**3. Materials and Methods**

The study used a quantitative approach to differentiate between the level of usage and investment in three digital technology categories (data acquisition, processing, and communication) across three classes (phases)—pre-COVID, during COVID, and post-COVID. In order to select the frequently used digital technology for the purpose of this research study, a thorough literature survey of published materials was conducted. This included journal articles, periodicals, and books that discuss the leading digital technologies currently in

use in the construction industry. In the next stage, a questionnaire survey was developed and distributed to construction professionals (owners, consultants, contracts, and project managers) in the UAE using online means. The survey was sent to 100 construction professionals, and 39 responses were collected in the end, representing a 39 percent rate. The survey consisted of 21 questions, and the duration of the survey was approximately 10 min. The questions were developed in a clear manner, where examples from the literature review about the different digital technologies in each category were given to make the questions comprehensible and easy to follow for the respondents. The questionnaire questions are presented in Appendix A. The first section of the survey consisted of general information about the respondents' companies or organizations. These included years of experience, respondents' role, and whether the organizations or companies are local or international. The remaining sections of the survey were designed to elicit the perceptions of the respondents on the frequency of usage and level of investment of the three digital technology categories across the three classes: pre-COVID, during COVID and post-COVID. The questions related to the level of usage of all three digital technology categories pre-COVID and during COVID followed a Likert scale of 1–5, where 5 represented "almost always" and 1 represented "never". The level of investment of all three digital technology categories pre-COVID and during COVID followed a Likert scale of 1–5, where 5 represented "very high" and 1 represented "very little". All questions related to the class post-COVID, as being a future state, followed a Likert scale of 1–5, where 5 represented "strongly agreewill increase greatly" and 1 represented "strongly disagree-will return to pre-pandemic level". We focus on the different digital technology (DT) categories rather than the specific technologies. The paper intentionally states only 3–4 examples within each DT category. The respondents are able to identify the example technologies within each category, as they were included in the survey questions. The survey questions are provided in Appendix A.

The data were then analyzed using both descriptive statistics and inferential statistics. The descriptive statistics mainly consisted of comparing the means of the three class populations in the frequency of usage and level of investment of the digital technologies. Additionally, the means of the small projects were compared against those of the large projects across the three class populations in frequency of usage and level of investment of the three digital technology categories, while for the inferential statistics, multivariate analysis of variance (MANOVA) was conducted, as the data comprised multiple dependent variables, namely, the level of investment in, and frequency of usage of the three digital technology categories, and a single independent qualitative variable designated as classes: pre-COVID, during COVID, and post-COVID. MANOVA is used to determine whether there is a difference between the level of investment and frequency of usage of the three digital technologies across these classes. In particular, the researchers were interested in determining whether, during COVID, a peak in the level of usage and investment in digital technology is evident, as expected in order to meet the demands imposed by the pandemic. The statistical analysis mainly examined whether or not the average of the dependent variables differs between the three categories of the class variable. MANOVA is an extension of the analysis of variance (ANOVA) in order to incorporate more than one dependent variable by combining the multiple dependent variables into a single optimum value to maximize the difference between the classes. Furthermore, MANOVA is also more appropriate than ANOVA, because it provides additional insights into the effects of the independent variables on the dependent variable [52,53].

MANOVA is a very effective statistical tool that has been previously used in the construction context in literature. For instance, Zhao et al. [54] have used MANOVA analysis to test whether or not different factors such as construction type, technology level, climate, and conditioned floor area significantly affect energy use and whether or not the effect changes over time. The authors tested the "between-subject" effect to analyze the factor's effect across all building units and used the "within-subject" effects to test the factor's effect over time. The authors modeled the "between-subject" effect by fitting the

sum of the repeated measures to the model effect, while the "within-subject" effect was modeled using a function that fits differences in the repeated measures.

There are three main assumptions when conducting MANOVA analysis. First, the data should be multivariate normally distributed, while the second assumption is concerned with the equality of the covariance matrices for all treatments in the study. The third and last assumption states that all observations must be independent of each other, as this will affect the significance level reported in results later on [55]. Furthermore, Wilk's Lambda test statistic (Λ) is used in MANOVA to test whether there are differences between the means of the three classes on the frequency of usage and on the level of the investment of the three digital technologies or, in other words, to test the two-null hypothesis H0 shown below.

$$\mathbf{H}\_0 \text{ for frequency of usage of the three digital technology} = \begin{bmatrix} \mu\_{11} \\ \mu\_{12} \\ \mu\_{13} \end{bmatrix} = \begin{bmatrix} \mu\_{21} \\ \mu\_{22} \\ \mu\_{23} \end{bmatrix} = \begin{bmatrix} \mu\_{31} \\ \mu\_{32} \\ \mu\_{33} \end{bmatrix} \tag{1}$$

$$\begin{bmatrix} \mu\_{11} \end{bmatrix} \quad \begin{bmatrix} \mu\_{21} \end{bmatrix} \quad \begin{bmatrix} \mu\_{21} \end{bmatrix} \quad \begin{bmatrix} \mu\_{31} \end{bmatrix}$$

H0 for level of investment of the three digital technologies = ⎣ μ<sup>12</sup> μ<sup>13</sup> ⎦ = ⎣ μ<sup>22</sup> μ<sup>23</sup> ⎦ = ⎣ μ<sup>32</sup> μ<sup>33</sup> ⎦ (2)

> where μip, i = (1,2,3) is the number of the three classes' populations (pre-COVID, during COVID, post-COVID), while "p" is the number of dependent variables, which, in this study, is three, since there are three digital technology categories (acquisition, processing, and communication) for the variable frequency of usage and the variable level of investment. Furthermore, to test whether there is no difference between the three population mean vectors, Λ will be used according to Equation (3) [55], as it is the test statistic preferred for MANOVA, where H0 will be rejected if Λ is small. Lambda is a value that ranges between zero and one; a null hypothesis would be rejected if Lambda was close to zero, but it should be considered in conjunction with a small *p*-value as well, where the *p*-value here represents the probability that measures the consistency between the data and the hypothesis being tested [51]. The alpha level that the *p*-value will be compared against in this study is 0.1, indicating a level of confidence of 90%. Wilk's Lambda test statistic is defined as:

$$\Lambda = \frac{|\text{Serror}|}{|\text{S}\_{effect} + \text{S}\_{error}|} \tag{3}$$

where

S*error* is the variation of the residual within the matrix of sum of squares cross product and

S*effect* is the variation of the treatment between the matrices of sum of squares cross product. Figure 1 below further illustrates the methodology that will be adopted in this research.

**Figure 1.** Methodology.

#### **4. Results**

A total of 39 responses were received from the online survey. The respondents' profiles showed that 15% of them worked in international offices, while 85% worked in local offices. Moreover, 13% of the respondents had more than 20 years of experience. Thus, a majority of the respondents were local contractors in the building industry with less than 5 years in business. In addition, it was observed that 56.4% of the respondents were involved in projects with an average project size of AED 50 million (considered large) or more, while 43.6% of respondents' average project size was less than AED 50 million (considered small). In this research, these two groups' (large and small) responses were further analyzed to gain important insights.

Descriptive analysis was conducted to compute the mean of the three populations in each variable (Table 4). The results show that the mean of frequency of using the three digital technologies (y1, y3, y5) in class 0 (pre-COVID) population is 3.29, while the mean of the level of investment (y2, y4, y6) in class 0 was 3.30. Similar calculations were performed for class 1 (during COVID) population, and the mean of frequency of usage and level of investment are 4.03 and 3.75, respectively, whereas the mean of frequency of usage in class 2 population (post-COVID) is 4.26, and the mean of the level of investment is 4.23. This indicates that there is in fact an increasing trend from pre-COVID to post-COVID in terms of mean frequency of usage and level of investment in digital technology for data gathering, processing, and communication. The results are summarized in Table 4.


**Table 4.** Means of the responses of the three classes.

In addition, the questionnaire contained three questions related to "project success". The questionnaire statements and the weighted average of responses, indicating a high degree of agreement, are listed below:


Thus, it is substantiated by the above findings that the digital technologies do indeed contribute to overall success of construction projects. Furthermore, the results presented in Table 4 also indicate that the frequency of using digital technology for communication

was at its peak (variable y5 with a value of 4.49) during COVID, as the demands imposed by the pandemic were higher than they were before the pandemic, indicated by a value of 3.21, and decreased somewhat from the peak to 4.28 in the post-COVID phase, as in-person communication is expected to return back somewhat but probably not at the pre-COVID level. This is because it is expected that the professionals in the industry will retain some of the communication technologies and conveniences, as they will be increasingly more familiar and comfortable with the technology. It was noticed that there was a slight drop in the use of communication technology from during COVID to post-COVID, apparently as some of the functions and operations are expected to return to "normalcy" after the pandemic. The results also indicate, predictably, that investment in communication technologies (variable y6) had increased during the pandemic, to 3.97 from 3.44 (pre-COVID), and is higher after the pandemic, at 4.26.

In an effort to compare between the responses based on the project size, two groups were created, as noted earlier: small projects with sizes less than 50 million AED and large projects with sizes 50 million AED or more. These two groups were almost equally represented in the survey.

In general, it should be noted that, while in class 0 (pre-COVID), both the average of the frequency of usage and level of investment were both greater in larger organizations than the smaller ones. The difference shrinks in both when compared to class 1 (during COVID) and class 2 (post-COVID), as the data in the last two rows of Table 5 indicate. This is a significant observation, as it substantiated two facts—(1) overall, there has been an increase in both the usage and investment in digital technologies in construction during pandemic, and the trend will continue after the pandemic; and (2) the extent of both usage and investment is significantly greater in smaller organizations than the larger ones at the during and post-COVID stages. This underscores the notion that benefits of digital technologies remained largely unknown or unexperienced by the smaller organizations, and once known they can become increasingly more acceptable in the industry.


**Table 5.** Means of the three class populations across small vs. large projects.

The results show that in class 2 (post-COVID), as illustrated in Table 5, the average of the frequency of usage and the level of investment were equal to each other in the large projects, but in small projects the average of the frequency of usage was higher than the

average of the level of investment. Furthermore, it can be seen that both the averages were greater in small projects than in large projects in class 2, and that could be due to the fact that large projects were ahead of small projects in terms of digital technology usage before the pandemic and, therefore, will not need to invest as much as the small projects to keep up with the demands created by the pandemic. Another observation that can be drawn is the fact that in class 1 the averages of the frequency of digital technology usage were almost the same in both small and large projects—in data gathering (variable y1, 3.69—small to 3.61 large) and communication (variable y5, 4.63—small to 4.40—large) technologies but not so different in processing technologies (variable y3, 3.69—small versus 4.13—large). Again, it is not surprising, as data/information processing technologies are more capital-intensive (high initial investment cost) compared to the other two—acquisition and communication technologies—and organizations dealing with smaller projects would not have investment capital available to them during the pandemic. This observation naturally leads to a closer look at what happens to expected investment responses after the pandemic. Results in Table 5 show that it is expected that organization with smaller projects would increase their investments significantly—in data acquisition (y2) from pre-COVID response of 3.00 to 4.38; in data processing (y4) from 3.00 to 4.18; and, lastly, in data communication (y6) from 3.25 from 4.25. The average of all three variables shows an increase in responses from 3.08 to 4.27. For organizations with larger projects, the extent of this difference in investment level between pre-COVID and post-COVID is much lower—from 3.45 to 4.23—although higher as predicted. These results are shown in Table 5.

To gain further insight, the responses from the survey were then analyzed using SAS statistical software as per the methodology explained in the earlier section. The MANOVA test was conducted using F approximation to test the hypothesis of no significant difference between the means of the three classes' populations (pre-, during, and post-COVID) on the frequency of usage and level of investment of the three digital technologies as shown in Tables 6 and 7, respectively. Table 6 shows that the *p*-value is less than alpha (0.1) for the null hypothesis of "no overall class difference"; therefore, it can be rejected, and it can be concluded with 90% confidence that there is in fact a significant difference between the means of the three class populations on the frequency of usage of the three digital technologies. Therefore, the analysis was taken a step further to conduct pairwise comparisons between the means of class 0 (pre-COVID) and class 1 (during COVID), class 0 (pre-COVID) and class 2 (post-COVID), and class 1 (during COVID) and 2 (post-COVID). The results have shown that in all three MANOVA tests, the *p*-values were less than alpha (0.1), which leads to the rejection of the null hypothesis and the conclusion with 90% confidence that there is a significant difference between the means of class 0 and class 1 and between the means of class 0 and 2 as well as between the means of class 1 and 2 on the frequency of usage of the three digital technology categories. Similar conclusions were drawn when MANOVA was conducted to compare between the means of the three class populations on the level of investment of the three digital technology categories, and the results are shown in Table 7.


**Table 6.** MANOVA Test: Wilk's Lambda for frequency of usage.

The first row in the tables above shows the value of Wilk's Lambda (refer to Equation (3)) in each MANOVA test. It measures how well each category of the independent variable (class) contributes to the model. The scale ranges from 0 to 1 as mentioned earlier, where 0 means total discrimination and 1 means no difference. Since all the Lambda values are less than 1 and are associated with a small *p*-value (significant at the 0.1 level), it can

be concluded with 90% confidence that there is in fact a difference between the means of the three classes on both the frequency of usage and level of investment of the three digital technology categories. The *p*-value represents the probability that measures the consistency between the data and the hypothesis being tested. The null hypothesis that there is no overall class effect and no difference between the classes in pairwise comparisons is evaluated with regard to this *p*-value. For a given alpha level, if the *p*-value is less than alpha then this null hypothesis is rejected. The alpha level used in this study is 0.1, and the tables show that all *p*-values are less than 0.1. It should be noted that some of the *p*-values are slightly higher for "no difference between classes 1 and 2" and "no difference between classes 0 and 1" in the level of investment, they are still less than 0.1. Therefore, it is safe to reject all the null hypotheses and conclude with 90% confidence that there is in fact a significant difference between the means of the three classes on both the frequency of usage and level of investment of the three digital technology categories. Future studies can focus on testing a larger sample size in order to achieve a smaller *p*-value and a higher confidence level.


**Table 7.** MANOVA Test: Wilk's Lambda for level of investment.

#### **5. Discussion**

The study on COVID-19 and its impact in the construction industry are fairly new. Nevertheless, the following paragraph compares the results of this study with recently published papers. The results of this study are in line with the conclusions of the research done by Cheshmehzang [56], who reported on the impact of COVID-19 on boosting digitalization in the built environment. The author stated that COVID-19 was a driver for the built environment to utilize available technological toolkits even further and formulate new policies on the use of digital technologies. Similarly, Ebekozien and Aigbavboa [57] have highlighted the role of the fourth industrial revolution technologies in curtailing the impacts of COVID-19 in Nigerian construction sites. The authors explained how the use of AI technologies such as RFID have compensated for the absence of workforces and the use of BIM has helped construction professionals to comply with the pandemic rules, such as physical distancing, and still be able to simulate construction site activities.

The study was conducted using a carefully crafted questionnaire survey in the UAE construction industry. The survey instrument (Appendix A) was designed to elicit responses from the professionals. Although it is common knowledge that digital technology contributes towards project success, the findings of this study confirmed the observation across all three categories of digital technology. Importance of data acquisition and processing technologies were rated at a higher level by the respondents than communication, indicating that in-person communication will continue to be used with less dependence on technology as compared with the other two. The study showed that smaller organizations will be using and investing in digital technologies at a greater extent than the larger ones.

One of the limitations of this study was the small size of data, as represented by only 39 respondents. It should be pointed out, however, that the respondents were drawn from a very specific population—the UAE construction professionals. Thus, the value of their responses was considered significant qualitatively for this study despite the small sample size. The second limitation was that the study was conducted in a specific country (UAE), and, thus, the results are difficult to generalize. While this is a shortcoming, the fact is that the UAE is a vibrant economy in the Middle East with a high projected growth rate in the construction sector and is considered a representative example in the global construction

industry. Nevertheless, the results and findings derived from this study conducted in the UAE provide valuable future directions and insights for further research on this topic.

The methodology developed and adopted in this study can be replicated in a larger country with a major economy and high volume of construction spending. A similar study can be undertaken with bigger sample sizes derived from different professional groups in the AEC industry. The variations and correlations between the regions and the professional groups will provide valuable insights regarding the use of digital technologies. Moreover, specific digital technologies can be investigated further, in addition to the three broad categories of digital technologies as conducted in this study. Nevertheless, this study provided a framework for future investigations on the use and investment in digital technologies in construction in the wake of catastrophic disasters, such as COVID-19.

The findings of this study can be useful from a practical standpoint by several measures. First, the study highlights the importance of digital technology in improving construction productivity. Second, this study differentiates between data acquisition, information processing, and communication technologies; this differentiation is helpful to identify and prioritize technologies for investing. Finally, and most importantly, this study underscores the importance of 'virtualization' in construction enabled by digital technology in the wake of catastrophes such as COVID-19. While the findings are general in nature in the context of the construction sector, the implications at the project level are imperative and indicative.

#### **6. Conclusions**

It is common knowledge that overall productivity increases in the construction sector, both regional and global, are minimal and significantly lower than the manufacturing and the service sectors. One reason frequently cited by the researchers and the practitioners alike is the slow adoption of technology, digital in particular, in construction. This study takes a deep look into this issue of adoption (usage and investment) of digital technology in construction in the context of the prevalent pandemic caused by COVID-19 since early 2020. COVID-19, ironically, presented an opportunity to investigate the status of use of digital technology adoption in the construction industry. Thus, this study is undertaken to determine the status in three levels, pre-COVID, during COVID, and post-COVID. The premise of the study is that digital technologies in three major categories—data acquisition, processing, and communication of data and information—are all impacted by the pandemic. There was a noticeable increase in the use of, and investment in, these digital technologies in the industry since early 2020 as a reaction to the restrictions put in place to reduce the spread of the COVID-19 virus. This crisis caused by the COVID-19 pandemic imposed a de facto mandate by instilling a sense of urgency among the construction professionals for digitalization of many processes and operations in construction and to perform them virtually. Digital technology, although not new and having existed for some time, was not utilized extensively prior to the "lockdown" situation caused by the pandemic. Oddly, COVID-19 provided the necessary impetus for digitalization in the construction industry. This crisis, as unexpected and undesirable as it is, offers a window of opportunity to improve and make the industry better positioned for the future.

This research study, despite the limitations, identified several significant facts and provided important insights. The most significant among them are that COVID-19 revealed that the use of digital technology, although remaining underutilized, is gaining wider acceptance in the industry. It also showed that the benefits of digital technology, once realized, will continue to be used. As a consequence, investment in digital technology in the construction industry will continue to increase. This will have long-term transformational and beneficial impacts on productivity in the construction sector. The methodology developed and employed in this study can be utilized to investigate certain selected technologies for use and investment decisions. A decision-making model for use in the industry can be developed using the categories as outlined in this study.

**Author Contributions:** Conceptualization, I.A. and S.E.-S.; methodology, O.E., S.A., I.A. and S.E.-S.; software, S.A.; validation, O.E., S.A., I.A. and S.E.-S.; formal analysis, O.E., S.A., I.A. and S.E.-S.; investigation, O.E., S.A., I.A. and S.E.-S.; resources, O.E., S.A., I.A. and S.E.-S.; data curation, O.E. and S.A.; writing—original draft preparation, O.E. and S.A.; writing—review and editing, I.A. and S.E.-S.; visualization, O.E. and S.A.; supervision, I.A. and S.E.-S.; project administration, I.A. and S.E.-S.; funding acquisition, I.A. and S.E.-S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by The American University of Sharjah, grant number FRG21-M-E71. The work in this paper was supported, in part, by the Open Access Program from the American University of Sharjah, and the APC was funded through Grant number: OAPCEN-1410-E00060.

**Institutional Review Board Statement:** The study was approved by the Institutional Review Board of the AMERICAN UNIVERSITY OF SHARJAH (protocol # 20-072 on 5 May 2021).

**Informed Consent Statement:** Written informed consent was waived for this study.

**Data Availability Statement:** Data are available upon request.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. This paper represents the opinions of the author(s) and does not mean to represent the position or opinions of the American University of Sharjah.

#### **Appendix A. Survey Instrument**

*Appendix A.1. General Information*

In this section, you will be asked about your general information.

Your company?


Years of experience in UAE


Project Type/expertise


Current role


Average size of projects


#### *Appendix A.2. Data and Information Gathering Technologies*

In this section, you will be asked to rate the level of usage, investment decisions, and the success potential for data and information gathering technologies (e.g., RFID, Barcode, Drones, GPS, laser scanning).

	- -Strongly Agree
	- -Agree
	- -Neutral
	- -Disagree
	- -Strongly Disagree
	- -Almost Always
	- -Often
	- -Sometimes
	- -Rarely
	- -Never
	- -Almost Always
	- -Often
	- -Sometimes
	- -Rarely
	- -Never
	- -Strongly Agree—Will greatly increase
	- -Agree—Will slightly increase
	- -Neutral—remain the same as already reached
	- -Disagree—Will slightly decrease
	- -Strongly Disagree—Will return to the pre-Pandemic level
	- -Very high
	- -High
	- -Normal
	- -Adequate
	- -Very little
	- -Very high
	- -High
	- -Normal
	- -Adequate
	- -Very little
	- -Strongly Agree—Will greatly increase
	- -Agree—Will slightly increase
	- -Neutral—remain the same as already reached
	- -Disagree—Will slightly decrease
	- -Strongly Disagree—Will return to the pre-Pandemic level
