*3.1. Variables*

As a result of the literature review, for developing a model for the comparative analysis of public and private development projects, the accessibility of the project site, distance, size, and development density were identified as major determinants. The variables also reflect the circumstances specific to South Korea, where the distance within a 5-km radius [11] from the project site and the number of facilities within this range were added as additional factors. This is because the most decisive factor in the site selection of each project entity is the demand of users for the facility after actual development. Therefore, the psychological distance is set according to the sphere of influence to ensure the validity of the comparative analysis. In sum, the variables used in previous studies were organised to accommodate our purpose (Table 3) [9,11,17–20,28,29].


**Table 3.** Variable selection.

Concerning project size, the public development projects were analysed based on the Central Investment Appraisal, a preliminary procedure for large-scale projects; thus, the size of a project was added as a variable. Moreover, because the size is, to some extent, linked to the gross floor area or building-to-land ratio, these factors were also selected as variables for the public and private sectors, respectively. For public development projects, it is common practice to develop projects on public land; in the case of private development projects, a similar variable, which is the rate of project site purchase, was selected. For accessibility to roads and public transportation, the range was limited to subways and train stations within a 5-km radius, considering the nature of most projects located in the Seoul metropolitan area. In addition, regarding the development level around the project site, which is closely related to the development density, this variable was divided into cultural/convenience facilities and educational facilities. These variables are detailed in Tables 4 and 5.


**Table 4.** Variables for public development projects.

**Table 5.** Variables for private development projects.


#### *3.2. Analysis of Public Development Projects*

For the construction of analysis models of public development projects, analysis was performed for the projects [10] that were recommended for additional review in terms of location suitability and environmental factors, such as similar and competing facilities. The data used for the analysis were for 117 projects from 2016 to 2019, and environmental factors and conditions such as the development of basic infrastructure, including environmental factors within a 5-km radius [10] were identified, as well as competing facilities to determine which factors affected Central Investment Appraisal approval. As a result of conducting descriptive statistical analysis prior to the empirical analysis, 81 (68.6%) projects were approved in the Central Investment Appraisal, and

36 (31.4%) failed to obtain approval. The total project cost was KRW 29.7 billion on average, and the mean gross floor area was 9876.18 m2, which indicate that most projects were large-scale ones. For public transportation facilities, there were 68 project sites (58.1%) with subways and train stations within a radius of 5 km, mainly distributed in Seoul and its metropolitan area, which is believed to have contributed to the Central Investment Appraisal approval. In addition, 68 project sites (58.1%) were for large-scale supermarkets and culture/convenience facilities, and 103 (88.0%) project sites were for cultural facilities, thereby indicating that most cultural facilities are located within a 5-km radius. Elementary, middle, or high schools or colleges were located within 0.99 km from the project site on average, and the distance to the farthest facility was 3.86 km, which shows that the proximity to educational facilities was high (Table 6).


**Table 6.** Descriptive statistics.

In the analysis, the -2-log likelihood value was 27.944. In terms of the assessment of goodness-of-fit of the model, *Cox* and *Snell's R*<sup>2</sup> and Nagelkerke's *R*2, both with the same utility for the model fit decisions, accounted for 33.8% and 58.4% of the total variance, respectively. Additionally, the model fit was verified by the Hesmer–Lemeshow test: Pearson's chi-square statistic was 15.316 (*p* = 0.053), thereby indicating that the fit of the model was significant. The analysis showed that the significant variables were the total project cost (Exp(β) = 1.003), adjacent subways/train stations (Exp(β) = 0.929), and adjacent large-scale supermarkets (Exp(β) = 1.153). In summary, for large-scale projects with a large total project cost, the rate of Central Investment Appraisal approval was high and decreased with increasing numbers of subways/train stations in the adjacent area. Furthermore, the approval rate increased with the increasing number of large-scale supermarkets in adjacent areas (Table 7).


**Table 7.** Analysis results.
