Intro to Marketing Research: Measurement and Scaling Comparative Scales Paired Comparison Rank Order Constant Sum QSort

What It Is

Comparative Scales are a family of methods used in marketing research to compare and rank items, concepts, or attributes. One canonical example is the Rank Order Scaling used by the American Marketing Association (AMA) to rank the top 100 most effective advertising campaigns in the United States. This study helps marketers understand what works and what doesn't, informing future advertising strategies. By using comparative scales, marketers can make data-driven decisions to optimize their marketing efforts.

Key Terms & Concepts

• Paired Comparison: A method where respondents compare two items at a time to determine which one is preferred or more desirable.
  • Example: A brand comparing two new product flavors to determine which one to launch first.
  • Formula: None, but often used in conjunction with Constant Sum Scaling.
• Rank Order Scaling: A method where respondents rank a list of items in order of preference or importance.
  • Example: A market research firm ranking the top 10 most influential social media platforms for a brand's target audience.
  • Formula: None, but often used in conjunction with Constant Sum Scaling.
• Constant Sum Scaling: A method where respondents allocate a fixed number of points across a list of items to represent their relative importance or preference.
  • Example: A brand allocating 100 points across 10 product features to determine which ones to emphasize in marketing campaigns.
  • Formula: None, but often used in conjunction with Rank Order Scaling.
• Q-Sort: A method where respondents sort a list of items into categories or ranks based on their similarity or difference.
  • Example: A market research firm sorting a list of 20 product features into three categories: must-haves, nice-to-haves, and deal-breakers.
  • Formula: None, but often used in conjunction with Constant Sum Scaling.
• Triangular Choice Model: A method where respondents choose between three options to determine their preference or importance.
  • Example: A brand choosing between three different packaging options for a new product launch.
  • Formula: None, but often used in conjunction with Paired Comparison.
• Constant Sum Scaling Formula: Σx = 100, where x is the number of points allocated to each item.
• Rank Order Scaling Formula: None, but often used in conjunction with Constant Sum Scaling.
• Q-Sort Formula: None, but often used in conjunction with Constant Sum Scaling.
• Reliability: The consistency of a measure or scale over time or across different respondents.
  • Example: A brand using a consistent survey instrument to measure customer satisfaction over time.
  • Formula: Cronbach's alpha (α) = (k / (k - 1)) * (1 - (∑σ^2 / σ^2_total)), where k is the number of items, σ^2 is the variance of each item, and σ^2_total is the total variance.
• Validity: The accuracy of a measure or scale in measuring what it is intended to measure.
  • Example: A brand using a survey to measure customer satisfaction with a new product launch.
  • Formula: None, but often used in conjunction with Reliability.
• Type I Error: The probability of rejecting a true null hypothesis.
  • Example: A brand conducting a survey to determine if a new product feature is preferred by customers, but incorrectly concluding that it is not preferred.
  • Formula: α = 0.05 (common threshold for Type I error).
• Type II Error: The probability of failing to reject a false null hypothesis.
  • Example: A brand conducting a survey to determine if a new product feature is preferred by customers, but failing to detect a significant preference.
  • Formula: β = 0.20 (common threshold for Type II error).

Common Misunderstandings

• Misunderstanding: Constant Sum Scaling is a type of survey question.
• Correction: Constant Sum Scaling is a method where respondents allocate a fixed number of points across a list of items to represent their relative importance or preference.
• Misunderstanding: Rank Order Scaling is a type of regression analysis.
• Correction: Rank Order Scaling is a method where respondents rank a list of items in order of preference or importance.
• Misunderstanding: Q-Sort is a type of statistical analysis.
• Correction: Q-Sort is a method where respondents sort a list of items into categories or ranks based on their similarity or difference.

Quick Application / Identification

Scenario: A brand wants to determine which of three new product features to emphasize in marketing campaigns. The brand uses a Rank Order Scaling method to rank the features in order of importance. Which method is the brand using?

Answer: Rank Order Scaling. Explanation: The brand is using a method where respondents rank a list of items in order of preference or importance to determine which product feature to emphasize.

Scenario: A market research firm wants to determine which of five social media platforms is most influential for a brand's target audience. The firm uses a Constant Sum Scaling method to allocate points across the platforms. Which method is the firm using?

Answer: Constant Sum Scaling. Explanation: The firm is using a method where respondents allocate a fixed number of points across a list of items to represent their relative importance or preference.

Scenario: A brand wants to determine which of two new product flavors is preferred by customers. The brand uses a Paired Comparison method to compare the flavors. Which method is the brand using?

Answer: Paired Comparison. Explanation: The brand is using a method where respondents compare two items at a time to determine which one is preferred or more desirable.

Last-Minute Revision

• ⚠️ Cronbach's alpha (α) is a measure of reliability, not validity.
• ⚠️ Type I error is the probability of rejecting a true null hypothesis, not failing to reject a false null hypothesis.
• ⚠️ Constant Sum Scaling is a method where respondents allocate a fixed number of points across a list of items to represent their relative importance or preference.
• ⚠️ Rank Order Scaling is a method where respondents rank a list of items in order of preference or importance.
• ⚠️ Q-Sort is a method where respondents sort a list of items into categories or ranks based on their similarity or difference.
• ⚠️ The triangular choice model is a method where respondents choose between three options to determine their preference or importance.
• ⚠️ The formula for Cronbach's alpha is (k / (k - 1)) * (1 - (∑σ^2 / σ^2_total)).
• ⚠️ The threshold for Type I error is 0.05.
• ⚠️ The threshold for Type II error is 0.20.
• ⚠️ Reliability is the consistency of a measure or scale over time or across different respondents.
• ⚠️ Validity is the accuracy of a measure or scale in measuring what it is intended to measure.