Publication record
Title: Risk Assessment of Road Construction Projects Using the Monte Carlo Simulation Approach: Empirical Results From Tanzania
| Field | Detail |
|---|---|
| Authors | Mbona Mtala Luhida; Ismail W. R. Taifa |
| Journal | Journal of Engineering |
| Publisher platform | Wiley Online Library |
| First published | 12 June 2026 |
| Volume and issue | 2026(1) |
| Article | 3261435 |
| DOI | 10.1155/je/3261435 |
| Access | Open access research article |
The study
The researchers assessed uncertainty in road-construction projects in Tanzania. Their work combined a literature review, expert responses, statistical analysis, fitted probability distributions, and Monte Carlo simulation. The published abstract reports data from 55 road-construction projects and discusses operational, technical, managerial, and governance risks.
The goal was not merely to assign a single score to each risk. By modelling uncertainty through probability distributions, the study generated a range of possible outcomes and used those outcomes to prioritize risks.
Where Phitter enters the method
The article states that Phitter was used to build discrete probability distributions from expert ratings of risk likelihood and impact. Those fitted distributions became inputs to the Monte Carlo workflow.
The methodology names SPSS 27 for descriptive analysis, Phitter for distribution fitting, and Argo 4.1.3 for Monte Carlo simulation. It reports using Chi-square and Kolmogorov–Smirnov tests in the fitting process. Table 8 lists fitted models—including binomial and hypergeometric distributions—for the likelihood and impact variables attached to individual risk identifiers.
This is a concrete applied use of Phitter: observed ratings were converted into model inputs for an engineering risk study. The article cites Phitter as reference 57 and links to phitter.io.
Why the mention matters
The study demonstrates a workflow beyond a demonstration dataset. Distribution fitting supports a downstream decision model, and the selected distributions affect the uncertainty propagated by the simulation.
It also shows why transparent reporting matters. A reader should be able to identify the input data, candidate distributions, selection tests, fitted parameters, and simulation software in order to reproduce or challenge the result.