The DEMATEL Method

DEMATEL method
The DEMATEL method is developed for dealing with complicated and interrelated problems (Lin and Tzeng, 2009). The DEMATEL method by separating variables into cause and effect groups and accounting the causal relationships among these variables provide insights for managers to solve problems more efficiently (Li, 2009). The main purpose of DEMATEL method is to help managers to identify and construct interrelations among important criteria (causes of reduction in labor productivity in this study) (Tsai et al., 2010). Thus, in this study, the final product of the DEMATEL is an influence-relations map (IRM) that is a visual representation of the interdependencies which exists among the factors affecting construction labor productivity.

The first section contains general information about the respondents and the second section addresses causes leading to reduction of labor productivity. A list of the major causes of reduction in labor productivity obtained from the first group of experts was presented, and the respondents were asked to show a level that reflects the impact of one factor on another on the basis of the integrated causal loop diagrams among factors obtained from the hybrid framework. The scale of scores ranged in an ascending order from 0 to 4 that each number indicate different levels of influence. It would be better if it was possible to take two steps to survey the experts, selected by sampling. In the first step, they were surveyed to identify the most important factors based on the factors relative importance measure and after reaching a consensus on the selected factors, in the second step, they were surveyed to determine the level of impact each one can produce on the others. However because of some limitation, it was not possible to follow all these steps. Therefore, the experts were asked to add the missing factors in the questionnaire. However, they validated the opinions of the first group of local experts and the factors selected by them. All the experts chosen to be in this study were senior project managers of the related firms and had experience of participating in relatively large building construction projects in the Iranian

For example, if the sum of the experts scores is assumed 120, for an expert with score equal to 4 the weighting factor will be equal to 4/120 = 0.033. Then the impact of one factor on another factor can be calculated as I= ∑_(j=1)^n▒〖W_j×E_j 〗 , where I is the impact of one factor on another factor, j is the counter of experts. Wj is the weighting factor of jth expert and Ej is the level of impact that is stated by jth expert.
Table3: Experts features
The complex interrelated structure of factors affecting construction labor productivity
The complex interrelated structure among the affecting factors on labor productivity is shown in Figure 3 and each of its causal loops is reviewed briefly in this section. Note that to have better understanding of the model, the auxiliary variables such as the work done and the remaining work in each loop were also taken into account, which were not mentioned in Table 2 .Since based on the SD principle the normal sense of direction of variables should be positive Sterman (2000), the name of some factors were also changed in the causal loop diagrams.

Figure 3: Integrated causal loop diagrams among the factors affecting labor