Surveys ask a representative sample of respondents for information using verbal or written questioni

1
Chapter 8 Survey Research

• Surveys ask a representative sample of
  respondents for information using verbal or
  written questioning

2
Advantages
Survey Criteria

• Quick/Timely
• Inexpensive(costs)
• Flexible/Efficient
• Accuracy (/-)

Problems

• Poor Design (Bias)
• Improper Execution
• Sample Participation

Reliability - Validity - Representative ness
3
EXHIBIT 8.1 Categories of Survey Errors
4
Tree Diagram of Total Survey Error
Random Sampling error
Total Error
Administrative error
Systematic error (bias)
Respondent error
(aka Non Sampling Error)
5
Random Sampling Error

• A statistical fluctuation that occurs because of
  change variation in the elements selected for the
  sample (cannot be eliminated)

Systematic Error (Non Sampling Error)

• Systematic error results from some imperfect
  aspect of the research design or from a mistake
  in the execution of the research

6
Sample Bias

• A sample is biased when the results consistently
  deviate in one direction from the true value of
  the population parameter
• Key Point One must estimate or know the true
  value
• Self-Selection Bias - A bias that occurs because
  people who feel strongly about a subject are more
  likely to respond to a survey than people who
  feel indifferent about it.

7
Random Sampling Slackers!
Example
Marketing News Pg. 4, 7 July 2003
8
(No Transcript)
9
Tree Diagram of Total Survey Error
Non response error
Respondent Error . . A type of sample bias due to
some respondents action or inaction
Response bias
10
Non response errorstatistical difference between
actual respondents a perfect,100 response rate

• Non respondents - People who refuse to cooperate
  Factors affecting response
• 1. Perceived amount of work required
• 2. Intrinsic interest of the topic
• 3. Characteristics of the sample
• No contacts/Not-at-homes

11
Tree Diagram of Respondent Error
Acquiescence bias
Response Bias. . . respondents consciously
or unconsciously misrepresent the truth
Extremity bias
Interviewer bias
Auspices bias
Social desirability bias
12
Acquiescence Bias

• A category of response bias caused by a tendency
  to agree with all or most questions or to concur
  with a particular position.

Extremity Bias

• The tendency of some individuals tend to use
  extremes when responding to questions.

13
Auspices Bias
Interviewer Bias

• A response bias when the presence of the
  interviewer influences answers.

• Bias in the responses of subjects caused by the
  respondents being influenced by the organization
  conducting the study.

14
Social Desirability Bias

• Bias in responses caused by respondents desire,
  either conscious or unconscious, to gain prestige
  or appear in a different social role.

15

• Deliberate Falsification - people give false
  answers to appear intelligent, conceal personal
  information, avoid embarrassment
• Average-person hypothesis
• Individuals may prefer to be viewed as average,
  so they alter their responses to conform more
  closely to their perception of the average person.

• Unconscious Misrepresentation people
  consciously try to be truthful but response bias
  occurs from the question content or some other
  stimulus.
• Misunderstanding the question, Unable to recall
  details, Inability to translate feelings into
  words, after-event underreporting

16
Social Desirability example Religious event
participation
17
Tree Diagram of Administrative Error
Data processing error
Administrative Error
Sample selection error

• Research mistakes/blunders
• Confusion
• Neglect
• Omission
• Improper execution

Interviewer error
Interviewer cheating
18
(No Transcript)
19
Administrative Error

• Interviewer error - field mistakes or deliberate
  cheating - filling in fake answers or falsifying
  interviewers
• Data processing error - incorrect data entry,
  computer programming, or other procedural errors
  during the analysis stage.
• Sample selection error - improper sample design
  or sampling procedure execution
  - Self-selection bias
• Over represents extreme positions
• Under represents indifferent

20
Case example
Study Examines Decisions to End Ventilator Support

• Study results Families considered doctors
  feelings (about what patients wanted) in
  determining whether to stop life-support.

• Life-support is four times as likely to be
  stopped when doctors believe that patients would
  rather not be on support.

• Study Sponsor Canadian Critical Care Trial Group

WSJ,18Sept2003
21
Unit of Analysis

• 851 Patients
• 15 intensive-care units
• Canada United StatesSweden Australia

22
Family Decisions

• Family members may be uninformed of patients
  wishes.
• A patients best way to make their wishes known
• Discuss matter with a close family member
• Advanced directives

10 of patients in the study had documentation
prior to admission.
Advanced Directives- Documents that detail
which life-support measuresa patient would like
at the end of life.
23
Doctors opinions took precedence over patients
age or degree of sickness.
Non Determinant Factors

• Age
• Measures of illness
• i.e. Health of organs

24
Influences on Decision

• Chances of Survival
• Patients taking drugs to boost their blood
  pressure.
• Patients whose physicians believed they would
  suffer from brain damage.
• Patients whose physicians believed they had less
  than a 10 survival rate.

2.0
Times as likely to be removed from breathing
support
2.5
3.5
25
Ethical Issues

• Doctors feel that their own biases about a
  patient will influence the families decisions.
• Family may be influenced to stop life-support due
  to benefits of death.
• i.e. Inheritance
• Pain/suffering

26

27
Methods of Survey Communication
Internet
TELEPHONE/Fax
MAIL
PERSONAL
INTERVIEW
28
Classifying Survey Research Methods

• Structured and Disguised Questions
• Structured question imposes a limit on the
  number of allowable responses.
• Unstructured question does not restrict the
  respondents answers.
• Undisguised (direct) questions assume the
  respondent is willing to answer.
• Disguised (indirect) questions assume the
  purpose of the study must be hidden from the
  respondent.

29
Survey Classification by Time Periods

• Cross Sectional study collect data at 1point in
  time
• Longitudinal Study - A survey at different times
  to analyze response continuity changes over
  time. - Tracking
  study uses successive samples to compare trends
  changes in variables such as brand image
• Consumer panels -A longitudinal survey of the
  same sample of people to record (in a diary)
  their attitudes and/or behavior over time.
  -
  Internal validity (3-5 months)
  - a true panel uses
  different questions each time period - an
  omnibus panel uses the same questions

30
Harris Pole Longitudinal Research
Source Americans Feel More Isolated, Less
Empowered, Poll Shows, The WSJ, Dec 8, 2005,
http//online.wsj.com.
31
Example
32
Consumer Panel Disadvantages

• Non Representative
• Mortality (drop outs)
• Cost ( plus time)
• Testing Effects (tune in, sensitivity,validity)
• Selection bias (initial refusals)

33
Total Quality Management and Customer
Satisfaction Surveys

• Total quality management -A business philosophy
  that emphasizes market-driven quality as a top
  organizational priority.
• Quality is determined by the customer/buyer.
• The degree to which a good or service corresponds
  to buyers expectations.

34
Determinants of Physical Product Quality

• Performance
• Features
• Conformance with specifications
• Reliability
• Durability
• Serviceability
• Aesthetic design

35
Determinants of Product Service Quality

• Access to service
• Communication
• Competence
• Courtesy
• Reliability
• Credibility

36
Stages in Tracking Quality Improvement
Continuous Quality Improvement
Initial quality improvement
Commitment and Exploration
Bench- marking
37
Stage 1 Commitment and Exploration
Tracking Quality Improvement

• Management makes a commitment to total quality
  assurance
• Marketing researchers explore external and
  internal customers needs, motivations and
  problems.

38
Stage 2 Benchmarking

• Research establishes quantitative measures as
  benchmarks or points of comparison
• Overall customer satisfaction ratings and product
  quality ratings of specific attributes
• Employee actual performance and perceptions

39
Stage 3 Initial Quality Improvement

• Tracking wave 1 measures audience trends
• Establishes a quality improvement process within
  the organization by translating target audience
  quality issues into the internal vocabulary of
  the organization.
• Establishes performance standards and
  expectations for improvement.

40
Statistical Data
Example Customer service

• 2/3 of those experienced rage
• 16 said that they desired revenge
• 3 took legal action

41
Stage 4 Continuous Quality Improvement

• Consists of many consecutive waves with the same
  purposeto improve over the previous period.

42
Table 1Families in Panel Purchasing Each Brand
43
Table 2 Number of Families in Panel Purchasing
each Brand in each Period
At Time T 0 one
At Time T o