📊

PLS-SEM Analysis and Reflective Models

Apr 10, 2025

Statistics Tutorial: Reflective Measurement Model and PLS-SEM Analysis

Introduction

  • Welcome to the statistics tutorial shared by the presenter, referred to as Mrs. Statistics.
  • Focus on the latest application of PLS-SEM (Partial Least Squares Structural Equation Modeling) and the reflective measurement model.
  • The session aims to discuss reading outputs from the Smartfren application.
  • Research focuses on empowerment and religiosity affecting motivation and job satisfaction in local government settings.

Reflective Measurement Model

  • Evaluation Criteria:
    • Convergent Validity:
      • Loading factors should be above 0.70.
    • Composite Reliability (CR):
      • Assesses the reliability of the measurement items.
    • Discriminant Validity:
      • Use HTMT (Heterotrait-Monotrait Ratio) and Fornell-Larcker criterion for assessment.
    • Cross Loadings:
      • Check loadings of indicators on their respective constructs.

Smartfren Application Tutorial

  • Creating a New Project:
    • Import data and set up the project in Smartfren.
    • Drag and drop features to build the model.
  • Model Evaluation:
    • Check loading factors and discard items below the threshold (e.g., 0.70).
    • Analyze composite reliability and AV (Average Variance Extracted).
    • Assess discriminant validity using HTMT values.
    • Evaluate structural models using t-statistics and p-values.

Structural Model Analysis

  • Multicollinearity Check:
    • Ensure VIF (Variance Inflation Factor) is below 5 to avoid multicollinearity.
  • Statistical Significance:
    • Use bootstrapping to determine significance.
    • Check if empowerment, religiosity, and motivation significantly influence satisfaction.
  • Mediation Analysis:
    • Assess indirect effects of variables through mediation.

Model Goodness-of-Fit

  • R-Square Values:
    • Classify effects into low, moderate, and high based on R-square values.
  • Goodness-of-Fit Metrics:
    • Evaluate RMSE (Root Mean Square Error) and predictive power.
    • Compare PLS model predictions with linear regression to assess accuracy.

Reporting and Conclusion

  • Suggests reporting on PLS analysis for statistical research.
  • Highlight the importance of discriminant validity and predictive power.

Final Notes

  • Presentation aims to provide comprehensive understanding of PLS-SEM.
  • Offers statistical resources and ongoing support for learners engaging in advanced statistical modeling.
  • Contact Information:
    • Mention of a comprehensive 1700-page statistics book available for further learning.
  • Channel Subscription:
    • Encouragement to subscribe for updates and continuous learning resources.

This guide summarizes the key aspects of using the Smartfren application for PLS-SEM analysis focusing on reflective measurement models, including evaluation criteria, model setup, and result interpretation.

Full transcript