Intro to Marketing Research: Factor Analysis Confirmatory Factor Analysis CFA Introduction

What It Is

Confirmatory Factor Analysis (CFA) is a statistical method used to test the validity of a measurement model by examining the relationships between observed variables and their underlying latent constructs. A classic example of CFA in marketing research is the study by Fornell and Larcker (1981) on the measurement of customer satisfaction. They used CFA to validate a scale that measured customer satisfaction with a product, which is crucial for marketing decision-making as it helps businesses understand their customers' needs and preferences.

Key Terms & Concepts

• Confirmatory Factor Analysis (CFA): A statistical method that tests the validity of a measurement model by examining the relationships between observed variables and their underlying latent constructs.
  • Example: Fornell and Larcker (1981) used CFA to validate a scale that measured customer satisfaction.
• Latent Construct: A theoretical concept that cannot be directly observed, but is measured through observed variables.
  • Example: Customer satisfaction is a latent construct that is measured through observed variables such as satisfaction with product quality and price.
• Measurement Model: A statistical model that describes the relationships between observed variables and their underlying latent constructs.
  • Example: A measurement model for customer satisfaction might include observed variables such as satisfaction with product quality, price, and customer service.
• Factor Loading: The strength of the relationship between an observed variable and its underlying latent construct.
  • Formula: Factor loading = (observed variable - mean) / (standard deviation of observed variable)
  • Example: A factor loading of 0.8 indicates that the observed variable is strongly related to the latent construct.
• Cronbach's Alpha: A measure of the reliability of a scale.
  • Formula: Cronbach's alpha = (k / (k - 1)) * (1 - (Σσ^2_x / σ^2_y))
  • Example: A Cronbach's alpha of 0.9 indicates that the scale is highly reliable.
• Goodness of Fit Index (GFI): A measure of how well the measurement model fits the data.
  • Formula: GFI = (χ^2 / df) / (χ^2 / df) + (1 - (χ^2 / df))
  • Example: A GFI of 0.9 indicates that the measurement model fits the data well.
• Type I Error: The probability of rejecting a true null hypothesis.
  • Example: A Type I error occurs when a researcher concludes that a measurement model is invalid when it is actually valid.
• Type II Error: The probability of failing to reject a false null hypothesis.
  • Example: A Type II error occurs when a researcher concludes that a measurement model is valid when it is actually invalid.
• Maximum Likelihood Estimation (MLE): A method of estimating the parameters of a statistical model.
  • Example: MLE is used to estimate the factor loadings and other parameters of a measurement model.
• Structural Equation Modeling (SEM): A statistical method that combines CFA and path analysis to examine the relationships between latent constructs.
  • Example: SEM is used to examine the relationships between customer satisfaction, loyalty, and retention.
• Observed Variable: A variable that can be directly observed, such as a survey question or a product rating.
  • Example: Satisfaction with product quality is an observed variable.
• Latent Variable: A variable that cannot be directly observed, but is measured through observed variables.
  • Example: Customer satisfaction is a latent variable.
• Measurement Error: The error that occurs when an observed variable does not accurately measure the underlying latent construct.
  • Example: Measurement error occurs when a survey question does not accurately measure customer satisfaction.

Common Misunderstandings

Misunderstanding: CFA is only used for exploratory research.
Correction: CFA is used for both exploratory and confirmatory research. In exploratory research, CFA is used to identify the underlying latent constructs, while in confirmatory research, CFA is used to test the validity of a measurement model.

Misunderstanding: CFA is only used for survey research.
Correction: CFA can be used for any type of data, including survey, experimental, and observational data.

Misunderstanding: CFA is only used for marketing research.
Correction: CFA is used in many fields, including psychology, sociology, and business.

Quick Application / Identification

A marketing researcher wants to examine the relationships between customer satisfaction, loyalty, and retention. Which statistical method would be most appropriate for this study?

Answer: Structural Equation Modeling (SEM)

Explanation: SEM is a statistical method that combines CFA and path analysis to examine the relationships between latent constructs, making it the most appropriate method for this study.

Last-Minute Revision

⚠️ CFA assumes that the data is normally distributed.
⚠️ CFA assumes that the measurement model is correct.
⚠️ Cronbach's alpha is a measure of reliability, not validity.
⚠️ Goodness of Fit Index (GFI) is a measure of how well the measurement model fits the data.
⚠️ Type I error occurs when a researcher rejects a true null hypothesis.
⚠️ Type II error occurs when a researcher fails to reject a false null hypothesis.
⚠️ Maximum Likelihood Estimation (MLE) is a method of estimating the parameters of a statistical model.
⚠️ Structural Equation Modeling (SEM) is a statistical method that combines CFA and path analysis.
⚠️ Observed variables are variables that can be directly observed.
⚠️ Latent variables are variables that cannot be directly observed.
⚠️ Measurement error occurs when an observed variable does not accurately measure the underlying latent construct.
⚠️ Factor loading is the strength of the relationship between an observed variable and its underlying latent construct.
⚠️ Cronbach's alpha = (k / (k - 1)) * (1 - (Σσ^2_x / σ^2_y)) ⚠️ GFI = (χ^2 / df) / (χ^2 / df) + (1 - (χ^2 / df))