Mediation Analysis: Concept, Methodology, and Application

  

Mediation analysis is a statistical technique used to explore and explain the mechanism or pathway through which an independent variable (X) affects a dependent variable (Y) via a third variable, known as a mediator (M).

 Conceptual Framework

A mediation model answers:

“Does variable X affect Y directly, or through some indirect pathway via M?”

Basic Mediation Structure:

X → M → Y

  • X = Independent Variable (predictor)
  • M = Mediating Variable (intervening mechanism)
  • Y = Dependent Variable (outcome)

 Paths in a Mediation Model:

Path

Meaning

a

Effect of X on M (X → M)

b

Effect of M on Y (M → Y, controlling for X)

c

Total effect of X on Y

c′

Direct effect of X on Y (controlling for M)

ab

Indirect (mediated) effect of X on Y through M

c=c′+abc = c' + ab

 Steps in Mediation Analysis (Baron & Kenny, 1986 method)

  1. Step 1: Regress Y on X to test path c (total effect).
  2. Step 2: Regress M on X to test path a.
  3. Step 3: Regress Y on both X and M to test:
    • path b (M → Y controlling for X)
    • path c′ (direct effect)

Significant ab = indirect/mediated effect

 Testing Mediation:

Method

Description

Sobel Test

Tests significance of indirect effect (ab) using normal approximation

Bootstrapping (preferred)

Re-samples data to estimate indirect effect confidence intervals

Monte Carlo method

Simulates distribution of ab to test mediation

 Types of Mediation:

Type

Condition

Full mediation

c′c' not significant, but abab significant

Partial mediation

Both c′c' and abab significant

No mediation

abab not significant

 Example in Finance (Behavioral):

Research Question: Does financial literacy (X) improve investment decision-making (Y) through risk perception (M)?

  • aa: Does financial literacy influence risk perception?
  • bb: Does risk perception affect decision-making?
  • c′c': Is there a direct effect of financial literacy on investment decisions after accounting for risk perception?
  • abab: Is there a significant indirect effect through risk perception?

 Interpreting Output:

Effect

Meaning

Desired Significance

Direct (c′)

X → Y directly

Significant or not

Indirect (ab)

X → M → Y

Should be significant for mediation

Total (c)

X → Y overall

Generally significant

 Summary

Component

What it Tests

Path

a

X affects mediator

X → M

b

Mediator affects Y (controlling for X)

M → Y

ab

Indirect effect

X → M → Y

c′

Direct effect

X → Y (with M)

c

Total effect

X → Y

 

Comments

Post a Comment

Popular posts from this blog

Shodhganaga vs Shodhgangotri

  Shodhganga and Shodhgangotri are both initiatives by the INFLIBNET Centre (Information and Library Network Centre) in India. They aim to promote research and knowledge sharing in the academic community but serve different purposes. Here’s a comparison: Feature Shodhganga Shodhgangotri Purpose Repository for completed theses and dissertations. Repository for research proposals and synopses. Content Full-text doctoral theses and dissertations submitted by universities. Approved research proposals and synopses of doctoral students. Focus Stage of Research Final stage – after the research is completed. Initial stage – before the research begins. Target Audience Researchers looking for completed, peer-reviewed research work. Scholars looking for examples or ideas for research topics and methodologies. Utility ...
Read more

PLS-SEM is a variance-based modeling approach that has gained popularity in the fields of management and social sciences due to its capacity to handle small sample sizes, non-normal data distributions, and complex relationships among latent constructs. explain

 Your sentence accurately summarizes the advantages of Partial Least Squares Structural Equation Modeling (PLS-SEM) . Here's a clear, expanded explanation of each part of the statement: ✅ “PLS-SEM is a variance-based modeling approach…” Variance-based means PLS-SEM aims to maximize the explained variance (R²) in the dependent (endogenous) variables. Unlike Covariance-Based SEM (CB-SEM) , which focuses on reproducing the entire covariance matrix and overall model fit, PLS-SEM emphasizes prediction and explaining target constructs . ✅ “…gained popularity in the fields of management and social sciences…” PLS-SEM is widely used in: Management research (e.g., leadership, organizational behavior) Marketing (e.g., customer satisfaction, brand loyalty) Information systems (e.g., tech adoption models like TAM/UTAUT) Education and psychology Its popularity stems from its flexibility and suitability for exploratory research and theory building , which ...
Read more

Researches in Finance Area

 Research in finance has distinct characteristics that differentiate it from research in other fields. Here are some key differences:     1. Nature of the Subject Matter -  Finance Research : Focuses on financial markets, investment analysis, corporate finance, financial instruments, risk management, and economic factors influencing financial decisions. -  Other Research Areas : Can cover a wide range of topics, including social sciences (psychology, sociology), natural sciences (biology, physics), humanities (literature, history), etc. Each of these fields has its own set of principles and methodologies tailored to their subject matter.     2. Quantitative Emphasis -  Finance Research : Often heavily quantitative, relying on statistical methods, econometrics, and mathematical modeling to analyze data and test hypotheses. Financial research frequently involves numerical data from markets, economic indicators, and financial ...
Read more