Source of Error

8/6/2019 Source of Error

1/19

Marketing Research Exam 1 Lecture 6 1

Marketing Research

Dr. David M. AndrusDr. David M. Andrus

Exam 1Exam 1Lecture 6Lecture 6


2/19


Themes of My Presentation Sources of Error in Research DesignsSources of Error in Research Designs

Random Sampling ErrorRandom Sampling Error NonSampling ErrorNonSampling Error

Response ErrorResponse Error

Nonresponse ErrorNonresponse Error


3/19


Potential Sources of Error in

Research DesignsTotal

Error

Random

Sampling

Nonsampling

Error

Response

Error

Nonresponse

Error


4/19


Potential Sources of Error in

Research DesignsResponse

Error

ResearcherErrors

InterviewerErrors

RespondentErrors

-Surrogate Information

-Measurement

-Population Definition

-Sampling Frame

-Data Analysis

-Questioning error

-Recording error-Cheating error

-Inability error

-Unwillingness error


5/19


Field Work/Data Collection

ProcessSelectionof Field Workers

Training of Field Workers

Supervision of Field Workers

Validation of Field Workers

Evaluation of Field Workers


6/19


Errors Related to Field WorkInterviewer

Errors

Nonresponse

Errors

Refusals Not-at-Homes

Respondent Questioning Recording Cheating

Selection Error Error Error


7/19


Methods for Handling Field

Errors Match background characteristics of interviewer andMatch background characteristics of interviewer and

respondentrespondent

Make interviewer instructions clear and writtenMake interviewer instructions clear and written Conduct practice training sessionsConduct practice training sessions

Examine interviewers understanding of the studysExamine interviewers understanding of the studys

purpose and procedurespurpose and procedures

Have interviewers complete the questionnaire and examineHave interviewers complete the questionnaire and examinethe replies to see if there is any relationship.the replies to see if there is any relationship.

Verify a sample of each interviewers interviews.Verify a sample of each interviewers interviews.


8/19


Methods for Handling

Nonresponse Errors NotNot--AtAt--HomesHomes

Have interviewers make advanceHave interviewers make advance

appointments.appointments.

Call back at another time on a different day.Call back at another time on a different day.

Attempt to contact designated respondentAttempt to contact designated respondent

using another approach.using another approach.


9/19



Nonresponse Errors Refusals:Refusals:

Attempt to convince respondent of the value of research andAttempt to convince respondent of the value of research andimportance of their participation.importance of their participation.

Provide advance notice, set a time limit to complete theProvide advance notice, set a time limit to complete thequestionnaire and identify the sponsor.questionnaire and identify the sponsor.

Guarantee anonymity and include only vital questions.Guarantee anonymity and include only vital questions.

Provide incentive for participation.Provide incentive for participation.

Try to get a footTry to get a foot--inin--thethe--door.door.

Use personalized cover letters and followUse personalized cover letters and follow--up contact.up contact. Adjust the results to account for nonresponse.Adjust the results to account for nonresponse.

Give the questionnaire a professional appearance and use shortGive the questionnaire a professional appearance and use shortquestions.questions.

More expensive mailing methods decrease refusalsMore expensive mailing methods decrease refusals


10/19



Noncoverage Errors Improve basic sampling frame using otherImprove basic sampling frame using other

sources.sources.

Select sample to reduce ineligibles on aSelect sample to reduce ineligibles on a

list.list.

Adjust the results by appropriatelyAdjust the results by appropriately

weighting the subsample results.weighting the subsample results.


11/19


Methods for Handling Office

Errors Use field edit to detect the most glaringUse field edit to detect the most glaring

omissions and inaccuracies in data.omissions and inaccuracies in data.

Use a second edit to decide how dataUse a second edit to decide how data

collection instruments containingcollection instruments containing

incomplete answers and obviously wrongincomplete answers and obviously wrong

answers are to be handled.answers are to be handled.

Use closed questions to simplify the codingUse closed questions to simplify the coding


12/19


Methods for Handling Office

Errors When openWhen open--ended questions with multiple codersended questions with multiple coders

are used, divide the task by questions.are used, divide the task by questions.

Have each coder code a sample of the othersHave each coder code a sample of the otherswork to ensure a consistent set of coding.work to ensure a consistent set of coding.

Follow established conventions when coding theFollow established conventions when coding thedata for computer analysisdata for computer analysis

Prepare a codebook that lists the codes for eachPrepare a codebook that lists the codes for each

variable and the categories included.variable and the categories included. Use appropriate methods to analyze the data.Use appropriate methods to analyze the data.


13/19


Improving Response Rates

Reducing

Refusals

Reducing

Not-At-Homes

Prior

Notification

MotivatingRespondents

Incentives

Questionnaire

Design andAdministration

Follow-up

Other

Facilitators

Callbacks

Prior

Notification

Motivating

RespondentsFollow-up

IncentivesQuestionnaire

Design and

Administration

Other

Facilitators

Callbacks


14/19


Potential Survey Problems Total Error:Total Error: Difference between the true informationDifference between the true information

being sought and the collected information resultingbeing sought and the collected information resultingfrom the measurement process.from the measurement process.

Systematic Error:Systematic Error: Caused by a constant bias in theCaused by a constant bias in thedesign or implementation of the measurementdesign or implementation of the measurementinstrument.instrument.

Random Error:Random Error: Caused by inconsistency inCaused by inconsistency inrespondents or their circumstances.respondents or their circumstances.

Collected InformationCollected Information = True information += True information +Systematic error + Random errorSystematic error + Random error


15/19


Pre-Sampling Errors Population Definition Error:Population Definition Error: Results when theResults when the

population is wrongly defined.population is wrongly defined.

SampleSelection Error:SampleSelection Error: Occurs if we select aOccurs if we select a

sample that is not representative of thesample that is not representative of thepopulation.population.

Measurement Instrument Error:Measurement Instrument Error: OccursOccurswhen the survey contains questions orwhen the survey contains questions or

statements that bias respondents or make itstatements that bias respondents or make itdifficult for the researcher to clearlydifficult for the researcher to clearlyunderstand intended responses.understand intended responses.


16/19


Present-Sampling Errors

Procedure Error:Procedure Error: When sample subjects areWhen sample subjects areimproperly selected.improperly selected.

Nonresponse Error:Nonresponse Error: When a high percentage ofWhen a high percentage ofrespondents do not participate.respondents do not participate.

Response Error:Response Error: When respondents answerWhen respondents answerparticular questions or statements incorrectly.particular questions or statements incorrectly.

Interviewer Error:Interviewer Error: Whenever a personWhenever a person

administers a survey, there is a chance that theadministers a survey, there is a chance that theinterviewer will influence responses by providinginterviewer will influence responses by providingadditional information.additional information.


17/19


Post-Sampling ErrorsInformationInformation--Processing Error:Processing Error: Caused byCaused by

mistakes in coding or inputting data into amistakes in coding or inputting data into a

computer for analysis.computer for analysis.


18/19


Selecting A Data-Collection

Method Budget AllocationBudget Allocation

Time AvailabilityTime Availability

Demographic Composition of Targeted RespondentsDemographic Composition of Targeted Respondents

Communication Vehicle AvailabilityCommunication Vehicle Availability

Need for Product or Service DemonstrationNeed for Product or Service Demonstration

Survey Content and LengthSurvey Content and Length

Survey reliability, validity, and generalizabilitySurvey reliability, validity, and generalizability

Survey StructureSurvey Structure

Desired Response RateDesired Response Rate

Desired Sampling PrecisionDesired Sampling Precision


19/19


Response Rates Mail surveys usually produce lowerMail surveys usually produce lower

response rates than telephone surveys.response rates than telephone surveys.

Telephone surveys usually produce lowerTelephone surveys usually produce lowerresponse rates than personal interviewresponse rates than personal interviewsurveys.surveys.

Well designed surveys can produceWell designed surveys can produceresponse rates of 70% (mail), 75%response rates of 70% (mail), 75%telephone, and 80% (personal interviews).telephone, and 80% (personal interviews).

Source of Error

Documents