2020 journal article

How Does Experience with Delay Shape Managers' Making-Do Decision: Random Forest Approach

JOURNAL OF MANAGEMENT IN ENGINEERING, 36(4).

By: Y. Zhang, A. Javanmardi n, Y. Liu, S. Yang, X. Yu, S. Hsiang*, Z. Jiang*, M. Liu n

co-author countries: China 🇨🇳 United States of America 🇺🇸
author keywords: Making-do; Random forest; Decision tree; Information theory; China; United States; Construction
Source: Web Of Science
Added: June 15, 2020

Making-do, a decision to start a construction task despite knowing that its preconditions are not fully ready, is a complex dilemma for construction managers. Managers’ previous making-do decisions and the resulting consequence, delay, can have a significant impact on future making-do decisions. To understand how managers’ experience with delay impacts their making-do decision and how it is handled differently in different countries, two surveys were administered, one in China and one in the United States (US), and 260 usable responses were collected. This study used: (1) the Mann–Whitney U test to examine whether delaying task starting time, when lacking precondition readiness, pays off with shorter delays; (2) a random forest approach to find important causes of delay that contribute to a making-do decision; and (3) an entropy-based decision tree to determine how much uncertainty in making-do decisions can be reduced by knowing managers’ experience with delays in past projects. Results showed that in the United States, managers who preferred the making-do approach experienced up to 60% less task duration delay; whereas Chinese managers who preferred making-do experienced up to 100% more task duration delay due to lack of readiness in labor, equipment, material, management, and information flow. The contributions to the body of knowledge are the development of a random forest approach to quantitatively examine the relative importance of the causes of delay to the making-do decision and to reveal the fundamental differences in culture and management traditions that cause the difference between the two countries. The methods presented in this study will enable others to use a similar random forest approach repetitively for classification, prediction, and variable selection problems in civil engineering. The findings of this study will help project managers better understand underlying factors that trigger making-do decisions in China and the United States, and have more efficient collaboration and communication when they work on projects located in a foreign country.