Identifying and ranking the causes of rework in construction projects: a Relative Importance Index (RII)-based pilot survey

Abstract Rework is understood as the avoidable labor spent on doing an operation or activity that was done incorrectly earlier, and it is still among the main contributors to cost overruns and delays in the construction industry. According to various sources, rework can cost between 5% and 20% of the original contract price. The focus of this analysis lies on discovering, evaluating, and ranking the causes of rework on construction sites via a survey based on the Relative Importance Index (RII) method. The survey, involving 27 causes of reworking grouped into six categories, namely design and drawings, human resources and supervision, contractor/construction services, material and equipment, client-related, and external/environmental factors, was conducted with 30 construction specialists (contractors, consultants, designers, and site engineers) operating in Egypt, Saudi Arabia, Qatar, and the UAE using Likert-type scale. Thus, this sample can be considered as a pilot study. The findings show that failure to comply with specifications (RII = 0.960) and design modification during construction (RII = 0.953) are among the most influential factors of rework. According to this analysis, lack of coordination among project stakeholders (RII = 0.927), poor supervision and quality control on site (RII = 0.927), and ineffective use of quality control checklist, are identified as the major causes of rework in this pilot study. It is analyzed based on the RII where design and documentation (mean RII = 0.856) and client related issues (RII = 0.855) come top of the causes ahead of contractor/construction (RII = 0.849) and human resources and supervision related issue (RII = 0.835) while issue related to material/equipment (RII = 0.799) and external/environmental (RII = 0.769) issue seems to be at relatively lower level. Spearman rank-correlation method was utilized in order to analyze the item level impact and it was found that there was moderate to strong agreement among different stakeholders on the issues causing the problems with least agreement between designers and consultants. Internal-consistency (Cronbach alpha) tests with this sample provided unreliable and mainly negative values, but the paper discusses this openly as a methodological flaw of the study instead of hiding it.