Agile Governance: Examining the Impact of DevOps on PMO: A Systematic Review

Appendix A.1. Risk of Bias in Included Studies

The risk of bias in the 35 included studies was qualitatively assessed by reviewing each publication's methodology, sample size, study design, and reporting clarity. Due to the diverse nature of the literature—ranging from empirical case studies to conceptual and theoretical frameworks—a formal tool such as the Cochrane Risk of Bias Tool was not applied. However, articles lacking transparency in data sources or those relying heavily on anecdotal evidence were noted as lower in methodological rigor. Most studies published in peer-reviewed journals presented moderate to high methodological quality, suggesting an acceptable level of credibility for inclusion in the synthesis.

Appendix A.2. Summary Statistics and Effect Estimates

Given the qualitative nature of this systematic review and the diversity of study designs and outcomes, no meta-analytic summary statistics or effect estimates were calculated. Instead, thematic synthesis was employed to extract patterns, implications, and theoretical contributions from the body of literature. Where possible, qualitative comparisons were drawn across domains such as deployment frequency, time-to-market, and organizational agility improvements, though these remain descriptive rather than statistical.

Appendix A.3. Characteristics and Bias Summary of Contributing Studies

The included studies (n=35) span various geographies (UAE, US, Europe, India) and industry contexts (finance, e-commerce, software services). A minority of studies (approximately 20%) did not disclose clear sampling methods or participant selection criteria, indicating potential risks of selection or publication bias. Nevertheless, the remaining studies offered rich insights with robust theoretical grounding, thereby balancing potential weaknesses in a few. Table 4 (List of Selected Studies) presents detailed characteristics including authorship, year, focus, and contribution.

Appendix A.4. Statistical Syntheses

This review did not employ statistical synthesis techniques such as meta-analysis due to heterogeneity in study designs, metrics, and methodological frameworks. The review instead adopted a qualitative thematic analysis approach, grouping findings into emergent themes across governance transformation, cultural adaptability, automation, and organizational performance.

Appendix A.5. Investigation of Heterogeneity

No formal subgroup analysis or statistical heterogeneity tests were conducted. However, thematic variations were observed between studies focusing on developed economies (e.g., EU, US) and those from emerging economies (e.g., UAE, India). Regional differences were evident in how organizations adopted DevOps principles, with regulatory environments and infrastructure maturity serving as moderating factors. These findings are discussed under the "Regional Variations" subsection.

Appendix A.6. Sensitivity Analyses

Sensitivity analysis was not conducted as no meta-analytic procedures or effect size calculations were performed. The robustness of findings was ensured through the use of a well-defined inclusion/exclusion framework, multiple database sources, and triangulation of themes across a large sample of studies. Future research employing meta-analytic techniques could benefit from incorporating sensitivity analysis to test the stability of outcomes under varying assumptions.

Appendix A.7. Risk of Bias Due to Missing Results

While no quantitative synthesis was performed, potential risk of reporting bias was mitigated through broad and systematic searches across four major academic databases, including grey literature where possible. However, it is acknowledged that publication bias may persist, especially in underreporting of failed DevOps integration efforts or non-English studies excluded due to language constraints.

Appendix A.8. Certainty of Evidence

Due to the narrative synthesis approach and absence of effect size aggregation, a formal GRADE assessment or other certainty grading was not feasible. Nevertheless, consistency across the studies in key findings—such as the benefits of automation, CI/CD, and cultural transformation—supports moderate confidence in the derived conclusions. The integrated use of both ADRM and DCF further enhances the theoretical robustness of the synthesized evidence.