Analysis of Factors Affecting Adoption of Volunteered Geographic Information in the Context of National Spatial Data Infrastructure

Spatial data infrastructures (SDIs) have been implemented for the last four decades in most countries. One of the key objectives of SDIs is to ensure the quick availability and accessibility of spatial data. The success of SDI depends on the underlying spatial datasets. Many developing countries such as Pakistan are facing problems in implementing SDI because of the unavailability of spatial data. Volunteered Geographic Information (VGI) is an alternate source for obtaining spatial data. Therefore, the question is what factors hamper the adoption of VGI for making it part of SDI in Pakistan. The intention behind this paper is to explore such factors as the key research question. To do so, we make use of the Technology–Organization–Environment (TOE) framework along with the partial least square structural equation model (PLS-SEM) to empirically analyze the factors impeding VGI from becoming part of SDI in the country. The study concludes that many technical, organizational, and environmental factors affect the adoption of VGI to be part of SDI in Pakistan.


Introduction
SDIs are the kind of infrastructure developed to enable coordination among producers and users regarding the creation, access, and use of geospatial data in an electronic environment [1]. VGI is generated by non-professional volunteers having little or no knowledge of spatial data collection. Although VGI is a potentially good source of free information, its quality cannot be defined easily [2].
Both SDI and VGI are introduced for the quick generation, availability, accessibility, and use of geospatial data. The growing costs of official mapping processes, including field surveys as well as ground verification, and the convenience of low-cost, easily generated, and up-to-date VGI, have placed VGI as a valuable geospatial data source for SDIs. VGI can be used in multiple ways, e.g., completing, improving, or updating the existing official mapping databases or collecting fresh data where no official data exist.
Pakistan is in the process of implementing the National Spatial Data Infrastructure (NSDI). In this regard, recently a Request for Proposal (RFP) has been advertised to seek consultancy firms for the execution of the feasibility study for the establishment of an NSDI for Pakistan. Pakistan is struggling through economic crises and implementation of the NSDI would require billions of rupees. Data comprise one of the core components of the NSDI. Successful implementation of the NSDI would require the availability of updated geospatial data from multiple organizations and disciplines to harness it for development, planning, policy-making, or research purposes in the governments, private sector, academia, and general public. VGI, being an attractive source of free, updated, and locally consistent geographical information, can be a potential data source of the NSDI for Pakistan.
The paper is organized as follows. Section 2 provides the background and importance of VGI for the NSDI. Section 3 describes the research model and hypotheses development. Materials and methods employed in this research are detailed in Section 4, whereas results are described in Section 5. Discussions are written in Section 6 and finally Section 7 outlines the conclusion and future work.

Theoretical Background
VGI platforms have been established for producing updated geospatial data. The data collected through VGI can be a vital source of information for enriching, updating, and mapping the missing areas in the context of NSDI developments especially in developing countries such as Pakistan. The synopsis of VGI in the context of the NSDI is presented below with the help of a literature review of spatial data, VGI, and the relationships of VGI with NSDI.

Spatial Data
Spatial data are a kind of data that associate objects to a location on the earth's surface and are also known as geographical, geospatial, location-based data, or geo-referenced data [3]. Almost all of the decision-making processes are centered around spatial data as a fundamental component. The availability and accessibility to spatial data have always been problematic as the acquisition of spatial data involves huge investments. Spatial data collected and maintained by government organizations are called authoritative data whereas spatial data collected by volunteers are known as VGI. Spatial data have great value for socio-economic development. As far as Pakistan is concerned, the instrumental value of the spatial data in the socio-economic development of the country has been realized moderately late as compared to the developed nations of the world [4]. Having the limited interaction of policymakers with geographic information systems, they had undervalued the socioeconomic significance of systematic spatial data collection, maintenance, and dissemination [5].

Spatial Data Infrastructure
SDIs are the kind of infrastructures developed to ensure quick availability and accessibility to spatial data held by various organizations. SDIs are being developed around the world for economic development and better governance [6]. SDIs are composed of five components, namely, people, policy, standards, data, and access mechanism, as defined by Rajabifard et al. [7]. This underpins the importance of spatial data as well as the people (user) for SDIs. Three generations of SDI development and implementation, namely, data-driven SDIs, process-driven SDIs, and user-driven SDI, are asserted by the scientific community [8]. The third generation being centered around the users highlights the increasing involvement of users in SDIs. The government of Pakistan (GoP) has been in the process of implementing spatial data infrastructure at the national level, i.e., national spatial data infrastructure (NSDI), since 2014 [9,10]. Despite the fact that the GoP is making efforts to implement the NSDI, no tangible progress has been achieved thus far, mainly due to the unavailability of authoritative spatial information. This scenario necessitates the need to look for other alternatives for obtaining spatial data, such as Volunteered Geographic Information (VGI).

Volunteered Geographic Information
Crowdsourcing geographic information is a type of geographical information collected by the mass crowd [11]. Crowdsourcing geographic information [12] contains both VGI [13] and involuntary geographic information [14]. VGI is a term that was first devised by [13]. VGI refers to geographical information generated with the help of volunteers equipped with GPS-enabled mobile devices. VGI initiatives emphasize gathering spatial data by the citizen. For example, an initiative such as OpenStreetMap (OSM) gathers the location of and information about different types of infrastructures, i.e., education [15], health [16], transportation [17], etc. Over the past 10 years, OSM has been used around the world to implement sustainable development goals by stitching data gaps [18]. On the other hand, initiatives such as Flickr and Picasa are centered around the collection of georeferenced images and videos.
VGI is being utilized in several countries. In the context of Pakistan, a few studies have been conducted, such as on the use of VGI in educational planning [19], in the adaptation to climate-change effects [20], and in promoting heritage tourism [21]. However, the importance of VGI has not been comprehended by the Government of Pakistan (GoP) yet, due to various reasons which will be assessed, being the objective of this study in the next sections.

SDI and VGI
Since the introduction of the VGI concept, its relationship with SDI remains on the research agenda in many parts of the world. There are similarities and differences existing between the two concepts regarding data, as discussed in [22]. The combination of VGI and SDI will improve what is available to the end-users and be convenient for the decisionmaking process. However, certain challenges need to be considered to devise a suitable solution for these challenges. Different authors discussed the integration of SDI and VGI in different prospects while mentioning the advantages and disadvantages of VGI in the context of SDI.
The authors in [23] discussed the integration of VGI and SDI in the context of developed and underdeveloped countries by mentioning some advantages and disadvantages of VGI. The authors asserted that the quality of VGI needs to be addressed by developing specialized tools before integrating them into SDI. The advantage of VGI being the datagap-filler tool where the data coverage is not uniform in SDI is also detailed in [23]. An SDI framework based on OGC standards is proposed in [24] to create and manage interoperable VGI using SDI. To use VGI as an additional source of spatial data in SDI, the concept of Volunteered Spatial Data Infrastructure (VSDI) is proposed by the authors in [25]. The VSDI implementation has been demonstrated in the field of transportation planning in a collaborative environment to disseminate real-time update information to decision makers to make better decisions. The results of the study favor the use of VGI to provide updated information for transportation planning.
Availability of timely information in disaster management is still a challenge to planners, decision makers, and rescue authorities. VGI as updated geographical information of the area is rendered as one of the solutions to timely feed data in disaster-management systems [26][27][28][29][30][31]. A framework has been proposed by [32] to effectively use the value of VGI for disaster management to support the authoritative data in the context of SDI. The authors endorsed some issues in the use of VGI by the government, such as privacy, copyright, and data ownership. The authors suggested that the government needs to address these issues by re-evaluating the relevant policies so that the integration of VGI with authoritative data could be possible. Similarly, the authors of [33] discussed the use of VGI in SDI in the context of disaster management. However, the authors suggested improving the data quality of VGI by semantically processing the locational details of VGI along with available spatial data of that geographical area. The authors of [34] discussed the development of a new framework to integrate SDI and VGI in Indonesia to better respond to disaster-management activities.
Land administration in developing countries is still a dilemma to better utilize land resources for economic development. The use of VGI in land administration has been proposed by many authors [35][36][37]. The quality of VGI hinders its utilization for official purposes such as in land administration. To minimize the ground-truthing of VGI to assess its quality, a Kappa methodology-based solution has been proposed in [38] by offering a proxy quality of VGI through agreement of consensus. Similarly, other proxy measures of VGI quality using trust and reputation modeling have been proposed by [39]. To efficiently and quickly update China's national spatial data infrastructure, a crowdsourcing-based solution to collect and update geospatial data with the help of the general public is proposed. The prototype system is developed to demonstrate the solution with the help of crowdsourced data collection of POIs, roads, and residential areas [40].
Different national SDI also started to adopt VGI to support their official mapping operations. INSPIRE is an SDI implementation in most European countries. Many studies have suggested and explored ways to induct the value of VGI into the INSPIRE. A transformation of current SDIs into future evolutionary models as data spaces is presented in [41]. The architecture of a data space contains many technology artifacts of the modern era including crowdsource information. The concept of data spaces is built on top of the INSPIRE to support the use of many data generators such as VGI, sensors, etc.
The discussion in the above paragraphs concludes that most of the studies on the integration of VGI in SDI are centered around one theme such as disaster management, land administration, transportation, etc. However, a general discussion on the integration of VGI and SDI with the description of the advantages and disadvantages of the integration is also discussed.

Research Model
The TOE framework was presented by [42]. The TOE framework provides the theoretical base, understanding, and empirical support for research into the use of technology and information systems in the workplace [43]. TOE divides the technology, organization, and environment as the three classes of factors to study organizational perspective on adopting technology [44,45]. The TOE framework has three contexts, i.e., organizational, technological, and environmental; however, for this study, socioeconomic context has been added to the model to better understand the adoption of VGI in the public sector within the context of SDI.

Reasons behind the Selection of the TOE Framework
To investigate the adoption of new technologies within the organization, there are many theories in practice [46], for example, Diffusion of Innovation (DoI) [47] and Institutional Theory [43]. However, the TOE framework has some advantages that are listed below:

•
Most of the adaptation theories available that deal with information technology and communication aspects emerge from the TOE framework. These theories either shorten or extend the contexts of the TOE framework. For example, the Institutional Theory [43] divides TOE aspects and deals mainly with the environmental context of the TOE framework. In the same pattern, the DoI theory [47] describes the technology and organizational context of the TOE framework.

•
As far as the adoption of new technology within the organization is concerned, TOE is the most utilized framework to test such a kind of adoption [46].
Based on the facts of the TOE framework mentioned above, a lot of research studies employed the TOE framework to examine the adoption of new technologies within the organization, such as cloud computing [48,49], blockchain [50], e-commerce [51], mobile apps [15,41,52,53], social media [54], etc. Hence, the selection of the TOE framework to test the adoption of VGI is justified.

Hypothesis Development
The research model in Figure 1 explores the links between technology, organization, environment, and socioeconomic aspects in the public sector in terms of VGI readiness and suitability for SDI.

Technology Context
The technological landscape is changing quickly. From a technological standpoint, the organization should analyze both the technologies in the market and the technologies currently employed in the operations, to effectively adopt new technology. The current system will be one of the most significant factors to examine when evaluating the adoption of the newest technologies. The technical context encompasses both internal and external technologies pertinent to the organization, encompassing both present procedures and equipment as well as available technologies outside of the enterprise [43]. The considered factors of the technology context are compatibility, complexity, data quality, and technological competence [55][56][57]. Complexity refers to the level of difficulty in executing the innovation within the organization [46]. For the current study, complexity means the level of difficulty of using VGI in the public sector in the context of the NSDI. In the public sector, generally, data from conventional sources are used to craft the mapping products. VGI is a new kind of data source and comes from multiple platforms, and the choice to choose from, searching the relevant data, and handling these sorts of data is challenging for conventional users. Thus, it is good to gauge complexity in the context of the adoption of VGI. Compatibility is a match of innovation with existing rules, prior experiences, and contemporary expectations of the organization [46]. Compatibility means the congruity of VGI with the past, current, and future mapping rules, standards, and practices within the stakeholders of the NSDI. Data interoperability is a challenging task for data managers and data administrators. Fusion of a new type of dataset with the current database schemas, existing datasets, and data formats must be considered before the inclusion and use of VGI datasets. Hence, the it is essential to test the compatibility factor regarding the adoption of VGI in the NSDI. Data quality is characterized as "the measure of the agreement between the data views presented by an information system and that same data in the real world" [58,59]. Data quality has multifaceted aspects [60] such as accuracy, completeness, consistency, and currency [61,62]. In the VGI context, data quality is described as the quality and accuracy of the VGI dataset. VGI data are generally considered as of low quality as compared to the authoritative dataset. Moreover, VGI has limited metadata and the VGI creation process follows minimal data and metadata standards. Thus, the consideration of data quality factors in the context of VGI usage within the stakeholders of the NSDI is compelling. Technological Competence is illustrated as an organization's ability to increase its throughput [63,64]. Technological competence means the capacity of the stakeholders of the NSDI to handle the utility of VGI datasets within the organization workflow in terms of technical infrastructure. Technological competence involves handling VGI datasets in the GIS analysis, accessibility of online VGI platforms, processing of VGI datasets, etc. Thus, it is imperative to consider the factor of technological competence in judging the adoption of VGI within the stakeholders of the NSDI. Complexity, compatibility, data quality, and technology competence variables of the technological context are hypothesized to have influences on the suitability of adoption of VGI for the public sector in the context of SDI.

Organization Context
Organizational context means the internal features of the organization [65]. The organization must analyze its resources to adopt and support the technology [52,66]. The organizational context consists of components that enable technology adoption within the organization. Organization factors considered for this study are resources, collaboration, reliability, and management support [55,67]. Management support is defined as the support from executive bodies or management to absorb the new technologies within the organization [68,69]. Management support means an endorsement from top management of the use of VGI within the organization. The support from top-level management in prioritizing and facilitating the VGI consumption within the organization reduces the chance of failures in the processes of planning and implementation of VGI use. Hence, management support is also an important factor to be considered to inspect the adoption of the VGI dataset. Collaboration refers to practices of facilitating and functioning in multi-organizational setups to address issues that are too big for a single organization to manage [70]. Collaboration in the context of the current study refers to the collaboration of stakeholders of the NSDI to benefit VGI. Collaboration among stakeholders to encourage, grow, and share experiences as well as skill sets in the production and processing of VGI is good to empower the NSDI implementation. Thus, inclusion of the factor of collaboration seems vital to test. In this study, reliability refers to the reliability of VGI data, i.e., that the VGI data are error-free and consistent. The utilization of VGI data in official work remains in question as it is produced by volunteers, in comparison with data produced by professionals of the public sector organizations. The reliability of VGI is an important factor to be considered in the adoption of VGI in the NSDI. Resources refer to an organization's ability to meet future demands and adopt to changing circumstances. The resources are concerned with the organization's infrastructure and how effectively that infrastructure can support the adoption of new technologies [71]. Additional resources may be required to process and consume the VGI dataset within the existing workflow of the stakeholders. The inclusion of resources factor is helpful to test the adoption of VGI in the NSDI context. Management support, reliability, resources, and collaboration, which characterize the organizational context are hypothesized to have influences on the suitability and readiness of adoption of the VGI.
Hypothesis (H5). Management support impacts the suitability of VGI.

Environment Context
Environmental context is related to the impact of the external and internal factors of the environment in which an organization works. Environmental factors such as government funding and competitor rivalry influence an organization's adoption of new technologies [65]. Facilitation conditions are also another environmental factor that impacts the adoption of new technologies [72]. The environmental context entails the working environment of the organization. Environmental factors considered for this study are regulatory policy, legal liability, and consumer perception [55]. Regulatory policies are framed by a government to regulate the use of some technologies within the country [73]. Therefore, regulatory policy is an indispensable aspect in deciding the use of technology within the organization under the laws and policies of governments. To what extent are the regulatory policies of the country helpful in the use of VGI and the inclusion of VGI data in the official working environment in the context of the NSDI? Is there any new policy required to regulate the use of VGI? These are the questions that trigger the consideration of the factor of regulatory policy to be tested regarding the adoption of VGI in the official work of stakeholders of the NSDI. Legal liability is described as the responsibility that somebody has for their activities [74]. For the current study, legal liability refers to responsibility for any incorrectness and inaccuracies in the VGI data as VGI creations process generality involves virtual environment and actual identity, as well as sources of data, may be unknown and anonymous in some platforms of VGI. Responsibility for incorrect VGI data may not be fixed and intellectual property and licensing issues are also associated with VGI. Therefore, this factor of legal liability is also equally important to test whether or not legal liability of VGI is a barrier for its adoption in the NSDI. Consumer perception means the response of the end-users towards new technologies [75]. For the current study, consumer perception refers to the response of the users of the data or products produced by the NSDI stakeholders with the help of VGI data. A consumer may argue that VGI data are produced by amateurs, so it is not valuable and thus should not be mixed with official data. Therefore, the inclusion of this aspect of consumer perception is justified to test the supposition of consumer perception in the context of VGI adoption. Legal liability, consumer perception, and regulatory policy factors of the environmental context of the model are hypothesized to have influence on the suitability of adoption of VGI in public sector organizations.

Socioeconomic Context
Community participation, job creation, and expected profitability are the considered factors for this study regarding the socioeconomic context. Community participation is delineated as the involvement of people in a joint action to solve a common problem [76]. In this study, community participation means the participation of citizens to collect and manage spatial data on different VGI platforms. Community support in VGI creation is equally important as the motivation of the contributors of VGI. To make the VGI project sustainable, it is challenging to continuously involve the citizen in the VGI creation process. Hence, the importance of community participation aspects is vital to test the adoption of VGI. For this study, expected profitability means the degree to which an organization feels a technology would be beneficial for the organization [47]. It may refer to the additional amount that can be generated through the adoption of VGI. It is imperative to consider this factor to judge the return on investments that can be consumed in the process of adoption of VGI. Job creation in this study refers to opportunities created for new jobs as a result of the adoption of VGI in the NSDI. It is argued that new jobs of data processing, quality control, and analysis would be generated as a result of the inclusion of VG in the official working environment. However, there are chances that jobs may be reduced. Thus, the factor of job creation is also essential to be considered. Job creation, community participation, and expected profitability variables of the socioeconomic context are hypothesized to have impacts on the readiness of the adoption of VGI in the public sector organizations in the context of SDI.

Suitability
Suitability can be characterized as a concept that stimulates organizations to implement new technologies. When work is very suitable, organizations prefer to incorporate new technologies and services [77][78][79]. The common observation is that VGI is not fit for official work. There is a need to advocate for top management and policymakers to use the VGI. In the current study, suitability means the appropriateness of VGI to be adopted for the official working environment of the different stakeholders of the NSDI to use VGI. Accordingly, the following hypothesis is established as a result of this research.

Readiness
Readiness means the consumers' proclivity to use emerging technology to accomplish their objectives [80]. Thus, the implementation of new services and technologies by an organization depends on the readiness of that organization for that technology. Readiness also refers to the readiness of human and financial resources for the adoption of new technology within the organization [80]. Bigger organizations are more likely to adopt new technologies due to the existence of required resources within the organization as compared to smaller organizations [81]. Ifinedo [82] argued that organizations with trained human resources are more readily accepting of new technologies. In the current study, readiness means the willingness and existence of human and financial resources to use VGI in the official working environment. Considering the above, this study establishes the following hypothesis.

Materials and Methods
The path model of Partial Least Square-Structural Equation Modeling (PLS-SEM) is applied to verify the correlation of the model proposed for this research.

PLS-SEM
The Partial Least Square (PLS) technique is applied to test the proposed model and hypotheses. For quantitative data analysis, many analysis techniques exist such as correlation, regression, analysis of variance, etc., besides the PLS-SEM technique [83]. However, PLS-SEM has certain advantages, mentioned below: • As compared to other Structural Equation Modeling (SEM) methods such as (LIS-REL) [84] and EQS [85] that use a maximum likelihood estimation function, the PLS-SEM method uses a least-squares estimation procedure [86]. • PLS-SEM is quite useful to simultaneously evaluate the factors of the measurement model and structural model [86]. • PLS evades many restraining assumptions such as multivariate normality and large sample size [87,88].

•
As the size of the sample is small, the PLS-SEM method is used because it can perform well even when the sample size is small [87]. • PLS-SEM helps to investigate the relationship of observed and unobserved variables, while other above-mentioned methods work on observed variables only [89]. • Moreover, the effects of moderators can be directly observed with the help of the PLS-SEM technique [83].
Based on the facts, the PLS-SEM model was used in this study.

Sample and Measures
A survey questionnaire (see Appendix A) was developed based on the questions formulated using the established hypothesis. The research model was tested using the data collected through a questionnaire survey. There were various aspects considered while designing the questionnaire used in this study.

•
We have used already-established factors such as complexity, compatibility, data quality, technological competence, collaboration, management support, reliability, resources, regulatory policy, legal liability, consumer perception, community participation, suitability, and readiness from the available research studies. The factors are fine-tuned to meet the requirement of this study. The factors of expected profitability and job creation are self-developed for this study. Table 1 delivers details about the abbreviations for latent variables and a description as well as code of item for each factor. • We keep the questions simple and avoided the confusing and more technical terminologies in establishing the questionnaire. • Questions were developed by getting feedback from the researchers working on the concepts of the NSDI and VGI. • Questions were proofread by language experts to avoid any grammar mistakes.
Generally, the Likert scale has been used in the research in two formats: five-point scale and seven-point scale. However, the accuracy of research output is enhanced by employing a seven-point scale as this scale version provides more options to respondents of the questionnaire as compared to the five-point scale version [90,91]. Due to the more flexible nature of the seven-point scale, we found it more suitable for our study. In the questionnaire, all questions are evaluated against a seven-point Likert scale, where 1 = strongly disagree and 7 = strongly agree. The questionnaire was circulated among 150 users of data in the geospatial sector of Pakistan. The users were selected based on their understanding of GIS data. From the collected questionnaires, 141 valid responses were received. An 82% proportion were male, whereas 18% were female. Among users, 57% were from the public sector, 11% were from the private sector, and the rest (32%) belonged to academia. Overall, 28% were postgraduate, 48% graduate, and the remaining 24% were undergraduate. As far as map usage is concerned, 37% used maps frequently, while 49% used maps often. Table 2 reviews the descriptive statistic of the respondents. The sample size utilized in this research is as per the guidelines of the 10-time rule and guidelines endorsed by Hair Jr. et al. [102]. The 10-time rule is a popular and frequently used method to gauge the sample size to be used in the PLS-SEM method. The 10-time rule narrates that "the sample size should be greater than 10 times the maximum number of inner or outer model links pointing in the model" [86]. As per guidelines endorsed by Hair Jr. et al. [102], The sample size for the PLS-SEM strategy is determined by the wellknown 10-times rule, which specifies that the sample size should be equal to the larger of: (a) 10 times the largest number of formative indicators utilized to assess a single construct, or (b) 10 times the largest number of structural paths aimed at a specific construct in the structural model. There are fourteen formative indicators in all, and thus, the minimum sample size should be 140. Therefore, 141 valid responses are sufficient.

Factor Analysis
To uncover patterns in a set of variables, factor analysis employs mathematical approaches for the simplification of correlated measurements [103]. There are two types of factor analysis: Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) [104]. To test the conceptual model proposed in the study and alignment of questionnaire items, external factor loading was used to perform EFA. According to the evaluation criteria in [105], each indicator's factor loading should be greater than 0.6-0.7 and higher than the factor loading of other indicators. The factor loading findings are reported in Table 3. In this study, the values of factor loading (in bold) meet the criteria mentioned above.

Reliability Analysis
The dependability of each variable of the external model, which is the measurement model, is determined via reliability analysis [106]. Internal consistency is checked by evaluating reliability. Different parameters are used such as Cronbach's α value, DG's rho value (composite reliability), and the eigenvalue [105]. The threshold value of Cronbach's α value is 0.7 or higher [105,107]. Similarly, the threshold for DG's rho value is 0.7 or higher [105]. Table 4 describes the result of the reliability analysis. In this study, the values of Cronbach's α and DG's rho meet the criteria mentioned above.

Validity Analysis
A validity analysis is generally of two types, i.e., convergent validity and discriminant validity [106]. The convergent validity is analyzed with the help of the average variance extracted (AVE) value [108] and, to secure the convergent validity, the AVE should be higher than 0.5 [106], whereas the discriminant validity is evaluated with the help of the square root of the AVE values. The square root of the AVE values of latent variables must be more than or equal to the correlation value of the latent variable and other latent variables [86,108]. Table 5 describes the AVE values of this research, whereas Table 6 describes the square root of the AVE values. In this study, values of AVE and the square root of AVE satisfy the criteria and secure the validity.  Table 6. Discriminant validity analysis.

Hypothesis Testing
To produce statistically significant path coefficients from a path analysis, the PLS structural equation employs a nonparametric test methodology with the bootstrapping method [106]. In this research, 13 out 17 hypotheses were adopted and four were rejected, as described in Table 7. The first result presented is based on the hypothesis on the relationship between technical characteristics and suitability. Hypothesis 1, which narrates that complexity impacts suitability, was accepted, with a p-value of 0.029 and path coefficient of 0.113. Hypothesis 2, which describes that compatibility impacts suitability, was accepted, with a p-value of 0.008 and path coefficient of 0.143. Hypothesis 3, which communicates that data quality impacts suitability, was rejected, with a p-value of 0.137 and a path coefficient of 0.066. Hypothesis 4, which expresses technical competence impacts suitability, was accepted, with a p-value of 0.03 and path coefficient of 0.112.
Secondly, the outcome of the hypotheses on the relationship between organizational characteristics and suitability is noted. Hypothesis 5, which asserts that management support impacts suitability, was accepted, with a p-value of 0.027 and path coefficient of 0.115. Hypothesis 6, which communicates that reliability impacts suitability, was accepted, with a p-value of 0.031 and path coefficient of 0.111. Hypothesis 7, which states that collaboration impacts suitability, was rejected, with a p-value of 0.068 and path coefficient of 0.089.
The results of the hypotheses regarding the links between organizational characteristics and readiness are described next. Hypothesis 8, which declares that collaboration impacts readiness, was accepted, with a p-value of 0.019 and a path coefficient of 0.124. Hypothesis 9, which narrates that resources impact readiness, was accepted, with a p-value of < 0.001 and path coefficient of 0.327.
Next, the outcome of the hypotheses on the correspondence between environmental characteristics and suitability is communicated. Hypothesis 10, which describes that regulatory policy impacts suitability, was accepted, with a p-value of 0.016 and path coefficient of 0.128. Hypothesis 11, which posits that legal liability impacts suitability, was rejected, with a p-value of 0.098 and path coefficient of 0.077. Hypothesis 12, which asserts that consumer perception impacts readiness, was accepted with the p-value of 0.043 and path coefficient of -0.103.
The outcome of the hypotheses on the connection between socioeconomic characteristics and readiness is asserted next. Hypothesis 13, which communicates that job creation impacts readiness, was accepted, with a p-value of 0.035 and a path coefficient of 0.108. Hypothesis 14, which expresses that community participation impacts readiness, was rejected, with a p-value of 0.074 and a path coefficient of -0.087. Hypothesis 15, which states that expected profitability impacts readiness, was accepted, with a p-value of 0.037 and path coefficient of -0.107.
Lastly, the results of the hypotheses on the association between suitability, readiness, and intention to adopt VGI are noted. Hypothesis 16, which states that suitability impacts adoption, was accepted, with a p-value of <0.001 and path coefficient of 0.2. Hypothesis 17, which states that readiness impacts adoption, was accepted, with a p-value of <0.001 and path coefficient of 0.655.

Discussion
This study empirically investigated factors that affect the adoption of VGI by the public sector in Pakistan. We applied the TOE framework along with the partial least square structural equation model (PLS-SEM) to empirically analyze the factors impeding VGI from being part of SDI in the country. To examine causal linear and non-linear relationships with latent constructs, the applied PLS-SEM blends process-driven and evidence-based techniques. The PLS-SEM model can reveal factors that may not have a direct impact on VGI adoption but do have an indirect impact on other factors. For example, Hypothesis 13, which states that job creation impacts readiness, was accepted with a p-value of 0.035 and path coefficient of 0.108.
Based on the results, we found that different technological, organizational, environmental, and socioeconomic factors influence the adoption of VGI by the public sector in the NSDI context. A survey questionnaire comprising 56 questions was developed and used to collect data from the respondents to test the research model. The questionnaire was based on the questions formulated using the established seventeen hypotheses, i.e., H1 to H17. All questions were evaluated against a seven-point Likert scale. The scale provides two moderate opinions along with two extremes, two intermediate, and one neutral opinion to the respondents. Out of seventeen hypotheses, four, i.e., Hypothesis (H3) regarding the impact of data quality on the suitability of VGI, Hypothesis (H7) related to the impact of collaboration on the suitability of VGI, Hypothesis (H11) dealing with the impact of legal liability on the suitability of VGI, and Hypothesis (H14) positing that community participation impacts the readiness of VGI, could not be validated and therefore were rejected. The outcomes of the study extend the TOE framework by adding new socioeconomic dimensions and the inclusion of new factors such as job creation and expected profitability. In the context of the TOE framework, the results of the study, their explanations, and their implications are presented below.

Technology Context
We find that complexity, technical competence, and compatibility in the technical characteristics pointedly impact the suitability of VGI for SDI. However, the VGI data quality does not significantly affect its adoption by the users, as found by the data collected in the context of Pakistan, although the effect of VGI data quality on its adoption in the public sector may differ from region to region. This result may be influenced by the fact that VGI usage in Pakistan is quite low in the public sector. The adoption of VGI data in some public sector projects may be supported by better knowledge of the quality of VGI, since the notion of unreliability may hamper its adoption. It is suggested that organizations willing to adopt VGI should improve and construct VGI supportive data and metadata standards. The findings highlight that compatibility influences the suitability of VGI. Attributes in VGI are characterized through the key/value pairs also called "tags". However, the semantics of VGI tags is ambiguous in some cases [109]. Moreover, different names of tags describe similar meanings, and the same name of the tag designate different meanings [109]. VGI data can be generated through different techniques and technologies. Data generated from different VGI platforms come in structured, unstructured, and semi-structured formats. Moreover, culture, geography, and education level may also be different for different VGI contributions [110]. Thus, a careful understanding of the existing database schema and interpretability of the VGI with another GIS format within the organization is needed before deciding the addition of VGI in the official working environment, as the quality of VGI data improves with the perfection in practical understanding and incremental enhancement to the VGI data collection [111]. This study found that the complexity of VGI influences the organization's conclusion to adopt VGI, although, considerable success has been noticed in the use of VGI in governments [112]. However, in cases such as Pakistan [19] where VGI is not being used regularly in the government sector, the development and maintenance of the tools required to collect, analyze, integrate, and validate VGI need special consideration to avoid complexities in the VGI data-collection implementation phase. The study showed that technical competence impacts the suitability and adoption of VGI. It is suggested that stakeholders need to remove their deficiencies in technical competence and employ best practices. Technical competence enhances confidence among the stakeholders of the NSDI, and resultantly, stakeholders would facilitate the adoption of VGI.

Organizational Context
Organizational factors such as reliability and management support pointedly impact the suitability of VGI for SDI. Collaboration and resources factors have a significant impact on readiness. However, the study finds no impact of collaboration on suitability. Thus, we can say all factors of the organization influence the adoption of VGI in the context of Pakistan. Resources are an essential part of the pre-requirements for the use of VGI. Retention of well-trained manpower capable of handling the VGI issues involved in the adoption in the public sector is a challenge [113]. The shortage of skills and resources can hamper the initiation of VGI adoption [111]. It is recommended that due resources should be arranged before the adoption of VGI so that the initiation and implementation process may run smoothly, as unavailability of due resources significantly affects the readiness of VGI to be part of SDI. As far as top management is concerned, our study finds that management support impacts the adoption of VGI. Organization priorities may not help the adoption of VGI if support from top management is not found [95]. This is because top management regulates the adoption of new technologies and allocates due resources within the organization. It is suggested that top management should be wellbriefed about the value of VGI and its potential for NSDI so that their support could be channelized to set the priorities for the adoption of VGI in the NSDI initiative. The results of this study reveal that the reliability of VGI has a profound effect on its suitability. The mindset of organizations towards the reliability of VGI hampers it from being part of SDI in the country. However, Zarin Khan and Peter Johnson [95] argued that "reliability of VGI can be increased as organizations gain experience collecting and analyzing citizen data". It is suggested that more and more practices and experiences should be gained and shared to increase the trust and reliability of VGI among organizational top management. Moreover, new or modified versions of relevant guidelines, policies, and regulations need to be developed to improve the reliability of VGI data, as asserted in [97,114]. We find that collaboration does not significantly affect VGI suitability but has a significant impact on readiness. As readiness impacts the adoption of VGI, and thus collaboration, has some significance for the adoption of VGI in the government sector in the context of the NSDI. Collaboration among the organizations helps to share and enhance experience and skills [96,97]. Collaboration provides expert guidelines that resultantly lead to a clear direction for the adoption of VGI as well as to minimizing the time consumed in government approvals and permissions [95]. It is recommended that, for good initiation of the project related to the adoption of VGI in the government sector, collaboration among relevant stakeholders should be established to grow experience and skills within government organizations and motivate other organizations too.

Environmental Context
Environmental factors such as regulatory policy and consumer perception pointedly impact the suitability of VGI for SDI, whereas legal liability does not influence the suitability of VGI. Existing regulatory information policies do not apply to the current rise of the adoption of social media by the government [115]. The same is true for VGI adoption by the government. Thus, it is suggested that existing policies need to improve and enhance to make these policies more flexible towards the adoption of VGI. However, such policies should not be restricted to a specific technology because new technologies may emerge in the future and existing policies can remain relevant to the new technologies [115]. The results of this study reveal that consumer perception has a profound effect on the adoption of VGI. Data products may be produced by integrating VGI into the official dataset. Consumer perception of data quality of such products remains doubtful due to the inclusion of VGI as consumers think such products are produced by non-professionals. Thus, it is recommended that quality aspects of VGI should be included with the data products to build the trust of the consumer in products created with the help of VGI data. We gather that legal liability does not affect VGI adoption. However, VGI is a comparatively new technology for many government organizations, hence, the inclusion of VGI in official working may raise questions on the liability of VGI data [98,116]. Hence, it is suggested that the initiation and implementation of the adoption of VGI should be handled carefully so that all stakeholders benefit [97,117]. Moreover, the platform developed for the collection of VGI needs to have fully authenticated contributors only, so that anonymous contributions be discouraged. Intellectual property rights associated with VGI contributors and licensing issues associated with freely available VGI contents should be understood well before the adoption of VGI to avoid any liability concerns [117].

Socioeconomic Context
The socioeconomic characteristics of job creation and expected profitability have a significant effect, whereas community participation does not influence the readiness of VGI. The creation of new jobs in government sector in developing countries generally does not suit the government due to economic crises. However, the inclusion of new technology within the organization always provides an opportunity to offer some jobs to grow the new technology adoption. The adoption of VGI may create new jobs in data integration, data analysis, data management, and web development. Thus, it is suggested that job aspects should be considered by the organization before the adoption of VGI. This study discloses that the expected profitability has a profound effect on the adoption of VGI. A lot of investment is required at the initiation of VGI projects. However, expected profits cannot be calculated without the successful implementation of the VGI projects. Thus, it is recommended that efforts should be made for the successful implementation of VGI projects. We find that community participation has no significant impact on VGI adoption. However, motivation of the contributors is required so that contributors may complete the tasks to make fruitful contributions. Hence, it is suggested that direct liaisons with the contributor should be established to sustain VGI projects administrated by the government.

Implications
Some key theoretical and practical implications of the study are presented below.

Theoretical Implications
First, based on our best knowledge, this is the first study carried out about the factors influencing organizational adoption of VGI within NSDI which employed a statistical model to determine the factors but also statistically gauged the level and extent of those factors, although there are studies that discussed the adoption of VGI in NSDI concerning technical aspects. However, previous studies descriptively identified the factors impeding VGI from being part of SDI [97,98,111,113]. They did not use any statistical model to find and validate the factors. Second, this study tested the extended version of the TOE framework by the addition of extra socioeconomic context. Third, our study explored the factors that affect the adoption of VGI in NSDI for developing countries, i.e., Pakistan. The fourth implication is that our study generally enriched the existing literature on different factors affecting the adoption of VGI. Fifth, our research described a validated research model for the organizational adoption of VGI in NSDI. Finally, as an outcome of the study, we have established a measuring scale for the job creation and expected profitability factors, which would support future research on VGI and NSDI.

Practical Implications
First, our research would offer guidelines to the policymakers employed on NSDI for Pakistan to develop policies to encourage VGI adoption among stakeholders of the NSDI. Second, our research is the first kind of study in the aspect of organizational adoption of VGI in NSDI, and there are a lot of opportunities for academia to conduct further research on different dimensions and inclusion of VGI and SDI as a subject in their academic programs. Third, based on the study, data-production organizations can develop flexible models or modify existing models to include the VGI in their working environment. Fourth, we have found that organizations feel reluctant to adopt VGI due to the lack of established standards of VGI. Therefore, this finding is important for governments and private regulatory bodies to develop relevant standards to remove the uncertainties hindering VGI adoption. Finally, the study highlights the role of an organization's top management in VGI adoption. Therefore, the organization's top management should be determined and focused on the adoption of new technology. Top management's clarity towards value creation is conducive to achieving successful adoption of VGI in the organization.

Limitations
However, this study has some limitations that can be the focus of future research too. First, as this research was focused on organizational adoption of VGI, a similar kind of study may be conducted to test the individual perception of the adoption of VGI. The concept of VGI in governments is not matured, but perhaps at the individual level the results may be similar or different from the organization adoption. Secondly, mediating factors were not used to test the relationship of various factors with suitability and readiness. Lastly, limited dimensions were tested for technology, environment, organization, and socioeconomic context, testing of more dimensions of these constructs can produce more interesting facts.

Conclusions
This study applied a statistical model to test multiple hypotheses related to the adoption of VGI in the local context of Pakistan. We validated 17 hypotheses to examine impacts on a range of outcomes including data quality (QTY), collaboration (COL), and reliability (REL). The study has attempted to determine the factors impeding the adoption of VGI by applying the TOE framework along with the partial least square structural equation model (PLS-SEM) to empirically analyze the factors impeding VGI from being part of SDI in the country, being a significant contribution to VGI research literature.

Future Directions
The objective of the study has been achieved, however, to improve its application and validity, more empirical research is needed. For example, this study was carried out in the local context of Pakistan; its validity can be enhanced if a similar kind of study is conducted for other countries, which have different policies, regulations, cultures, and technology setup. Secondly, more studies can be taken up to investigate the factors of VGI adoption by integrating TOE with other frameworks such as the Technology Adaptation Model or even using another framework autonomously. Thirdly, the adoption of VGI coupled with the partial least square structural equation model (PLS-SEM) for use-cases such as land management and agriculture are also recommended for future work. Moreover, future study is required to develop a model for a similar construct that can be sustained over the period. Data Availability Statement: Data sharing is not applicable to this article.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A. Survey Questionnaire
Please reply to each question by using the scale presented below. 1 = Strongly Disagree 2 = Disagree 3 = Somewhat Disagree 4 = Neutral 5 = Somewhat Agree 6 = Agree 7 = Strongly Agree