Kilbert (1994) [5] first acknowledged the need for sustainable construction during the first international conference on sustainable construction that was held in Tampa, Florida, United States of America where sustainable construction was defined as ‘creating and operating a healthy built environment based on resource efficiency and ecological design’ [6]. Since then, others have contributed to the discourse. For example, CIB (1999) concluded that sustainable construction is the minimisation of resource consumption, maximisation of resource reuse, use of renewable and recyclable resources, protection of the natural environment, creation of a healthy and non-toxic environment and pursuit of quality in creating the built environment. Similarly, G. Ofori [7] recognised sustainable construction as ‘creating construction items using best-practice clean and resource-efficient techniques, from the extraction of the raw material to the demolition and disposal of its components.’ From the perspective of a developing country, the Agenda 21 for Sustainable Construction in Developing Countries defined sustainable construction as ‘a holistic process aiming to restore and maintain harmony between the natural and the built environments and create settlements that affirm human dignity and encourage economic equity’ [8].

Airport Cooperative Research Program [9] recognised sustainable construction as ‘practices that have sustainability benefits during the construction stage of a project, including those benefits that may result from decisions made during the planning and design stages of a project.’

O’Connor [10], defined ‘sustainable construction’ as the processes, decisions and actions during the construction stage of capital projects that enhance current and future environmental, social and economic needs whilst considering project safety, quality, cost and schedule.

According to the study [11], sustainable development has three essential aspects: i) environmental responsibility, ii) social awareness and iii) economic profitability. Achieving the optimum balance between these factors supports true sustainability (Kibwami & Tutesigensi, 2016). Similarly, sustainable construction embraces three aspects, namely, social, economic and environmental considerations; by contrast, the traditional perspective is concerned with the economy, utility and durability [12, 13]. However, several researchers, such as [6] and [14], added a technical pillar as the fourth issue; the technical pillar includes durable, reliable and functional building structures with the inherent desire to inextricably link quality in all project processes.

In social and natural sciences, every element has a life cycle, which represents maturational and generational processes driven by mechanisms of reproduction in the natural population [15]. Ritz [16] suggested that the construction project life cycle encompasses the conceptual stage through project definition, execution, operation and finally, demolition. According to the study [5], the project life cycle is a process of planning, development, design, use, maintenance and deconstruction [17, 18]. However, Shen and Tam [19, 20] decomposed the project life cycle into inception, design, construction, operation and demolition.

Weaver et al. [21] suggested that the construction stage transfers the project design plans into reality. However, O’Connor et al. [10] defined the ‘construction stage’ as all fabrication/jobsite/field activities and decisions, starting with construction/fabrication contracting and planning for site mobilisation continuing to initial operations, final performance testing and handover of the completed facility.

Several factors affect sustainability performance during the construction stage. These factors can be conveniently examined through the three lenses of economic, social and environmental sustainability factors.

The target population in this study is all 38 consultants with valid registration in the Association of Engineers (AE, 2017). The following formula in Eq. (1) was used to determine the sample size of unlimited population [47].

(1)

Where, SS = Sample Size; Z = Z value (e.g. 1.96 for 95% confidence level); p = percentage selecting a choice, expressed as a decimal (0.5 used for sample size needed); and c = confidence interval, expressed as a decimal (e.g. 0.05 = ±5).

Correction for finite population:

Where, New SS = corrected Sample Size.

The sample size for the 38 consultants was calculated as follows:

Therefore, the calculated sample size is 35 consultants based on a 95% confidence level.

The questionnaires were consequently distributed to 35 consultants asking their contribution in ranking the determined 55 factors by using an ordinal scale. The ordinal scale used was 1 = very low influence, 2 = low influence, 3 = moderate influence, 4 = high influence and 5 = very high influence. A total of 31 completed questionnaires were received from the consultants, representing a good response rate of 88.57%. Tests of validity (internal and structure) and reliability of the questionnaire were conducted. In addition, the Kolmogorov-Smirnov test for normality was used to ascertain the parametric nature of data distribution.

The collected raw data were first sorted, edited, coded and then entered into computer software. Two software were used, the Excel sheet and SPSS. The ordinal scale is a rating data, which uses integers in ascending or descending order. The Relative Important Index (RII) was used in the analysis. Moreover, Analysis of Variance (ANOVA) tests, frequencies and percentiles were used. The RII method has been widely used in construction research for calculating and formulating attitudes with respect to surveyed variables. Then, the RII was computed using the following formula in Eq. (2) [48-51]:

(2)

Where, W is loading given for every factor by the respondent, between 1 to 5, (n1= number of respondents for who agreed strangely, n2 = number of respondents who don't agree, n3 = number of respondents for neutral, n4 = number of respondents for agree, n5 = number of respondents who agreed strangely). A is the first load (i.e. 5 in the study) and N is the total number of samples. RII value is between 0 and 1. The analyzed data was finally presented using descriptive methods for easy interpretation and to enable and to make comparison easy.

Table 2 illustrates that the first run had 16 variables, and the KMO value was 0.469 (should be > 0.50); thus, several weak variables were omitted to meet statistical requirements for sample adequacy. In addition, the Bartlett’s test for sphericity with chi-square = 244.857 at a significance level of p-value = 0.00 (should be < 0.05) meets the requirements. Cronbach’s alpha was 0.70, which was accepted. After the final run, six variables remained, the KMO value was 0.600, and the Bartlett’s test for sphericity (with chi-square = 244.857) at a significance level of p-value < 0.001, hence meeting the requirements. Cronbach’s alpha was 0.739, which was accepted.

Following the tests and using principal component analysis, six variables remained, and 10 variables were deleted because their factor loading values (less than 0.5) were considered very low (Table 3).

The mean of the economic sustainability variable equals 3.10 and RII = (62.04%) (Table 4). The mean is close to the neutral, indicating low awareness of the economic sustainability amongst the respondents.

The results clarify that the variable ‘professional fees of engineers and consultants’ was ranked first with mean = 3.29 and RII = (65.81%). These results are consistent with those of the study [23] who reported that this variable is important because it has a managerial focus, that is, engineers and consultants control most other variables affecting sustainability performance during the construction stage of project development. The factor ‘energy cost’ came second with mean = 3.23 and RII = (64.52%). These findings agree with those of [52], who noted that on average, green buildings use 25% less energy and emit 34% less carbon dioxide than commercial buildings. Moreover, this study [53] considered that green buildings conserve energy, land, water and materials. In the third position was the variable ‘use of full equipment capacity’ with a mean = 3.03 and RII = (60.65%). The use of equipment without full capacity can reduce their default life and expose them to depreciation in a short time, increasing cost and consuming new resources, which in turn affects sustainability [10]. Other variables were right-sizing of construction equipment, inspection and maintenance of construction equipment and cost of repairing errors and defects.

Table 5 illustrates that 18 variables were entered into the first run, and the KMO value was 0.499 (should be > 0.50); several weak variables were omitted to meet statistical requirements for sample adequacy. In addition, the Bartlett’s test for sphericity has Chi-square = 352.705 at a significance level of p-value = 0.00 (should be < 0.05), hence meeting the requirements. Cronbach’s alpha was 0.839, which was accepted. After the final run, the 12 variables remained, the KMO value was 0.600 and Bartlett’s test for sphericity with chi-square = 352.705 at a significance level of p-value < 0.001, thereby meeting the requirements. In addition, Cronbach’s alpha was 0.864, which was accepted.

Following the tests and using principal component analysis, 12 variables remained, and six variables were deleted because their factor loading values were less than 0.5, which was considered very low (Table 6).

The mean of the social sustainability variable = 3.00 and RII = (59.95%) (Table 7). The mean is equal to the neutral level, and most of them are below the neutral. Therefore, respondents are unaware of the social sustainability amongst respondents. This lack of awareness is a result of the traditional definition of environmental sustainability.

The results illustrate that the variable ‘availability of knowledge and skills in the labour force’ was ranked first with mean = 3.45 and RII = (69.03%). Given that sustainability is a relatively new concept within the Gaza Strip, consultants lack pertinent knowledge and skills, and further effort should be made by industry leaders and the government to educate professionals within the sector (c.f) [32]. The variable ‘increased burden on infrastructure as a result of the use and depletion of natural resources’ was ranked second with mean = 3.23 and RII = (64.52%). Any project developed can invariably increase the demand for supporting infrastructure (i.e. water, road, energy, services and space), thus placing a further burden on achieving sustainable construction practices. In the third position was the variable ‘public awareness’ with mean = 3.19 and RII = (63.87%). This result is consistent with that of [32] who illustrated that increasing public awareness is essential to delivering sustainable practices. Other factors were promotion and development of capacity and skills for the labour force, working conditions, physical space of the building, aesthetic options of the building, participation of all parties in project monitoring and decision making, health and safety in the workplace, directly creating jobs for local employment, indirectly creating jobs for local employment and project control guidelines.

Table 8 illustrates that the first run had 22 variables, and the KMO value was 0.605 (should be > 0.50), which meets the statistical requirements for sample adequacy. In addition, the Bartlett’s test for sphericity with chi-square = 431.325 at a significance level of p-value = 0.00 (should be < 0.05), meeting the requirements. Cronbach’s alpha was 0.85, which was accepted. After the final run, 13 variables remained, the KMO value was 0.600 and Bartlett’s test for sphericity with chi-square = 431.325 at a significance level of p-value < 0.001, hence meeting the requirements. Cronbach’s alpha was 0.891, which was accepted.

Following the tests and using principal component analysis, 13 variables remained, and nine variables were deleted because their factor loading values were less than 0.5, which was considered very low (Table 9).

The mean of the ‘environmental sustainability’ variable = 3.14 and RII = (62.88%) (Table 10). Most variables have a mean more than 3, indicating awareness of the environmental sustainability amongst respondents. This value is accepted with the traditional definition of environmental sustainability in developing areas.

The results exhibit that the variable ‘environmental regulations’ was ranked first with mean = 3.52 and RII = (70.32%). This finding concurs with that of [38] who proposed that environmental regulations support the implementation of a government-backed sustainability agenda, and sustainability cannot be readily achieved without them. This view is recently supported by [23], who expressed that environmental regulations are the most important factors affecting sustainability performance. The factor ‘pollution generation’ came second with mean = 3.42 and RII = (68.39%). The most adverse environmental effect of construction activities is the pollution generated [10, 15, 45], and the sector infamously contributes from 40% to 50% of the world’s greenhouse gas emissions [54]. In the third position was the variable ‘waste generation’ with mean = 3.32 and RII = (66.45%). These results concur with previous findings [25, 41, 44] that waste generation is an important factor that affects sustainability performance during the construction stage. Other factors are reuse of products, changes in the environment leading to the discomfort of people and the biological system, institutional interest on the environmental aspect, communication of environmental management information, recycling of products, use of sustainable temporary facilities (such as desks and bathrooms) during the project, environmental management technology, use of recyclable/renewable materials and ecology preservation.