**3. Methodology**

This study uses a qualitative research strategy coupled with a case study approach because the study aims to understand the contextual reasons for urban blight in Sub-Saharan Africa (SSA) in the case of Ghana (see Appendix A for analytical framework). According to Yin [51], the niche of a case study approach is to investigate a contemporary phenomenon that is labelled as "the case" in the real world. This research approach enables in-depth understanding of the pertinent contextual circumstances associated with the case. The choice of Ghana makes the study an interesting one to be investigated in SSA. This is because Ghana is the first SSA country to ascertain independence from colonisation in 1957 and thus the premier country to gain full control of urban planning and land managemen<sup>t</sup> by an aboriginal governmen<sup>t</sup> [52]. Additionally, the concept of urban blight is known in Ghana's urban policy, Land Use and Spatial Planning Act 2016 (Act 925), as previously mentioned. The research will provide a contextual understanding of urban blight to assist city planners and governmen<sup>t</sup> agencies in operationalising urban policies.

#### *3.1. The Study Area*

This study is conducted in East Legon, a first-class suburb of Accra. The area measures approximately 4.95 square kilometres and falls under the Ayawaso West Municipal Assembly (AWMA), as shown in Figure 1. It is classified as a first-class area by the Assembly mainly due to good infrastructural facilities, such as roads, water, and electricity, inclusive of modern developments and major commercial activities. Additionally, it is recognised as one of Accra's most expensive areas, with an influx of international businesses and expatriates [53]. Despite being noted as an affluent area with a high demand for urban space, the area is experiencing blight, and some areas have been in the same condition for many years [54]. To understand the nature of urban blight in East Legon, the study was conducted in phases. These are described in the next section.

**Figure 1.** Map of East Legon, Accra-Ghana. Source: Esri Topographical Map and Land Use and Spatial Planning Authority.

#### *3.2. Criteria for Identifying Blighted Properties*

To identify blighted properties within East Legon, we used a two-stage criteria. The first stage entails definition and characterisation of urban blight, which paves a way for the second stage, a visual selection process based on the outcomes of the first stage. In the first stage, we traced the characteristics of urban blight from the existing local law: Land Use and Spatial Planning Act 2016 (Act 925), as cited earlier in Section 2.1. Characterising urban blight according to the laws of Ghana is necessary to understand the case of Accra as a typical emerging southern city. A global characterisation would otherwise not match the local characteristics and thus have the tendency to blur or misrepresent the local reality/context of how urban blight is understood. We, however, acknowledge that a comparison between local and global perspectives is imperative to position our study within the discourse on urban blight. Thus, based on the characteristics of urban blight stipulated in Section 103 of the Land Use and Spatial Planning Act 2016 (Act 925), we categorised the characteristics into four forms of urban blight according to common descriptions found in literature [23,32,55]. These categorised forms of blight are shown in Table 2.


**Table 2.** A summary of the criteria for selecting blighted properties in this study.

In the second stage, we used the categorised forms of blight to identify blighted properties within East Legon using a virtual neighbourhood audit technique on the Google Earth aerial image. Neighbourhood audit on the general land use of an area could be reliably conducted with Google street view since the viewer is given a virtual feeling of about 15 m resolution [56]. However, the limitations of this remote observation were the fact that it could only provide the spatial perspective of the blighted properties, which was significantly dependent on the spatial resolution. Furthermore, the coverage was constrained because not all the streets and landed properties in the area could be viewed in the aerial images in 3D. Additionally, Pratomo et al. [57] argue that there are uncertainties regarding the spatial analysis of blighted areas because of non-observable indicators such as land tenure. Additionally, Kohli, Sliuzas, and Stein [58] acknowledge that the accuracy of remote sensing techniques for city deterioration requires some level of tacit knowledge. Thus, we augmented the Google Street view with tacit knowledge of the study area and physical inspections (field investigations).

First, different spots of each category of blight were visually detected on the Google image of the study area using visual image interpretation elements such as pattern, shape, and location/association. According to Bakx et al. [59], pattern depicts the spatial arrangements of the buildings where there is repetition of form, style, or relationships; shape takes into consideration the two or three-dimensional projection of the property with Google Street view; association takes into account the relationship between recognisable features and other structures. In this study, we used the element of shape to identify uncompleted structures within the study area; we also used the association of the blighted property with regards to its surroundings to determine single dilapidated/degraded and uncompleted buildings. Finally, we used pattern and location to determine clusters of disordered settlements and vacant plots, respectively. The four categorised forms of blight are illustrated in Figures 2–5 in the subsequent segment. However, during the field visits, some of the properties initially identified as blight were being developed into ultra-modern structures. This enabled us to further narrow our selection to properties that truly match the different forms of blight, as categorised in Table 2.

#### 3.2.1. The Aerial Views of the Four Forms of Urban Blight ClusterofDisorderedSettlements

Figure 2 illustrates an aerial view of clusters of disordered settlements with blue dots. The selection of this form of urban blight is based on irregularity of plots, overcrowding and lack of access to habitable dwellings.

**Figure 2.** Aerial view of cluster of disordered settlements. Source: Google Earth 2018 and parcel plan from Land Use and Spatial Planning Authority.

#### Vacant Plot of Land

The aerial view of a vacant plot of land surrounded by well-developed properties is shown in Figure 3 with a green dot. The criterion for the selection of vacant plots of land is their undeveloped nature.

**Figure 3.** Aerial view of a vacant plot of land. Source: Google Earth 2018 and parcel plan from Land Use and Spatial Planning Authority.

Single Dilapidated Property

The selection of dilapidated properties is based on identification as old, obsolete buildings that are degraded or fallen into disrepair. This is shown by the two red dots in Figure 4 below.

**Figure 4.** Aerial view of dilapidated properties. Source: Google Earth 2018 and parcel plan from Land Use and Spatial Planning Authority.

Uncompleted Structures

Figure 5 shows an aerial view of an uncompleted structure. Aerial selection is based on the foundations of building constructions on site, as shown by the orange dot.

**Figure 5.** Aerial view of an uncompleted structure. Source: Google Earth 2018 and parcel plan from Land Use and Spatial Planning Authority.

#### *3.3. Data Collection*

Both primary and secondary data were collected for the study. For the primary data, a field investigation was carried out in December 2019/January 2020. This involved purposeful identification of the current distribution of blighted properties. To visually engage and stimulate interest and understanding of urban blight in the respondents, as well as evoke deep reflections, a photograph showing the general description of the mixture of well-developed properties and blighted properties in the study area was used in the interviews (see Appendix B). According to Bryman [60], the photo-elicitation method in qualitative interviews serves as an anchor to trigger, excite and evoke the thoughts, views and perceptions of the respondents to provide a meaningful context for the subject matter of discussion. Furthermore, electronic devices—a digital camera and an audio recorder—were used to capture photographs of blighted properties and conversations with the respondents, respectively. Alternatively, there were narrative recordings and jotting down of notes when respondents were uncomfortable with audio recordings.

Secondary data, on the other hand, were information gathered from scientific articles, journals and aerial images: Google Earth 2018, orthophoto 2016 and the land use plan of the study area obtained from Land Use and Spatial Planning Authority (LUSPA) and Accra Metropolitan Assembly (AMA). The Google Earth aerial image was selected based on spatial resolution to help with the visual image interpretation of the blighted properties. Additionally, the orthophoto gave a better spatial resolution of 0.2 m, as well as spatial data from LUSPA. Furthermore, the boundary of the study area (shape file) was acquired from LUSPA, which was used to locate the study area on the aerial images. The Google Earth image was then exported as a kmz file and subsequently converted to kml files in ArcGIS software and geo-referenced accordingly. Finally, the land use plan assisted with the boundaries of the parcels, which was very helpful to the visual interpretation of the Google Earth image, as shown in Figures 2–5.

#### *3.4. Sampling Technique*

A study conducted by Galster [37] described the four main actors of neighbourhoods: households, businesses, property owners and local government. In this study, the four key stakeholders considered are experts from statutory agencies, residents/households, property owners and real estate developers. We used non-probability sampling techniques (purposive and convenience sampling) for the selection of respondents. A purposive sampling technique was used to obtain information from the experts. Purposive sampling is the judgment a researcher uses regarding who can provide the needed and required data for a study [61]. In order to understand the dimensions of land tenure and administration, as well as how they feed into the emergence of urban blight, we interviewed four (4) divisional heads of the Greater Accra Regional Lands Commission. Additionally, given that local authorities in conjunction with the Land Use and Spatial Planning Authority (LUSPA) in Ghana have the prerogative of spatial and land use planning, we interviewed four (4) experts from these authorities to determine their perspectives on urban blight, as well as the root causes, characteristics and implementation dynamics that have featured so far in the regulation of property development in East Legon. In this category, we interviewed eight (8) experts for this study.

Additionally, it is important to recognise that urban blight is perceptive in nature and thus may vary across stakeholders/actors. Therefore, we needed to ge<sup>t</sup> a clear understanding of how other actors perceive and define urban blight, the socio-cultural practices that surround property holdings, and how such dynamics influence the overall attitude of property managemen<sup>t</sup> and development within the study context. To find respondents for this category, we used convenience sampling, also known as accidental sampling, which is based on the researcher's ease of accessing, contacting and reaching respondents. Kumar [61] describes convenience sampling as a technique based on suitability and ease of accessing the respondents for a study. Additionally, response saturation (repetition) was used as a guide for our sample size. As explained by Bryman [60], the saturation point is reached by a researcher when either there are no new discoveries of information or any new information is negligible regarding the objective of the study. For this category, we interviewed 22 respondents, which included residents, property owners and real estate developers in the study area. Overall, a total of 30 respondents were interviewed using the two sampling techniques, as presented in Table 3.


**Table 3.** Summary of respondents and sampling strategies.

#### *3.5. Data Analysis*

The primary data collected via audio recordings were first transcribed into text using Microsoft Word documents. Subsequently, the transcribed documents were uploaded into Atlas.ti software for thematic analysis via open coding in order to identify emerging perceptions and reasons for urban blight. Bryman [60] describes open coding as the process of analysing qualitative data where the researcher remains open-minded to generate as many ideas as possible as well as make meaning out of the data collected by breaking down, comparing and categorising the data into themes. Additionally, the secondary data obtained from the land-use plan (local plan) and the aerial images were used to generate maps using ArcGIS software, as illustrated in Figures 2–6.
