Marketing Research Ch 6 Traditional survey research

Auspices bias

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

1 how is it limited?

Error cannot be avoided but can be reduced by increasing sample size choose a subset of population elements that is representative of the whole population

1 What is random (sampling) error,

Error that results from chance variation. Chance Variation: The difference between the sample value and the true value of the population mean

1 why does it occur

Even if all procedures in choosing a sample are adhered to rigorously, there will still be error due to chance

2) mall intercept personal interviewing,

Face-to-face interviews conducted by intercepting people at high traffic locations. Pros - direct customer contact, can test products, human interaction. Cons - expensive, not always representative of the population, bad image.

3) telephone surveys,

Interviews conducted by calling respondents from a central location. Pros - good segmentation, can verify respondent as right person. Cons - cost, response rates slipping, need to have short surveys.

self selection bias

Is a type of nonresponse bias over-represents extreme positions e.g., very satisfied and/or very dissatisfied customers under-represents more moderate positions

surrogate information error

Is discrepancy between the information needed to solve a problem and that sought by the researcher Results from incorrect or poor problem definition leading to a incorrectly stated research objectives

Extremity bias

A category of response bias that results because response styles vary from person to person; some individuals tend to use extremes when responding to questions. can be found in the way some people respond to scales Strong tendency to check end points, e.g., Like 1 2 3 4 5 6 7 Dislike

Acquiescence bias

A category of response bias that results because some individuals tend to agree with all questions A problem on NPD Research

3 processing error

Processing Error results from the incorrect transfer of information from a survey document to a computer A big problem with paper and pencil surveys Data must be validated (double-checked) Can be reduced using scanner technology Internet-based surveys reduce this source of error because of direct data entry

6) on-line survey

Pros - cheap, direct access, can get quick information, can download results can contact hard-to-reach people, good responses rates for web. Cons - response rates low (or e-mail); can't verify responder; security issues not always representative of the population.

1) door-to-door surveys,

Pros - direct customer contact, can demonstrate products, responses rates still good. Cons - expensive, bad image, liability, safety issues.

5) self-administered surveys,

Pros - not too expensive, can get targeted mailing lists, can have longer questionnaires cons-no interviewer to show explain or probe sample may be poor because of nonresponse who actually complete the questionnaire cannot be controlled

4) mail surveys,

Pros - not too expensive, can get targeted mailing lists, can have longer questionnaires. Cons - low response rates, mailing list often out of date, not sure who completed the survey, manual data entry common.

7. What is a disguised survey and what is the reason for the method?

Respondents not initially informed of purpose and/or sponsor of survey Researcher expects response bias to occur Must debrief respondents after questions are answered

systematic error can be these three types

Sample Selection Error results from incomplete or improper sampling procedures or not following appropriate procedures. Population Specification Error results from mis-defining the population of interest Frame Error results from using an inaccurate (incomplete, out of date) sampling frame Sampling frame is a list of the population elements

what is systematic error

Systematic error results from some imperfect aspect of the research design or from a mistake in the execution of the research Exists when results for a sample show a consistent (systematic) tendency to vary in one direction (higher or lower) from the true value of the population parameter being estimated. f

Social desirability bias

caused by respondents' desire, either conscious or unconscious, to gain prestige or appear in a different social role Examples "How often do you donate blood?" "How many hours each week do you spend watching television?" "How many books did you read last year?" "How many hours of 'quality time' do you spend with your children in a typical week?" Use projective (third-person) technique

3 response bias

occurs when respondents tend to answer questions with a certain slant that consciously or unconsciously misrepresents the truth.

3. What are the forms of measurement error

processing error response bias Acquiescence bias Extremity bias Auspices bias Social desirability bias non response bias self selection bias measurement instrument bias interviewer error Surrogate Information Error

non response bias

results from a systematic difference between those who do and do not respond to the survey Huge problem with surveys (esp. cold call mail) Increases with lower response rates Response rate = (# of respondents ÷ # contacts) Use incentives and follow up contacts to increase response rates Always report response rates Try to assess nonresponse bias by capturing a small sample of initial nonrespondents

measurement instrument bias

results from the design of the questionnaire or measurement instrument; also known as questionnaire bias We will examine its causes in detail when we discuss measurement and questionnaire design Limited by using validated measures and Careful pretesting

interviewer error

results from the interviewer's influencing - consciously or unconsciously - the answers of the respondent

6. What factors should be considered in choosing a survey method? Why?

sampling precision, budget, need to expose respondent to various stimuli and have respondent perform specialized tasks, quality of data required, length of questionnaire, incidence rate, degree of structure of questionnaire, time available to complete survey