Chapter 11 Mktg Research
Learning Objectives
- Understand the various types of errors in marketing research.
- Know how to manage and control data collection errors.
- Learn what constitutes data quality issues in datasets.
Key Concepts in Marketing Research Process
- Establish the need for research: Identify underlying motivations for conducting research.
- Define the problem: Clearly articulate the problem to be studied.
- Establish research objectives: Set specific aims to achieve through the research.
- Determine research design: Choose the method (qualitative/quantitative) for conducting the research.
- Identify information types and sources: Decide what data is necessary and where it can be obtained.
- Design data collection forms: Create tools for gathering information from respondents.
- Determine sample plan and size: Decide on how many and which individuals to include in the study.
- Collect data: Execute the plan and gather responses.
- Analyze data: Evaluate the collected data to derive insights.
- Communicate insights: Report and present findings effectively.
Types of Errors in Survey Research
- Sampling Error: Error due to the size and selection of the sample.
- Non-Sampling Error: All other errors that occur not related to the sample methodology, including:
- Nonresponse errors
- Data gathering errors
- Data handling errors
- Data analysis errors
- Interpretation errors
Field Data Collection Errors
- Fieldworker Errors: Mistakes made by those administering the questionnaires.
- Respondent Errors: Mistakes made by the respondents themselves, which can be intentional or unintentional.
Types of Fieldworker Errors
Intentional Errors:
- Cheating: Fieldworkers fabricating responses or misrepresenting data.
- Leading respondents: Influencing the answers given by respondents.
Unintentional Errors:
- Interviewer characteristics: Bias introduced by the fieldworker's attributes (e.g., accent, demeanor).
- Misunderstanding: Mistakes due to incorrect administration of the survey.
- Fatigue: Errors that arise when fieldworkers are tired.
Types of Respondent Errors
Intentional Respondent Errors:
- Falsehoods: Deliberate inaccuracies in answers.
- Nonresponse: Decisions to skip questions or not participate.
- Speeding: Rushing through the survey, potentially leading to erroneous responses.
Unintentional Respondent Errors:
- Misunderstanding: Respondents not comprehending the questions.
- Guessing: Providing answers when unsure.
- Attention loss and fatigue: Not paying full attention or becoming tired of the survey process.
Quality Control Mechanisms in Data Collection
- Supervision: Overseeing fieldwork to ensure compliance with data collection protocols.
- Validation: Confirming interviewer activity through checks and reviews.
- Training: Conducting orientation and role-play to prepare interviewers.
- Anonymity and confidentiality assurances: Providing guarantees to respondents about data privacy to reduce inaccuracies.
- Use of incentives: Encouraging participation through cash or gifts.
- Validation checks and third-person techniques: Techniques to ensure truthful responses while maintaining respondent comfort.
Nonresponse Error
- Definition: Nonresponse refers to situations where prospective respondents do not participate in a survey or fail to answer specific questions.
- Types of Nonresponses:
- Refusal: Direct decision not to participate.
- Break-off: Stopping participation partway through the survey.
- Item Omission: Skipping a question but responding to others.
Measuring Nonresponse Error
- The industry standard for calculating the survey response rate includes distinguishing between eligible respondents, completions, and nonrespondents (CASRO Response Rate).
- Various formulas can quantify response rates by considering different levels of response and completion.
Data Quality Issues to Monitor
- Incomplete Responses: Identifying incomplete surveys that signify break-offs.
- Yea-saying and Nay-saying: Detecting response patterns where participants consistently agree or disagree, which can skew results.
- Middle-of-the-road Responses: Noting excessive neutral or no-opinion responses which could indicate disengagement or indecision.
Dataset and Coding
- Dataset: Arrangement of data points in rows and columns for analysis.
- Data Coding: Assigning codes to responses for ease of analyzing and interpreting results through a coded system.
- Data Codebook: A document that defines the datasets, variable names, and the coding scheme associated with survey questions.