DEVELOPMENT OF A CONSTRUCTION QUALITY ASSESSMENT TOOL FOR HOUSES IN SOUTH AFRICA

Housing is a critical socio-economic driver in the vast majority of developing countries, including South Africa. It involves many aspects such as construction quality, affordability, geographic location, long-term financing, and the environment. A key research concern is the quantification of the construction quality of houses and how this may be used to assist in the delivery of better quality houses. This article is based on studies undertaken on housing construction sites in South Africa. A construction assessment tool is developed using principles similar to those used by CONQUAS in Singapore and Malaysia. The tool thus developed is capable of measuring the quality of ‘as-built’ construction elements of a house against national technical standards and specifications, within reasonable time and cost. Studies on the quality of houses were then conducted on 700 houses (two low-income projects and one middle-income project). The results showed that the two low-income projects had average quality scores of 58% and 64%, while the middle-income project scored 80%. Details of the sub-elements of the scores indicated the construction quality assessment tool, the government and other authorities can make better informed decisions when awarding contracts. If introduced and implemented correctly, the quality of the houses delivered across the entire housing spectrum can be measured and monitored, and improvement measures put in place. The data collected through this quality assessment tool will be invaluable for national authorities, regulators, and Statistics South Africa to evaluate and report if the housing stock being delivered is consistently improving. Risk assessment studies will assist the regulators in developing proper quality management strategies. middel-inkomste-projek). Die resultate het getoon dat die twee lae-inkomste-projekte gemiddeld 58% en 64% gehad het, terwyl die middel- inkomste-projek 80% behaal het. Besonderhede van die sub-elemente van die tellings het die ontwikkelingsbehoeftes van die kontrakteurs wat by die projekte betrokke was, aangedui. Met behulp van die instrument vir die evaluering van kwaliteit van konstruksie kan die regering en ander owerhede beter ingeligte besluite neem wanneer hulle kontrakte toeken. As dit korrek bekendgestel en geïmplementeer word, kan die kwaliteit van die huise wat oor die hele behuisingspektrum gelewer word, gemeet en gemonitor word, en verbeteringsmaatreëls ingestel word. Die data wat deur hierdie instrument vir kwaliteitsbeoordeling versamel word, sal van onskatbare waarde wees vir nasionale owerhede, reguleerders en Statistieke Suid-Afrika om te evalueer en te rapporteer of die behuisingsvoorraad wat gelewer word, deurgaans verbeter. Risikobepalingstudies sal die reguleerders help om behoorlike kwaliteitsbestuurstrategieë te ontwikkel.

construction quality assessment tool, the government and other authorities can make better informed decisions when awarding contracts. If introduced and implemented correctly, the quality of the houses delivered across the entire housing spectrum can be measured and monitored, and improvement measures put in place. The data collected through this quality assessment tool will be invaluable for national authorities, regulators, and Statistics South Africa to evaluate and report if the housing stock being delivered is consistently improving. Risk assessment studies will assist the regulators in developing proper quality management strategies.
Sleutelwoorde: Boukwaliteit, behuisingskwaliteit, CONQUAS, kwaliteitsbeoordeling, laeinkomste behuising, Suid-Afrika INTRODUCTION The delivery of quality houses is a broad challenge in the vast majority of developing countries. In South Africa, the political process that led to democracy in 1994 provided an effective political platform for mobilising previously disadvantaged people in securing tenure of housing. Pre-1994, housing delivery was associated with sub-quality low-income houses, and the housing developments were in areas not suitable for human settlements and far from potential workplaces. To address these concerns, the 1994 South African democratic government formulated and implemented the national housing policy and established several state entities to assist in the fast delivery of quality houses. The national housing policy and the subsidy housing programmes accommodate several government housing delivery mechanisms (South Africa, 2010).
Since 1994, the South African housing market has been predominantly driven through the private sector, where the financial institutions primarily finance the middle-and high-income end users, and houses are delivered mainly by established homebuilders. The low-income housing market, defined as those households earning up to R3,500 per month, has been delivered through the local or provincial governments and is mainly dominated by small and emerging homebuilders. The governmentsubsidised housing market delivers products on a fixed house price, commonly referred to as the subsidy housing quantum, and the houses are mass-produced. The low-to middle-income end user, earning between R3,500 and R22,000, is partly financed by the government, and the houses are usually delivered through the social housing schemes (Butcher, 2020: 182). A low-income house is typically 40 square metres in footprint and has two bedrooms, a lounge, kitchen, and a bathroom (South Africa, 2009: 27). However, there is an apparent disparity in the quality of houses delivered in the two markets, despite the availability of information and construction guidelines.
Worldwide, housing is a critical socio-economic development driver and involves several aspects, including construction quality, affordability, geographic location, environment, and long-term financing. Despite the measures put in place by the South African government over the past few decades, the quality of houses is not yet up to acceptable standards, as witnessed in many low-and middle-income housing development projects (NHBRC, 2019: 44).
Several researchers and research documents (Sinha, Sarkar & Mandal, 2017: 337-340;Streimikiene, 2015: 140-145;Zunguzane, Smallwood & Emuze, 2012: 19-38, Statistics New Zealand, 2015 have defined housing quality as encompassing several aspects. Some of these aspects are reasonably objective and include the dwelling type, facilities, number of rooms, and the condition of the dwelling. Subjective aspects, which are also included in housing quality, include user needs, desires, and expectations. Other researchers (Acre & Wyckmansa, 2014: 183-204, Sima, 2015Streimikiene, 2015: 140) have gone beyond housing quality and have included user satisfaction in their analysis. They define user satisfaction in line with the user's needs and aspirations compared to what was delivered physically on the ground.
The minimum standards stipulated in the NHBRC Home Building Manual (NHBRC, 2015) and the national standards (SANS 10400, 2016) apply to all houses delivered in South Africa. However, the standards and guidelines are not prepared and presented efficiently, particularly for some emerging homebuilders. This substantiates a need to develop a tool that will enable the identification of training and developmental needs of homebuilders and assess if this intervention strategy does yield better results in terms of improving the quality of houses.
This article recognises the broader aspect of housing quality assessment, but it focuses on one critical element, 'construction quality'. There is minimal information on how the construction quality of houses is measured and quantified in the literature. The approach proposed in this article provides a formal, comprehensive, easy-to-use mechanism, in which housing construction quality can be quantified and measured. This article thus aims to outline local South African construction practices and benchmark international best practices on construction quality assessment of houses. The outcome of this will lead to the development of a comprehensive, straightforward, and effective assessment construction quality tool, which will enable users to assess the physical aspects of house construction that influence the quality of the 'end product', i.e., the housing top structure.
This article addresses the following key research questions:

1.
How can the physical construction quality of a house be measured and quantified and assure that the structural performance of the house meets the minimum requirements of the South African National Standards (SANS 10400, 2016)?

2.
How can construction quality be monitored, and is there a progressive improvement in quality, as new entrants and technologies come into the marketplace?
3. Does training of homebuilders lead to an improvement in construction quality? If this has a positive effect, how can the impact be measured?

4.
What is the difference in the quality of houses delivered for lowincome earners compared to those for middle-and/or high-income earners?
The answers to these questions require the development of an objective, systematic house "construction quality assessment" tool capable of measuring 'as-built' construction elements against technical standards and specifications. The quality assessment needs to be carried out systematically, rapidly, and at an affordable cost.

Housing delivery overview and quality of houses
Although there has been significant housing delivery in South Africa over the past decades, the trend has been declining (see Figure 1). Houses enrolled through the National Home Builders Registration Council (NHBRC) are a good indicator of houses delivered through the private sector. The NHBRC is a state entity, established through an Act of Parliament (Housing Consumers Protection Measures Act, Act 95 of 1998, as amended "The Act") (South Africa, 1998). The legislative mandate of the NHBRC is to: • Regulate the home building industry. The Act requires all homebuilders to register with the NHBRC; • Establish and promote ethical and technical standards. Every registered homebuilder is required to comply with the NHBRC code of ethics. All houses must be constructed in accordance with the NHBRC technical requirements, and • Improve structural quality in the interests of housing consumers and the homebuilding industry. To achieve this, all houses must be enrolled and inspected by the NHBRC. Although there has been a dip in the delivery of houses in the years 2009 and 2010, due to the financial crisis and economic meltdown, the growth in housing delivery has shown a steady increase over the past ten years.
On the other hand, there has been a decline in the houses delivered by the public sector (South African Government, 2019;NHBRC, [n.d.]: online). This explains why the housing backlog keeps on escalating, with the backlog being 2.3 million as of 2018 (Msindo, 2018: online). Since 1993, several studies have highlighted the challenges faced in the low-income housing delivery in South Africa, especially in metropolitan cities. These challenges include the impact of population growth, shortage of land, corruption, unaffordability, and poverty (Marutlulle, 2019;Bonner, Nieftagodien & Mathabatha, 2012;Jeffery, 2010;Bradley, 2003;Napier, 1993). Apart from these highlighted challenges, other researchers believe that the apartheid government initially caused the housing challenges (Setplan, 2008: 40-50;Baloyi, 2007;Eddy, 2010: 12-18). Several objections to the above statement have, however, been made and argue that these housing challenges are due to the appointment of less experienced contractors, lack of monitoring of the contractors, the poor performing construction sector, irregularities in municipalities, political issues, fraud, and corruption (Gibbon, 2010: 5;Lubisi & Rampedi, 2010: 2).
South Africa has a very sound legislative, regulatory environment and good technical standards for house-construction practices compared to other countries and states. However, despite the availability of all this information and enforceable regulations, poor-quality houses are still being delivered across the entire spectrum of housing (i.e., low-to high-income houses). The NHBRC Annual Report (NHBRC, 2019) highlights the root causes for poor-quality houses as due to any of the following: • The inadequate structural design caused by improper soil classification, resulting in an inadequate foundation solution; • Construction details that are not built in compliance with design specifications; • Use of unsuitable or poor-quality building materials that do not comply with South African National Standards (SANS 10400, 2016); • General poor workmanship; • Inadequate or non-existent service infrastructure such as stormwater systems; • Ineffective monitoring of homebuilders during construction, or • Complete ignorance and/or lack of experience of homebuilders. Gibbon (2010: 5) made similar observations related to substandard workmanship, inappropriate management systems, and lack of a monitoring mechanism on contractors operating in government-subsidised houses, thus contributing to poor-quality houses and delays in housing delivery. The appointment of emerging, less experienced contractors further exacerbates construction delays (Lubisi & Rampedi, 2010: 2). In a study conducted in five of the nine provinces of South Africa (i.e., Gauteng, Eastern Cape, Western Cape, Limpopo, and Mpumalanga), Mgida (2007) found that the application of unconventional building technologies, when used by emerging contractors, becomes an issue that impacts negatively on housing delivery.
Unconventional building technologies are commonly referred as Innovative Building Technologies (IBTs) or Alternative Building Technologies (ABTs) and refer to building products that are certified for compliance with building regulations through a performance assessment (South Africa, 1977). The performance assessment is conducted by Agrément South Africa (ASA, [n.d]:online). The emerging contractors who use these unconventional building technologies have limited knowledge, and there is no proper training to assist them in implementing these technologies. Zunguzane et al. (2012) further observed that municipalities impose unskilled labour on the contractors, thus causing further delays to housing delivery. There is also a minimum input from engineers to monitor the quality of the top structures.

Beneficiary expectation on quality of houses in South Africa
A beneficiary refers to a household that occupies a completed house, while satisfaction in this article refers to the degree to which the end product (i.e., the house) meets the beneficiary's needs, goals and expectations. According to a study conducted in Braamfischerville, Gauteng (Moolla, Kotze & Block, 2011: 138-140), most of the beneficiaries of low-income houses expressed their dissatisfaction with the quality of houses delivered. The vast majority of the houses had poorly built walls and unstable roofs, and the doors were poorly crafted, which resulted in them not functioning well. Most of the beneficiaries complained about poor ventilation, no air vents, and lack of kitchen and bathrooms. The study concluded that 55% of the beneficiaries were not satisfied with the functional aspects of the houses. In a study conducted in Diepsloot, Gauteng (Aigbavboa & Thwala, 2012: 13), most of the low-income beneficiaries were highly dissatisfied with the lack of proper plaster finishing on the inner and outer walls of their houses. The houses were built with no ventilation system to neutralise the inner air condition during the cold and warm seasons.
Studies conducted in Grahamstown, Eastern Cape (Kota, 2010: 26) showed that over 50% of the beneficiaries were unhappy with roof leakages. In this case, the municipal officials had to provide the beneficiaries with plastics to cover the roof and prevent roof leakages. In a further study (Zunguzane et al. 2012) conducted in Wentzel Park, Alexandria, beneficiaries had to use their finances to rectify and self-assure quality on their government-subsidised low-income houses. A high percentage (46%) of the beneficiaries had to use their finances to rectify the houses and, in general, over 50% of the beneficiaries were dissatisfied with the quality of the houses.
As far back as 1967, observations were made that beneficiaries consider the closeness of primary services and essential infrastructure to be of higher value than the actual physical display of the housing compartment (Turner, 1977). It was also noted that low-income houses are considered economical for beneficiaries if they are built close to places of economic activities (cities) and social infrastructure such as schools, hospitals, libraries, clinics, and recreational parks (Turner, 1977). Although the South African government adopted a similar approach in its Breaking New Ground Strategy (South Africa, 2004), it is still criticised for commencing subsidy housing projects for poor citizens in the improper informal settlements on the outskirts of cities far away from inhabitants' places of incomegeneration and primary facilities. Thus, low-income housing beneficiaries end up selling and renting their houses and relocating back to where they were initially residing, or moving to places where it is reasonably close to workplaces and other facilities (Napier, 2009: 71-97).

Quantification of quality of houses
Research on housing quality dates as far back as 1946 (Solow, 1946: 283). Since then, research on housing quality has progressed to include several aspects, including the impact of the market value and microneighbourhoods (Kain & Quigley, 1970: 540). Furthermore, over the years, housing quality has been broadly used to define the condition of a dwelling unit, the characteristics of the physical environment, and end user satisfaction (Streinikiene, 2015: 140-142;Mridha, 2015: 42-54).
In 2011, a study was done in the United Kingdom on the development of a Housing Quality Indicator (HQI, 2011). The HQI is an online toolkit designed to measure, evaluate and improve the building's design quality. The toolkit is broad and considers the location of the house, size, external environment, quality, and cost. Other house-quality tools in the literature include the Building for Life (CABE, 2019: online), which has 20 criteria measures compared to the HQI's ten criteria. In an article, Sinha et al. (2017: 337-347) provide a detailed literature review summary of techniques used to analyse housing quality.
However, all these housing quality assessment tools lack detail on the measurement and quantification of the construction quality of the house. The construction quality is influenced by the design of the house, workmanship during construction, and the quality of the materials used for the house (HQI, 2011: online). Although construction quality is a subset of housing quality, it forms an essential aspect, particularly in developing countries such as South Africa, where the construction quality is poor (NHBRC, 2019). In the context of this article, construction quality is defined as compliance of the construction building elements with technical specifications that are stipulated as minimum standards in the South African National Standards (SANS 10400, 2016).
The most established construction quality assessment tool found in the literature that quantifies the quality of a building is the construction quality assessment system (CONQUAS 21, 2003), launched in Singapore in 2003. After the launch, roughly 2,000 construction projects were assessed that year using the CONQUAS tool. CONQUAS is an assessment system used to measure and quantify the quality of construction building projects, referred to as the CONQUAS score. The assessment method uses a sampling technique, based on the size of the building, to measure the quality of 'selected elements' of the building. CONQUAS' latest edition focuses on quality assessment of three components, namely structural, architectural, as well as mechanical and electrical work.
The literature shows that the use of CONQUAS (2003;2017) has offered many benefits to the Singapore construction industry. After implementing the system in 2003, the CONQUAS score of Singaporean buildings improved from an average of 68% to 75% within eight years, and the target score for 2019 was set to 85.8% (BCA, 2021: online). CONQUAS (2003; 2017) seems to be a robust quality assessment tool that can consistently measure the construction quality of building projects. The principles of CONQUAS have been widely adopted in other countries such as China, Australia and Korea (Kamath & Jayaraman, 2013: 51-67). Of particular note is the adoption of CONQUAS by Industri Pembinaan Malaysia (CIS, 2014: online). The central concept used by the Malaysians is similar to CONQUAS, with the main difference being on the categorisation of buildings, the weightings of building elements, and the sampling guidance.
A survey on quality of houses conducted in 2018 as part of The South African General Household Survey (Statistics South Africa, 2018: 34) indicated that 13.6% of South African households lived in state-subsidised houses (low-income). The survey also included statistics of the construction quality of state-housing units delivered by the government from 1994 to 2018. The level of quality was measured subjectively in terms of the household's opinion whether the walls and roofs of the dwellings were very good, good, needed minor repairs, weak, or very weak. Based on the Stats SA survey, 10.2% of the households reported that their homes had weak or very weak walls, while 9.9% reported the same for their roofing structures.
However, Stats SA's approach is based on perceptions by households (end users) and 'lay-man' understanding of structural failures. For this research, it was noted that a direct application of Singapore's CONQUAS to South Africa would not be applicable as the two countries have different socio-economic conditions, geographic, technical, and political environments. Table 1 highlights and summarises these differences. The land surface is relatively flat, primarily urban with tall buildings.
1.2 million square kilometres, significantly larger than Singapore by roughly 1,700 times.
Geographic spread differs across the country, with some areas being mountainous and rural.
Most housing developments are single or double story.

Socioeconomic and technology (2020)
A population of roughly 5.9 million.
2.9% of GDP is spent on education.
A very low unemployment rate of roughly 2.2% and the standard of living is very high.
The skills base is high, and the quality of education is generally high and acceptable.
The usage and uptake of construction technologies to assist in construction delivery are high.
A population of roughly 54 million.

5.9% of GDP is spent on education.
A high unemployment rate (almost 30%), with more than 53% of youth unemployed.
Historical exclusivity limits the pool of skilled base, and the educational standards differ substantially across the country.
Limited use of technology in construction processes and resistance to adopt the use of innovative technologies.
Notwithstanding the above progress, no formal construction quality assessment systems have been developed in South Africa and other developing countries. Such a system should be able to take the economic and construction dynamics of developing countries into account. These dynamics differ substantially from Singapore and other developed countries and will be highlighted in this article.

PROPOSED HOUSEBUILDING CONSTRUCTION QUALITY ASSESSMENT TOOL
Based on the experiences and the situational analysis of the South African construction industry, the approach adopted in this article to develop the South African house building construction quality assessment tool was similar in principle to the methodology used in the CONQUAS model. The similarity was mainly for the division of building elements and assigning relative weights to these elements. Housing construction quality in South Africa is mainly associated with structural failures and poor workmanship, noticeably so in low-income houses. It was thus considered necessary to focus only on structural aspects affecting the integrity of the building and the roof structure. Concerning the structural integrity, some considerations on electrical works were included in the development of the construction quality assessment tool. The following were established as the objectives of the construction quality assessment tool: • A tool that is objective, simple, and practicable, with a capability of assessing and quantifying the quality of house construction and the performance of home builders; • A tool that will enable the differentiation of homebuilders based on their performance in house construction, and • A tool that will assist in the development of homebuilders to improve the quality of their work.

Elements
The approach adopted in developing the assessment tool suitable for the South African housing construction industry was to break down the housing structure into five building elements that were then further subdivided into sub-elements, as presented in Figure 2. This division forms a substantial departure from CONQUAS regarding the number of components, the inclusion of foundations, the exclusion of separate components related to architectural finishes, and the relative allocation of weighting percentages. In particular, the distribution of weightings shifts the emphasis of the assessment tool away from architectural issues towards structural works.  The weightings were based on observation of structural failures of houses in South Africa, and the actual weightings were determined using the Analytical Hierarchy Process (AHP) (Saaty, 1987). The AHP is a structured mathematical technique used to organise and analyse complex decisions. The technique was used to prioritise building elements that impact on the structural stability and integrity of a house. Thus, the performance of floors (15% weight), for example, is of less importance compared to the walls (25% weight). The electrical and plumbing works (10%) were found to have a small impact on the structure. The impact of electrical and plumbing works on the structure is due to the chasing of walls to make provision for the conduits or pipes. The highest weight of 30% was assigned to the foundation and 25% to the walling element.
The sub-elements were developed from each of the building elements, as shown in Figure 2 for the foundation element. For the other elements of a building, the sub-elements are shown in detail in Table 2. Each building element was broken down into sub-elements that would influence the overall construction quality of the element. This process was based on the performance of historically completed houses, their associated failure patterns, and root causes of failure, and on-site investigations. Saaty's AHP (Saaty, 1987: 161-176) approach, coupled with homebuilder interviews, was used to allocate the sub-weights of the sub-elements.

Assessment criteria
The assessment criteria used in CONQUAS (2003; 2017) employ a strict approach to score compliance items. A building element is assigned either a compliant and given a score of 1.0 or a non-compliant and given a score of 0. With this approach, most of South Africa's low-income houses will end up with a very low overall construction quality score. This will not be helpful, as hardly any information will be extracted from the data. An intermediate score of 0.5 was introduced to obtain reasonable scores and maintain consistency. This score allows some deviations (non-compliances) to occur, but only for those non-compliances that will not adversely affect the house's structural integrity. Other scoring systems such as Likert's five-point scale were considered inappropriate, as they allow too many non-compliances to be introduced and yield unreasonable results.
A set of assessment criteria were developed, using the minimum technical standards and the South African National Standards (SANS 10400, 2016). A large pool of construction technical information was identified. The most relevant, inclusive, and accessible to a homebuilder is the NHBRC's Home Building Manuals (NHBRC, 1999;. The information included in SANS 10400 (2016) tends to be sophisticated to an ordinary homebuilder and concentrates on details of issues pertinent to specific topics. Due to that, the difficulty of access, and the cost of the standards, the SANS 10400 documents are of less practical benefit to homebuilders.

RESEARCH
The proposed construction quality assessment tool developed was used to analyse the quality of 700 houses on three selected project sites located in Gauteng, South Africa. The study was undertaken on two low-income projects, shown in Table 3 as Project A (250 houses) and Project B (350 houses), and one middle-income project, shown as Project C (100 houses).
In Table 3, the homebuilders are indicated with the pseudo symbols A, B, and C, due to the sensitivity and confidentiality of the information. Thus, Project A had two contractors (homebuilders) shown in Table 3 as A1 and A2, each allocated 120 and 130 houses, respectively. Similarly, project B had three contractors, B1, B2, and B3, with 80, 110, and 160 houses, respectively. However, the middle-income project had one contractor (C1) allocated 100 houses.
Destructive investigations are usually costly and take longer to implement compared to visual assessment. In this study, a methodology was adopted in which an assessor would visit the site and assess (score) the various elements of the house structure without performing any destructive tests.

Assessment
Before the study, an on-site, hands-on training of the tool and an interactive calibration based on the results were conducted. The on-site inspection assessment was based on technical documentation in the form of drawings and specifications provided by the homebuilder. Several building elements were included for scoring, and assessors who used the tool were trained prior to the on-site assessment, in order to reduce subjectivity.
To make the assessment easy and quick to use, very basic construction tools and instruments were used for the assessment. These included using a spirit level to measure levelness, a Schmidt hammer to obtain compressive strength of concrete, a moisture meter, a measuring tape, and a camera. A trained assessor performed on-site assessment, using the quality assessment tool and verified, where applicable, with technical documentation (e.g., drawings) available on-site. The skills required for an assessor are a basic technical understanding of standards and experience of site inspection of a building during construction. The assessor is required to carry out the assessment of a building element and sub-element only once. This is in line with the recommendations of CONQUAS (2017) that encourages homebuilders to excel and deliver a good product the first time. Based on observed workmanship and previously recorded structural failures, the tool was also designed to handle complex designs, construction methods, and the usage of different materials and products (e.g., unconventional building technologies) and can be used during and post-construction.
Assessments of the foundation and roof sub-elements (see Table 2) of each house in the three selected project sites were conducted during construction. The final assessment (roof leak) was done six months after completion of the construction work. This process allowed the assessors to perform an objective analysis of the foundations during construction.
To calculate the quality score for each house, each building element was scored, and when aggregated with other elements, it gives an overall score for the house.

Limitations
Previous research has been conducted extensively on beneficiaries' perceptions on quality of houses, as highlighted in the literature review, with hardly any work on the actual measurement and quantification of the quality of the top structure. As such, the limitations of this research are as follows: • The quality assessment research is only for the physical top structure and foundations of a house, and • The assessment is limited to structural aspects and does not include architectural finishes, plumbing, and electrical designs. Table 3 presents the scores for each project, per homebuilder, and the average score for each project. The scores varied from 54% to 80%. The average score for the low-income houses was 64%, for the middle-income, 80%, and the aggregate average score for all houses was 66%. The houses were found to have various types of defects, ranging from inferior quality walls, substandard materials to roof leaks. Interestingly, the lowincome houses constructed by each homebuilder in each project showed similar patterns of defects. Average score 700 66 Figure 3 shows a plot of the frequency distribution of the quality of houses scores, where 100% indicates a near-perfect structurally defect-free house, complying with all relevant national standards. The scores in Figure 3 are based on the elements presented in Table 2. As shown in Figure 3, the data fits in a normal distribution curve, whereby most of the houses have scores falling between 60% and 75%. As part of this research project, a benchmarking exercise on quality scores was conducted on houses that the NHBRC identified in Eastern Cape and KwaZulu-Natal provinces (NHBRC, 2011). These houses were delivered through the governmentsubsidised programme and had structural problems that varied from minor to major defects. According to this exercise, scores below 50% indicated very poorly constructed houses with significant structural defects and would require the houses to be demolished and reconstructed. Scores between 50% and 60% indicated houses with major structural defects, and scores between 60% and 75% indicated houses with minor defects, while scores above 75% indicated houses with insignificant structural defects. The definitions of significant, major and minor structural defects were defined in the NHBRC Home Building Manual (NHBRC, 1999). In the NHBRC manual, the damage or structural defect to a structural element is defined in terms of ease of repair. Using the NHBRC definition in a slightly modified manner, the following was used in this research to define the defects in masonry walls of single-storey houses:

Findings
• Insignificant (hairline) cracks have crack widths in walls less than 0.25mm; • Minor defects have maximum crack widths in walls between 0.25mm and 5mm. Cracks occur internally and are not visible externally. Redecoration of the walls may be required; • Major defects have crack widths between 5mm and 25mm. Extensive repair works to the walls may be required, and • Significant crack widths greater than 25mm would require major repairs, involving partial or complete demolishing of the wall.
Roughly 8% of low-income houses fell below the 50% score and required to be demolished. Of the houses, 30% scored between 50% and 60%, indicating that these houses would require major rectifications. Of the houses, roughly 36% scored between 60% and 75% and these houses would require minor rectifications to meet the local, national standards. As indicated earlier, a score above 75% indicates a house of acceptable quality standards; only 25% of the houses fall into this category. The defects on houses that scored above 75% were cosmetic and could be addressed by the houseowner at minimum cost. Most of the low-income houses fell below the 75% score, with an average score of 64%.
On the other hand, middle-income houses scored above 75%, with an average score of 80%. The difference in scores shows the disparity in the quality of houses delivered in the two mainstream markets. Mainly small and or emerging homebuilders with limited housebuilding knowledge are involved in delivering low-income houses, while well-established homebuilders dominate the middle-income.  35  37  39  41  43  45  47  49  51  53  55  57  59  61  63  65  67  69  71  73  75  77  79  81  83  85  87  89  91  93   Of further interest to observe is the comparison of the quality of building elements of the houses. The building elements and the sub-elements associated with these elements were defined in Figure 2 and Table 2. Each building element has a score, and when aggregated with other elements, it gives an overall score for the house. For the low-income houses, the average quality score for the roofs and walls was 55% and 65%, respectively. This was not surprising, as most of the corrugated (IBR) roof sheets commonly used for low-income houses were torn, and as a result, the roofs were leaking within six months. Some of the roofs had boulders on top to prevent them from wind uplift, or cement blocks were used to anchor the roof beams (see Figure 4). Some of the walls were observed to have poor workmanship associated with skewness and substandard cement blocks. Samples of cement blocks taken from the site for lab testing indicated that the compressive strength was far less than the 3MPa minimum prescribed in SANS 10400 (2016). The compressive strength results ranged from 2.1MPa to 3.1MPa, with a mean value of 2.5MPa. An inadequate foundation system caused wide crack openings that were observed in some of the walls. The impact of crack width was determined as per the NHBRC Home Building Manual (NHBRC, 2015) and SANS 10400 (2016). Unacceptable cracks had a high impact on the house's overall score, which would require the houses to be demolished.
As shown in Figure 5, a log-normal plot represents the frequency distribution for the walls in one of the low-income housing projects. In this graph, approximately 20% of the walls scored below 50%, and roughly 45% scored more than 75%. From the data distribution in the graph, it appears that intervention strategies are required to improve the construction quality of houses in South Africa. Thus, to improve the quality of the walls, the graph must be shifted to the right, using mechanisms such as training, effective quality control, and monitoring, and the use of better quality materials. Upon embarking on these strategies, the construction quality assessment tool can be used to re-assess the quality of houses delivered by the contractors. If the interventions are effective, the graph should shift towards the right. Relevant regulatory authorities, Stats SA, the stakeholders in the housing industry, will then be able to arguably present the housing quality statistics for South Africa and determine whether indeed the quality of housing stock is improving with time. Much better insight is obtained by correlating building elements, as presented in Figure 6, which gives the scores for Project A. The scores of the walls are normalised against the scores of the floors. In the graph, a house is represented by a single data point, and a 45-degree regression line is drawn as shown. This line thus represents scores of equal magnitudes for both the walls and the floors. A point plotting below the unity-regression line denotes a house in which the quality of the floors is superior to the walls.
Conversely, a point plotting above the diagonal line reflects the opposite situation. Figure 6 shows that most of the houses had better quality floors than walls, as most of the data points plotted below the regression line. Such type of correlation graph is useful where the project has many subcontractors. The graphs would enable training interventions that are required and identify which aspects of the project should be addressed, in order to obtain a better quality house. Figure 7 presents a typical example of a poor sub-structure (floor) construction. The image shows that the floor level does not meet the minimum standard, which stipulates that the floor level must be at least 150 mm above the natural ground level. In the event of rain, the house is likely to flood, compromising the integrity of the foundation, the house, as well as the health and safety conditions of the occupants. Therefore, a developmental strategy for this homebuilder would be on the basic understanding of the impact of storm water on top structure performance. A similar comparison of walls and roofs for the same project suggests that the relative construction qualities of both elements are of a similar order of magnitude.

CONCLUSIONS AND RECOMMENDATIONS
South Africa has a dual housing market with an apparent disparity in the quality of houses delivered. A construction quality assessment tool was developed and presented in this article, taking into consideration the local South African conditions typical of the vast majority of developing countries. Although the tool is simple to use, it does provide an objective way of assessing and quantifying the quality of a house objectively and consistently. Implementation of the tool enables a comparison of quality outputs by various homebuilders and developers. It produces a consistent and statistically based measure of quality performance by the entire industry. When integrated with other housing quality systems, this tool will contribute to a holistic assessment of house quality.
Implementing the proposed 'construction quality assessment' tool for houses will benefit several stakeholders and role players in the homebuilding industry. The proposed assessment tool is capable of assessing and quantifying the quality of a house. An analysis of the data generated by the tool identified the developmental needs of the homebuilders. The regulatory authorities, contractor/building organisations, and associations can use the assessment tool to grade the homebuilders into different categories, depending on their historical quality assessment performance scores. Good performers can use their quality assessment score, based on their performance category, to improve the quality of their products and for marketing purposes. Therefore, a potential client can benefit by being able to differentiate and appoint competent homebuilders who can deliver a better quality housing product.
Using the construction quality assessment tool, the government and other authorities can make better informed decisions when awarding contracts. If introduced and implemented correctly, the quality of the houses delivered across the entire housing spectrum can be monitored, and improvement measures put in place. The data collected through this quality assessment tool will be invaluable for national authorities, regulators, and Statistics South Africa to evaluate and report if the housing stock being delivered is consistently improving. Risk assessment studies will assist the regulators in developing proper quality management strategies. 7.