Development of a Decision-Support Model for Agile Hybrid Engineering Project Delivery in Technology Driven Organizations
DOI:
https://doi.org/10.63125/v7jjv448Keywords:
Agile, Hybrid, Engineering, Decision-Support, PerformanceAbstract
This study developed and quantitatively validated a decision-support model for Agile hybrid engineering project delivery in technology-driven organizations by examining how project conditions, delivery-practice intensity, technological capability, governance, and human factors influenced multidimensional project performance. A quantitative, nonexperimental, cross-sectional explanatory and predictive design was applied to 1,586 eligible engineering projects from 54 technology-driven organizations, representing 2,438 project professionals across eight engineering and technology sectors. The final project-retention rate was 91.0%, data completeness reached 98.6%, and construct reliability coefficients ranged from .84 to .93. Hybrid delivery represented 45.0% of projects, compared with 31.0% Agile and 24.0% predictive delivery. Hybrid projects demonstrated the strongest overall performance, recording mean schedule variance of 5.8%, cost variance of 4.9%, budget adherence of 88.6%, milestone achievement of 91.3%, delivery reliability of 90.1%, and overall project success of 89.5%. Hybrid delivery exceeded predictive delivery by 9.8 adjusted performance points, producing a large effect size of Cohen’s d = 0.86. Multiple regression explained 59% of overall performance variance, with hybridization intensity emerging as the strongest positive predictor (β = .29), followed by team capability (β = .24), digital capability (β = .21), and governance strength (β = .19), all p < .001. Context-specific analysis showed that Agile-intensive hybrid delivery achieved a performance score of 90.2 under high technological novelty and requirements volatility, balanced hybrid delivery achieved 88.9 under combined high complexity and uncertainty, and predictive-intensive hybrid delivery achieved 89.1 among highly regulated projects valued above $10 million. Gradient boosting produced the strongest predictive performance, achieving 88.3% holdout accuracy, an F1-score of .877, ROC-AUC of .931, PR-AUC of .908, and R² of .72. Five-fold cross-validation accuracy remained stable at 87.6%. The findings demonstrated that Agile hybrid engineering delivery functioned most effectively as a configurable continuum in which adaptive practices and predictive controls were quantitatively aligned with project complexity, uncertainty, volatility, regulation, organizational capability, and technological maturity.


