Assessing statistical differences between parameters estimates in Partial Least Squares path modeling.

Area of Science:

• Business and Management
• Social Sciences
• Statistical Modeling

Background:

• Partial Least Squares Structural Equation Modeling (PLS-SEM) is a widely adopted statistical technique across disciplines.
• Existing methods for comparing PLS-SEM parameter estimates are limited to multi-group analyses of different subpopulations.
• There is a practical need for methods to assess statistical differences between parameter estimates derived from the same sample.

Purpose of the Study:

• To introduce a novel statistical technique for testing differences between parameter estimates within a single sample in PLS-SEM.
• To provide practical guidance on implementing this new method for researchers.
• To demonstrate the utility of the proposed technique using a validated model.

Main Methods:

• Development of a new statistical approach for comparing parameter estimates from the same sample in PLS-SEM.
• Application of the new technique to a reduced version of the Technology Acceptance Model.
• Comparison with existing parametric and non-parametric approaches for multi-group analysis.

Main Results:

• The study presents a viable method for assessing statistical differences between parameter estimates obtained from a single sample.
• The proposed technique extends the capabilities of PLS-SEM beyond multi-group comparisons.
• The application to the Technology Acceptance Model illustrates the practical implementation and benefits.

Conclusions:

• The developed technique offers a significant advancement for structural equation modeling using partial least squares.
• Researchers can now more rigorously compare parameter estimates within a single sample, enhancing analytical depth.
• This contributes to more robust and insightful statistical analyses in various research fields utilizing PLS-SEM.