Comparison of univariate and multivariate anthropometric accommodation of the northwest Mexico population

Area of Science:

• Ergonomics and Human Factors
• Anthropometry
• Workplace Design

Background:

• Ergonomic workstation design is essential for preventing work-related musculoskeletal disorders.
• Current design practices often rely on univariate anthropometric approaches.
• A gap exists in multivariate anthropometric analysis specifically for the Mexican population.

Purpose of the Study:

• To compare multivariate human accommodation approaches with the classical univariate method.
• To evaluate the effectiveness of principal component analysis (PCA), archetypal analysis (AA), k-means, and Ward's algorithm (WA) for Mexican anthropometric data.
• To determine which methods yield higher accommodation percentages for sitting workstations.

Main Methods:

• Comparison of multivariate techniques (PCA, AA, k-means, WA) against the univariate percentile-based approach.
• Application of Chi-squared goodness-of-fit and McNemar's tests for statistical validation.
• Analysis focused on the central 98% of anthropometric data for accommodation assessment.

Main Results:

• Multivariate approaches yielded a higher theoretical accommodation percentage than the univariate method for the central 98% of data.
• K-means and archetypal analysis (AA) demonstrated the highest and similar accommodation values.
• Ward's algorithm (WA) and principal component analysis (PCA) followed, showing significant accommodation improvements over the univariate approach.

Conclusions:

• Multivariate methods, particularly k-means and AA, offer enhanced human accommodation for workstation design in Mexico.
• These findings support the development of more effective ergonomic designs, reducing the risk of musculoskeletal disorders.
• The study provides a basis for assessing workstation designs and the impact of diverse anthropometric measurements.