Field Work & Research Errors

Field Work & Errors in Research

• On-site data collection is needed when the target population cannot be reached through other methods like phone or online surveys.
• This is used in situations like observing consumer behavior in stores or conducting interviews at events.
• Good data collection requires diverse locations to represent the population.
• Systematic sampling helps reduce bias.

Difference Between Sampling and Non-Sampling Errors

• Sampling Errors: stem from how the sample is selected; can be measured and controlled through probability sampling.
• Non-Sampling Errors: Result from human or process errors during data collection (e.g., respondent misunderstandings or interviewer mistakes). These cannot be precisely measured, but they can be reduced.

Types of Non-Sampling Errors

• Respondent Errors:
  • Intentional (e.g., social desirability bias, nonresponse).
  • Unintentional (e.g., misunderstanding questions, fatigue, distractions).
• Fieldworker Errors:
  • Intentional (e.g., fraud, falsifying data).
  • Unintentional (e.g., unclear instructions, fatigue).

Data Collection Quality Control

• Intentional Fieldworker Errors: Address through supervision, validation of responses.
• Unintentional Fieldworker Errors: Mitigate these with role-playing during training, enforced breaks.
• Intentional Respondent Errors: Prevent these by assuring anonymity, offering incentives, and using validation checks.
• Unintentional Respondent Errors: Address through clear instructions, reversed scales, and prompters.

Types of Nonresponse Errors

• Refusals: Participants declining to take part.

Database Structure, Coding, and Data Validation

• Assign unique numeric codes to each category to ensure consistency.
• Codes should be comma-separated without spaces.
• Data Sheet Stores raw data.
• Variables Sheet Contains description and coding for analysis.
• Validate data by checking for missing entries, formatting, value consistency, capitalization of labels.

Descriptive Data Analysis

• Descriptive Analysis: Summarizes data using averages and percentages.
• Inferential Analysis: Makes conclusions about the population from the data.
• Difference Analysis: Identifies significant differences between groups.
• Associative Analysis: Examines relationships between variables.
• Predictive Analysis: Forecasts future trends based on current data.

Measures of Central Tendency

• Mean: The average value.
• Median: The middle value in a dataset.
• Mode: The most frequently occurring value.

Measures of Variability

• Range: Difference between highest and lowest values.
• Standard Deviation: Measures data dispersion around the mean.

Grid Questions Analysis

• Use averages to identify trends.
• Calculate percentages to highlight significant categories.
• Use cross-tabulations for segment-specific insights.

Analyzing Open-Ended and Multiple Response Questions

• Categorize responses into recurring themes.
• Calculate percentages for each theme.

Calculating Sample Size

• Sampling Method: The process for selecting participants.
• Sample Size: The number of participants.
• Key Concepts:
  • Representativeness: Ensuring the sample reflects the population.
  • Confidence Interval Method: Defines the range of acceptable error.
  • Population Variability: Measures diversity within the sample.
  • Confidence Level: Typically 95% accuracy.

Associations Between Variables

• Understanding associations between variables is vital for identifying relationships, predicting outcomes, and understanding trends in data.
• Non-Monotonic: General association without a consistent direction.
• Monotonic: One variable consistently increasing or decreasing as the other changes.
• Linear: A straight line relationship between two variables.
• Curvilinear: A non-linear relationship between two variables.

Crosstab Analysis

• Purpose: Identifies how categorical variables (e.g., age, brand) relate to each other.
• Components:
  • Column Variable: Independent variable
  • Row Variable: Dependent variable
• Data presented in a table format with percentages.
• Significance is evaluated using Chi-Square tests.
• Presence: Look for any relationship via Chi-Square (p ≤ 0.05).
• Direction & Strength: Identify general trends. (e.g., higher percentages in specific groups).

Inference, Confidence Interval, and Hypothesis Testing

• Parameter Estimation: Estimates population parameters (e.g., average sales).
• Hypothesis Testing: Tests assumptions about populations (e.g., campaign effectiveness).
• Confidence Interval (CI): Defines the range of likely values for a population parameter.
• Hypothesis Testing Steps:
  • Formulate null and alternative hypotheses.
  • Collect data and calculate test statistics.
  • Decide whether to reject or fail to reject the null hypothesis based on the p-value.

Segmentation: Testing of Differences

• Testing for differences helps identify variations in preferences, or demographics.
• Provides actionable insights for marketing.
• Different data types uses different testing methods. (e.g. Metric vs. Non-metric data).

ANOVA and Conjoint Analysis

• ANOVA (Analysis of Variance): Compares means across multiple groups.
• Conjoint Analysis: Evaluates trade-offs among product features.
  • Identifying feature preferences, forecasting product success.

The Marketing Research Report

• Importance: Communicates findings and recommendations to stakeholders.
• Sections:
  • Front Matter
  • Body: Objectives, methodology, findings, recommendations.
  • End Matter: References, appendices.