

a. Personal Interviews (PI)

^. Procedure

^. Types of PI's

a. Door-to-door PI

b. Out-of-home PI

1. Traffic interview

2. Shopping mall intercept interview

^. Procedure

^. Advantages and disadvantages of PI's

b. Telephone Interviews (TI)

^. Procedure

^. Advantages and disadvantages of TI's

3. Noninteractive media - self-administered questionnaires

a. Mail Survey (MS)

^. Procedure

^. Advantages and disadvantages of MS

b. Fax surveys (FS)

^. Advantages and disadvantages of FS's

1 c. E-mail surveys (ES)

^. Advantages and disadvantages of ES's

d. Kiosk interactive surveys

e. Internet surveys (IS)

^. Advantages and disadvantages of IS's

· Selecting the Method of Communication

V. Pretesting Questionnaires

VI. Case Applications

Case 8.1: SAT and ACT Writing Tests

Case 9.2: Royal Bee Electric Fishing Reel

Case 9.1 National Do Not Call Registry

DATA COLLECTION: MEASUREMENT CONCEPTS

Questionnaire Design

(Related Reading: Chapter 15 and Appendix 15A)

I. Functions and Importance of Questionnaires (Qaire): a formalized structured set of direct

questions to be asked of respondents for eliciting information to achieve the research objectives.

· Research Instrument (Data Collection Form, Data Collection Instrument)

Questionnaire Structured Observation Form

Survey Research Experiment

Self-Administered Electronic Interactive Personal Interview Phone Interview

Door-to-Door Mall intercept

Audio: 2003 Mercuries #10: Target Market “What Would You Do?”; 2003 Mercuries #41 Target Market “What

Would You Do?”

II. Major Decisions

A. Preliminary Decisions

1. Required info.

a. Statement of management problem and res. purpose/res. objectives b. List of info. needs

c. Draft of statistical analysis plan (dummy tables)

2. Which respondents? (target population)

3. Interview technique (medium)

B. Determine Question Content: What you ask.

1. Is the question necessary? (relevance)

2. Is the question sufficient and specific enough to produce the needed info?

3. Is the resp. able to answer correctly?

a. Does resp. understand the question? (misunderstanding problem)

b. Is resp. informed? (ignorance problem):

c. Can resp. recall the info? (memory error/recall bias problem)

· Types of forgetfulness: omissions, creations/spurious awareness/ghost awareness/misattributions, telescoping

· Three levels of recall: unaided awareness (free recall), aided awareness, recognition

d. Can resp. articulate the answer? (latent needs, trouble expressing, superficial answers) n

4. Is the resp. willing to provide accurate info?

a. Will resp. have to do a lot of work to get the info?

b. Requests for personal info. - will resp. feel it is an invasion of privacy?

c. Requests for embarrassing/sensitive info. (social desirability bias: negatives to avoid)

d. Requests for prestige or normative info. (social desirability bias: positives to do)

C. Determine Response Format (Question format, Question structure): How you structure/frame

answers/response categories (if any).

1. Open-ended questions (unstructured questions, informal questions, free-answer questions)

a. Free-response questions (unprobing )

b. Probing

c. Projective techniques

2. Closed-ended questions (fixed-alternative questions, structured questions, formal questions, itemized

questions)

a. Dichotomous questions (simple-dichotomy questions, two-way questions, binary

questions)

b. Multiple choice questions (multiple category questions, multichotomous questions)

· Rules for writing multiple choice questions (nature and number of alternatives)

· Types of multiple choice questions:

Determinant-choice questions (single-response questions)

Frequency determination questions

Checklist questions (multiple-response questions)

c. Attitude and intention ranking and rating scales (scaled-response questions)

D. Question Wording/Phrasing: “The Art of Asking Questions”: how you ask

1. Define the issue clearly for the respondent

2. Should the question be subjective or objective?

3. Positive or negative statements

4. Avoid complexity: use simple, conversational language

5. Avoid leading and loaded questions

6. Avoid ambiguity: be as specific as possible

7. Avoid double-barreled questions

8. Avoid implied/implicit assumptions

9. Avoid implied/implicit alternatives

10. Avoid asking questions that tax the resp's memory or that pose a burdensome task

E. Question Sequence

0. Introduction to questionnaire

1. Lead-in question

2. Qualifying questions (screener questions, filter questions)

3. Warm-up questions

4. Specific questions

· Guidelines

· Position bias (order bias)

5. Classificatory info. (profile data)

F. Precoding

G. Questionnaire Layout and Design

· Layout of Internet Q’aires

H. Pilot Study: Pretesting and Revising

III. Case 15.1: Canterbury Travels

Basic Concepts of Measurement and Scaling

(Related Reading: Chapter 13)

I. The Concept of Measurement

· Measurement - the assignment of numbers or symbols to characteristics of entities (persons,

objects, states, or events: cases; observations), according to rules which allow those

characteristics to portray the entities.

- rules for assigning numbers to represent quantities or qualities of

characteristics of objects.

^{- Numbers/symbols – values or
scores}

^{- Characteristics – attributes,
properties; concepts or constructs; variables}

· 3 types of characteristics/variables/concepts/constructs: (1) states of being

(2) states of mind

(3) states of action

^{- Rules- specify how numbers or symbols
are assigned to the object’s characteristics:}

^{a. measurement scales (III. below) b. metrics (nits of measurement)}

II. Conceptual and Operational Definitions

· Concept (Construct, Variable, Characteristic) - an abstract idea generalized from specific

facts (if very abstract, called a construct)

·Conceptual definition (constitutive definition) - meaning/domain

· Operational definition (empirical definition)

III. Levels of Measurement (Rules for Measurement)

· Measurement scale - a plan or rule that is used to assign numbers or symbols to characteristics

of objects. It provides arrange of values corresponding to different values in the measured

concept.

· The properties of the four types of scales:

uniquely preserves equal natural

classifies order intervals zero

nominal

ordinal

interval (cardinal)

ratio

1. Nominal scale - equivalence; group membership: classificatory/categorical (categorized)

variables; nonmetric/qualitative (no unit

of measure)/discrete (finite number of

values)

2. Ordinal scale - rank order; relative values: nonmetric/qualitative/discrete

3. Interval scale - differences; relative values: metric/quantitative/continuous (large number of

values)

· Index numbers

Index number = current value/base period value x 100

4. Ratio scale - proportion; absolute values: metric/ quantitative/continuous

· Index measures (composite measures)

- Summated vs. weighted scores

IV. Components of Measurement Error

· Two major categories of measurement error:

1. Random error (nonsystematic error, variable error)

2. Systematic error (bias, constant error)

V. Scale Evaluation: Criteria for Good Measurement

· Accuracy = f (reliability, validity, sensitivity)

A. Criterion 1: Reliability (Precision)

1. repeatability (stability)

2. internal consistency (inter-item consistency, homogeneity, equivalence)

- Reliability Assessment

^. 2 approaches to measuring reliability:

(1) Stability approach

a. Test-retest reliability

(2) Equivalence approach

a. Split-half reliability

b. Equivalent-forms reliability

B. Criterion 2: Validity

- Validity Assessment

(1) Face validity

(2) Content validity

(3) Criterion validity

a. Concurrent validity

b. Predictive validity

c. Convergent validity

(4) Construct validity

(5) Discriminant validity

C. Criterion 3: Sensitivity

Attitude Measurement and Scaling

(Related Reading: Chapter 14)

I. Attitudes Defined

^. Attitude - a learned predisposition to respond in a consistently favorable or unfavorable way to a

specified class of objects, situations, or behaviors.

^. - a predisposition to think, feel, or behave in a positive or negative way toward an object,

issue, situation, or behavior.

^. Opinion – a verbal expression of an attitude.

^. Attitude is a predisposition, a hypothetical construct (intervening variable, latent variable)

Inference

Antecedents - Behavioral Response

Measurable

Observable

^. Characteristics of Attitudes:

1. Unidimensional or multidimensional: direction (polarity, valence) and strength

2. Attitude Structure: tricomponent attitude model

a. Cognitive

b. Affective

c. Conative (Behavioral)

3. Importance: Central attitudes vs. peripheral attitudes

4. Attitude strength (intensity)

II. Overview of Attitude Measurement

A. Observation

1. Perceptual responses (cognition)

2. Sympathetic nervous responses (affect) (physiological reactions)@@

3. Overt behavior

B. Interviewing/Questioning

1. Direct, nonstructured -

2. Indirect (disguised), nonstructured -

3. Direct, structured techniques (self-report measures)

III. Overview of Self-Report Attitude Scales

^. Scaling -

^. 2 major categories of Scales:

1. Basic rating scales (single-item format)

2. Attitude scaling techniques (attitude/attitudinal scales, multi-item format)

IV. Self-Report Attitude Scales

A. Basic Rating Scales (r.s.) (Single-Item Format)

1. Noncomparative r.s. (Monadic r.s.)

a. Graphic noncomparative r.s. (e.g.,"thermometer," "happy

face," and "ladder" scales)

b. Itemized noncomparative r.s.

^. Issues:

1. To what extent and how should we label the categories?

2. Number of categories

3. Balanced vs. unbalanced scale

4. Even or odd number of categories

5. Forced-choice or nonforced-choice scales

2. Comparative r.s.

a. Graphic and itemized comparative r.s.

b. Paired comparisons

c. Rank order r.s.

d. Constant sum scale

B. Attitude Scaling Techniques (Attitude Scales, Attitudinal Scales, Multi-item Format)

Specific Itemized Rating Scales

1. Semantic differential scale

2. Stapel scale

3. Likert scale

4. Thurstone interval scale

V. Measuring Behavioral Intentions (BI)

VI. Selecting the Appropriate Measurement Scale to Use

VII. Multivariate Attitude Measurement

A. Multiattitude Models of Attitude

B. Perceptual mapping

C. Nonmetric Multidimensional Scaling (MDS)

D. Conjoint Analysis

· Randomized Response Technique

DATA COLLECTION: SAMPLING AND FIELDWORK

Sample Design and the Sampling Process

(Related Reading: Chapter 16)

I. Introduction and Definitions

^. Sampling - using a subset of a set of all elements of interest to draw an inference (conclusion)

about characteristics of the set.

^. Population (universe, parent population, target population, statistical population) - the total

collection of elements sharing common characteristics about which we wish to make an

inference(s) based on sample info.

^. Sample - a subset or some part of a larger population that we select and measure or observe.

^. Population element (element) - an individual member of the population that contains the sought

information; the most disaggregated unit of analysis.

^. Sampling unit - a single element or group of elements subject to selection in the sample; a

collection of related population elements; possibly a more aggregated unit of analysis than the

population element.

^. Parameter - the true value of the population characteristic (C) in which we're interested.

^. Estimate - the measurement (M) or "sample statistic" that results from the sample we have selected,

which is out best knowledge of the population parameter. The estimate will likely differ from

the parameter, depending on the degree of presence of statistical/measurement errors.

^. Statistical errors (measurement errors): total survey error - the difference between the value of the

sample statistic (M) and the population parameter (C), composed of:

1. Sampling errors - random sampling error (rse) (margin of error.)

Lowers reliability/precision.

2. Nonsampling error (systematic error, bias)

Lowers validity.

· Nonsampling errors related to the sampling process:

a) Sample design errors (sample selection errors)

1. population specification error

2. sampling frame error (frame bias)

a. noncoverage error

b. undercoverage error (underregistration)

c. overcoverage error (overregistration)

b) Nonresponse error (a respondent error)

^. Accuracy (projectability, generalizability) - the degree of closeness of the sample estimate to

the population parameter; the extent to which the sample estimate is both valid and reliable,

i.e., a "true" measure.

^. Confidence - the degree of certainty we have regarding the accuracy of our sample statistic.

^. Census - an investigation of all the individual elements making up a population.

II. Reasons for Sampling

III. The Sampling Process

A. Define the Relevant Population and Parameters

· Complete operational population definition includes:

1. Population element (Who) - an individual member of the population.

2. Sampling unit (Where) - the unit for sampling; a single element or group of elements

subject to selection in the sample, where the population element can be located.

3. Extent - qualifying criteria (What the unit or element had to do or be to qualify to be in the

sample).

4. Time (When).

· Population specification error

B. Specify the Sampling Frame (frame, working population, operational population) - a list of

either individual population elements or broader sampling units from which a sample may

be drawn; a means of representing the population elements or sampling units.

· sampling frame error (frame bias)

a. noncoverage error: not included

b. undercoverage error (underregistration): underrepresented

c. overcoverage error (overregistration) : double counted and non-target population

elements included.

· Sources of sampling frames

C. Specify the Sampling Unit - the single element or group of elements subject to selection in the sample; the basic unit containing the elements of the population to be sampled.

· Primary sampling unit, secondary (tertiary) sampling unit, etc. (cluster and stratified

sampling)

D. Selection of the Sampling Method - the way the sampling units and/or population elements

are to be selected.

1. Choose between probability sampling and nonprobability sampling:

· Probability sampling (random sampling)

· Nonprobability sampling (nonrandom sampling)

2. Select specific sampling method:

a. Nonprobability samples (nonrandom samples)

1. Convenience sampling

2. Judgment sampling

3. Purposive sampling

4. Quota sampling (quota control sampling)

· Control characteristics

· Incidence rate

5. Snowball sampling (referral sampling)

· ^{Low incidence populations (rare
populations)}

^{6. Internet sampling:}

^a.^{Unrestricted Internet
sample}

b. ^{Screened Internet sample}

c. ^{Recruited ad hoc Internet sample}

d. ^{Panel sample}

e. ^{Opt-in list sample}

b. Probability samples (random samples)

1. Simple random sampling (srs) - single units (vs. clusters) are drawn from an

unstratified (vs. stratified) population with an equal (vs. unequal) probability in a

single-stage (vs. multiple-stage) procedure.

- the sampling procedure that uses a selection procedure that assures that each

population element has a known, equal chance of selection.

^. Ways to Increase Statistical Sampling Efficiency:

2. Systematic sampling (fixed interval sampling)

·^{Skip rate (skip
interval, sampling interval) = ____}

3. Stratified sampling (vs. unstratified sampling)

· ^{Stratum/strata - groups by
classification/stratification variable(s)}

· Proportional/proportionate vs. disproportional/disproportionate stratified sampling (optimal allocation sampling)

- Proportional stratified sampling (proportional allocation) - proportional to

- incidence rate

- Disproportional stratified sampling (disproportional allocation) –

- proportional to incidence rate and to standard deviation of critical dependent variable(s) (i.e., oversample diverse groups)

^. Way to Increase Economic Efficiency

4. Cluster sampling (vs. single-unit sampling)

· ^{Clusters (groups)}

· Area sampling

· Multistage area sampling (vs. single-stage area sampling)

E. Determine Sample Size

F. Specify Sampling Plan

G. Select Sample and Gather Information

H. Validate Sample

IV. Case 16.3: Action Federal Savings and Loan Corp.

Case 16.1 Who’s Fishing?

Review of Statistical Theory and Sample Size Decisions

(Related Readings: Ch. 17; Ch. 21, pp. 564-566)

I. Overview

Sample size= f (1.

c. )

Case 17.1 Pointsec Mobile Technologies

II. Terminology and Notation Review

Two branches of statistics (stats):

(1) Descriptive Stats

(2) Inferential Stats

(3) Notation: Sample Statistics (Estimates) Population Parameters

`x - sample mean m - population mean

s - standard deviation S or σ - standard deviation of a population

of a sample

s² - sample variance S² or σ² - population variance

p - proportion of a sample p = P - proportion of a population

n - sample size N - population size

x_i - i^th observation (a single observation on a variable)

III. Basic Concepts of Descriptive Statistics

A. Summary Measures

^. Raw data X₁, X₂,...X_n for n observations (data points) on some variable X - a data set on a

variable

^. Sorted data (arrayed data) X₍₁₎, X₍₂₎, ....X_(n)

^. Grouped data (tabled data): Set up class intervals (categories: categorical data):

^. Frequency distribution (frequency table)

e.g. Price f

<$75 7

75-80 12

80.1-85 15

85.1-90 13

>90 3

Total 50

^. Histogram (Bar Chart)

30 35 | 40 45

Class Mark

^. Percentage distribution - relative frequency for each class interval = f_i/∑f = f_i/n

eg. Cumulative frequencies

Value f % Greater Than % Less Than %

5 to 9.99 2 10 20 100 0 0

10 to 14.99 5 25 18 90 2 10

15 to 19.99 8 40 13 65 7 35

20 to 24.99 4 20 5 25 15 75

25 to 29.99 1 5 1 5 19 95

20 100 0 0 20 100

^. Probability (Pr.)

^. Random variable (vs. constant variable)

^. Probability distribution E.g.: Value Probability

5 to 9.99 .10

10 to 14.99 .25

15 to 19.99 .40

20 to 24.99 .20

25 to 29.99 .05

1.00

^. Probability histogram Pr.|

12.5% 37.5% 37.5% 12.5%

^. Proportion

^. Discrete random variable

^. Continuous random variable

B. Statistical Summarization

· Measures of Central Tendency (centrality)

1. Mean =Arithmetic Mean `x = ∑ X_i

Trimmed mean

2. Median M_d(50th percentile)

3. Mode M_o

^. Skewness - degree of symmetry

4. Fractiles:

. Quartiles

. Deciles

. Percentiles

· Measures of Dispersion (Spread, Variability)

1. The range

2. Interquartile range (Midspread) = Q₃– Q₁

Quartile deviation Q₃– Q_1/2

3. Variance σ² – the mean squared deviation of all the values from the mean.

– the extent to which a random variable is dispersed around its mean value.

4. Standard deviation σ – the square root of the variance.

IV. The Normal Distribution

^. Normal curve/distribution – a symmetrical (bell-shaped, mirror images) percentage or

probability distribution/curve for a variable.

^. Standardized normal curve/distribution

^{.
Standardized normal table}

^{.
Standardized values: Z = x -}`x

^.Calculating areas under the standardized normal curve

V. Miscellaneous Distributions Categorized by Variable Investigated

^. Population distributions - a frequency (or percentage or probability) distribution of the

elements of a population on some variable, X; a distribution of X's in a population.

^. Sample distributions - a frequency (or percentage or probability) distribution of the elements

of a sample; a distribution of X’s in a sample drawn from a population

^. Sampling distribution - a theoretical frequency (or percentage or probability) distribution of

the values of some sample statistic (e.g.,`x or p) calculated for each possible sample of a

given size, n, drawn from a particular population.

^. Sampling distribution of the mean - a frequency (or percentage or probability) distribution of

the values of sample means (`x ) calculated for each possible sample of a given size, n;

a distribution of means from all possible samples of a given size, n.

^. Standard error - the standard deviation of a sampling distribution.

- a measure of rse used in calculating confidence intervals.

^. Standard error of the mean - the standard deviation of the sampling distribution of the mean

= S`_x= s/Ön - rse for the sampling distribution of the mean.

. Central limit theorem (normal approximation rule) - for large simple random samples from a

population that is not normally distributed, the sampling distribution of the mean will be

approximately normal. As the sample size is increased, (1) the sampling distribution of the

mean will more closely approach the normal distribution, and (2) the standard error of the

mean will decline.

VII. Drawing Inductive (Statistical) Inferences Through Confidence Intervals (Confidence Bands)

^{.
Induction – the process of empirically deriving general principles from
particular facts or}

^instances.

^. Inductive inference (statistical inference, statistical estimation) - estimation of population

parameters that we do not know (e.g., m, S, P) from sample statistics that we do know (e.g.,

`x, s ,p).

^. Two kinds of estimation procedures:

1. Point estimate - a single number estimate of a population parameter from sample data. The

sample statistic (e.g.,`x, p).

2. Interval estimate (range estimate) - two points between which a population parameter is estimated to lie with some stated level of confidence = point estimate ∓ rse at (1 - a) %

confidence level.

^. confidence interval estimate = sample statistic ∓ rse: a measure of precision.

^. precision (reliability) - the degree of rse in a study's interval estimate.

a. absolute precision – units (for means)

b. relative precision – percentage points (for proportions)

^. confidence level (confidence coefficient) (1-a) - the degree to which one can be certain

that an interval estimate approximates the true value of the population parameter;

the probability that a particular confidence interval will include the true population value.

· Calculating a confidence interval:

A. Means (absolute precision)

- Steps: 1. Use`x as center of interval

2. Estimate S (e.g., 1/6^th of range)

3. Estimate S`_x= s/Ön= standard error of mean

4. Determine confidence level desired, i.e., determine the Z value associated with

the confidence level (1- a) desired. This confidence level (e.g., 1-.05=.95)

should be divided by 2 (e.g., .95/2=.475) to find what percentage of the area under

the curve must be included on each side of the mean (e.g., for .4750, Z=1.96).

5. Construct the confidence interval `x∓ Z_a_/2 x s/Ön ,where Z_a_/2 x s/Ön = margin of

error

or `