Implementing the AHP multi-criteria decision approach in buying an apartment in Jordan

Journal of Property Research ISSN: 0959-9916 (Print) 1466-4453 (Online) Journal homepage: http://www.tandfonline.com/loi/rjpr20 Implementing the AHP multi-criteria decision approach in buying an apartment in Jordan Mohammed Said Obeidat, Tarek Qasim & Aseel Khanfar To cite this article: Mohammed Said Obeidat, Tarek Qasim & Aseel Khanfar (2017): Implementing the AHP multi-criteria decision approach in buying an apartment in Jordan, Journal of Property Research, DOI: 10.1080/09599916.2017.1413588 To link to this article: https://doi.org/10.1080/09599916.2017.1413588 Published online: 08 Dec 2017. Submit your article to this journal View related articles View Crossmark data Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalinformation?journalcode=rjpr20 Download by: [46.185.213.128] Date: 08 December 2017, At: 04:54

Journal of Property Research, 2017 https://doi.org/10.1080/09599916.2017.1413588 Implementing the AHP multi-criteria decision approach in buying an apartment in Jordan Mohammed Said Obeidat, Tarek Qasim and Aseel Khanfar Department of Industrial Engineering, Jordan University of Science and Technology, Irbid, Jordan ABSTRACT Buying an apartment or a house is an important step in everyone s life worldwide, to reach a settled and stable life. Several criteria are considered when buying an apartment or a house. In Jordan, apartments are customers preferred choices because of financial circumstances. It is not easy for a person to decide on the apartment s specifications such as location, design, building design, and finances. This study assists people in selecting an appropriate apartment using the Analytical Hierarchy Process (AHP), which is considered an important multi-criteria decision-making approach. Data used in this study were collected in Jordan; however, people worldwide can benefit from this study. The methodology used is twofold. First, feedback was considered from five investors in the real estate sector in Jordan about specifications that customers consider when buying an apartment. Second, several customers were asked about their preferences in a dream apartment using a pairwise comparison questionnaire, which was collected from 305 participants to obtain the priorities of 10 different apartment alternatives found in the Jordanian real estate market. The AHP technique is used to analyse the collected data to assist customers in reaching the best purchase decision. 1. Introduction ARTICLE HISTORY Received 26 April 2017 Accepted 1 December 2017 KEYWORDS Apartment; sensitivity analysis; Jordan; analytical hierarchy process (AHP); multi-criteria decisionmaking (MCDM) Multi-criteria decision analysis is a tool that is applied for evaluation of candidate alternatives for ranking, sorting or choosing based on a number of quantitative and/or qualitative criteria and is associated with different measuring units (Özcan, Çelebi, & Esnaf, 2011). It can be applied to many complex decisions. The primary multi-criteria decision-making approaches are Analytic Hierarchy Process (AHP), Elimination and Choice Expressing Reality (ELECTRE), Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), and Grey Theory. Judgments and decision-making are an integral part of the average person s daily life. Moulding human judgment receives considerable attention in and out of the psychology science (Anderson, 1970); (Dawes, 1971); (Louviere, 1974). One important decision that a person may face is buying an apartment. Owning an apartment is an important step in most CONTACT Mohammed Said Obeidat 2017 Informa UK Limited, trading as Taylor & Francis Group msobeidat1@just.edu.jo

2 M. S. OBEIDAT ET AL. peoples life worldwide, so that people may settle down and have a stable life. This is also the case in Jordan, because this will influence the quality of the future life of a family. This study is concerned with the quantitative description of a person s judgment when buying an apartment that matches their requirements and preferences. People often tend to search for an apartment with specifications and qualities that they can find for the lowest cost possible. In Jordan, the most common way a person will decide upon purchasing an apartment is as follows: First, a person will summarise the required specifications and characteristics of the dream apartment. Second, the person starts the search by visiting several real estate buildings in the target city. This might take several months or sometimes about a year until finding some apartments with the required specifications or suitable options, especially with the rapid development of architectural structures that increase the variety of available alternatives. At this point, a complex decision-making problem evolves in that the person must choose one apartment out of the shortlisted candidates. In Jordan, close family members and friends influence decision-making during the process of selecting an apartment. However, this is not the best solution due to the lack of a systematic procedure in selecting an apartment alternative. Some problems may be faced when a final decision is made and the buying is accomplished. There might be some important characteristics and requirements in the apartment that were missed and not considered in the unsystematic purchasing procedure of this decision-making process, which may force the buyer to sell the apartment and restart the apartment searching procedure from scratch. This is a total waste of time, money and effort. To the best of our knowledge, there is no study in the literature that applies a methodological approach for selecting an apartment for a family using multi-criteria decision-making approaches. The main rationale behind this research is to help people choose an appropriate apartment using the AHP multi-criteria decision analysis approach. To investigate this approach, data were collected in Jordan from five investors in the real estate industry and from 305 participants who were searching for an apartment. The AHP was used to analyse the collected data to help real estate customers worldwide and in Jordan reach the best decision when buying an apartment. The rest of this paper is organised in the following manner. Literature is reviewed in the next section. The status of the housing market in Jorden is presented in Section 2.1. The methodology of this paper is described in Section 3. A brief summary of the questionnaire respondents is presented in Section 4. The results and discussion are given in Section 5. Lastly, Section 6 provides the conclusions. 2. Literature review While many studies have been conducted on finding optimal location alternatives for industrial and commercial sectors, general households and families do not have a comprehensive study to assist them in buying their dream home or apartment with peace of mind. Louviere and Henley (1977) performed related research on university students who were searching for a rental apartment based on three factors of student concern, namely price, distance from the campus and the quality of the apartment. Their results showed that students do not prefer to rent an apartment that is far away from campus (Louviere & Henley, 1977). The main driver of housing preferences for first-time home owners is finance, where decisions related to choice and timing of housing depend on site-specific factors (Reed & Mills, 2007).

JOURNAL OF PROPERTY RESEARCH 3 No scientific and analytic study that utilises multi-criteria decision-making tools was found in the literature to assist people in deciding upon the apartment they want to buy. No research was found on using multi-criteria decision-making tools to select an apartment for family use. However, a few studies were found that used AHP to choose a site for an industrial application. Examples include evaluating optimal locations for new agriculture product warehouses (García et al., 2014), prioritising relief logistic centre locations for natural disasters (Bozorgi-Amiri & Asvadi, 2015), selecting a retail store location (Erbıyık, Özcan, & Karaboğa, 2012), selecting a store location (Turhan, Akalın, & Zehir, 2013), and locating new shipyard zone locations in Turkey (Saracoglu, 2013). One study was found in which three multi-criteria decision tools were used together, namely TOPSIS, ELECTRE and Grey Theory, to select a warehouse location (Özcan et al., 2011). The general structure for selecting a site consists of the following: evaluating candidate alternatives according to specific criteria, describing relevant criteria in the process of decision-making, developing candidate alternatives for site selection, and evaluating the candidates and making a decision by choosing the best alternative (Erden & Coşkun, 2010; Ertuğrul & Karakaşoğlu, 2008). The objective of multi-criteria decision-making is to assist the decision-maker in choosing the best alternative out of several feasible alternatives (Erden & Coşkun, 2010). The application of multi-criteria decision-making methods is found in wide areas based on the nature of the decision. The most used multi-criteria decision-making approaches are ELECTRE, TOPSIS and AHP. ELECTRE methods (I, II, III, IV, IS, and TRI) have been selected as the best methods by pairwise comparison of alternatives within the decision problem (Özcan et al., 2011). ELECTRE methods have been used in previous studies; examples include but are not limited to support a group of decision makers with different value systems (Leyva-López & Fernández-González, 2003) in the area of solid waste management (RogerS & Bruen, 1998), in assessing an action plan for the diffusion of renewable energy technologies at a regional scale (Beccali, Cellura, & Mistretta, 2003), in the area of water resources planning (Raj, 1995) and in material selection under weighting uncertainty (Shanian, Milani, Carson, & Abeyaratne, 2008). The TOPSIS method generates a positive ideal alternative solution and a negative ideal alternative solution. This method is based on the concept that the selected alternative should simultaneously have the maximum geometric distance from the negative ideal solution and the minimum geometric distance from the positive ideal solution (Assari, Maheshand, & Assari, 2012); (Karim & Karmaker, 2016); (Benítez, Martín, & Román, 2007). The formation of a decision matrix and normalised decision matrix in the TOPSIS method is performed by using precise scores that each alternative receives from all the criteria (Karim & Karmaker, 2016). Positive and negative ideal solutions are found when considering the rates of all attributes. The preference order of the alternatives is determined by comparing the distance coefficient of each alternative. TOPSIS applications were shortlisted for financial investment decisions such as outranking of highway buses (Feng & Wang, 2001) and identifying new active investment opportunities (Kim, Park, & Yoon, 1997), and were also applied in the field of operations management such as in decision problems related to the selection of production processes of semiconductors (Chau & Parkan, 1995) and supplier selection in manufacturing industries (Vimal, Chaturvedi, & Dubey, 2012). The AHP is a suitable tool for working on decisions under certainty, where judgment in a logical manner is quantified and considered as a base of achieving a good decision (Taha,

4 M. S. OBEIDAT ET AL. 2011). The AHP is a structured technique for organising and analysing complex decisions, considering psychology and mathematics (Venkataraman, Vijaya Ramnath, & Kannapan, 2014). The AHP reaches decisions for complex problems by quantifying nonnumeric factors affecting decision-making such as emotions, feelings, ideas, etc., of the people involved in the decision-making process (Taha, 2011). Certain numeric values are assigned to these factors between 1 and 9 by decision makers, stakeholders and/or experts for pairwise comparison; a greater value indicates more importance of a factor (Taha, 2011). The values are represented in a matrix form to indicate the relative importance or influence of a factor with respect to other factors within the same hierarchical level (Abu Qdais & Alshraideh, 2016). An AHP model consists of four main steps to make a decision in an organised manner that enhances the process of generating priorities (Saaty, 2008). The first step is to define the problem and to determine knowledge research (Saaty, 2008). The second step is to construct a decision hierarchy from the top, by identifying the decision goal, the related objectives from a broad perspective that are assigned throughout the intermediate levels (which include different criteria that subsequent factors depend on) to the lowest levels of several candidate alternatives (Saaty, 2008). Afterwards, the construction of the set of comparison matrices starts. Each element in a higher level of the hierarchy is utilised to compare all elements in the level immediately under it with deference to the upper element (Saaty, 2008). If the comparisons are not perfectly consistent with each other, then it provides a mechanism for improving consistency (Triantaphyllou & Mann, 1995). The last step is to use the priorities obtained from the constructed comparison matrices to weigh the priorities of elements in the immediate level below (Saaty, 2008). AHP has been used intensively in a wide variety of decision-making situations in several studies. It has contributed to a wide range of applications. Vaidya and Kumar (2006) provided over 150 applications of the AHP. They categorised these applications into 10 areas: selection, evaluation, benefit-cost, allocation, planning and development, priority and ranking, decision-making, forecasting, medicine and quality function deployment (Vaidya & Kumar, 2006). Examples of these 10 applications are as follows: supplier selection (Dweiri, Kumar, Ahmed Khan, & Jain, 2016), evaluation (Fan, Zhong, Yan, & Yue, 2016), benefit-cost (Wedley, Choo, & Schoner, 2001), allocation (Yu & Tsai, 2008), planning and development (Chen & Wang, 2010), priority and ranking (Khan & Ahmad, 2017), decision-making (Özcan et al., 2011), forecasting (Wang et al., 2014), medicine (Moon et al., 2015) and quality function deployment (Chen, 2016). As can be seen, the AHP is a suitable analytical solution to the problem addressed in this study, especially when taking into account both parties (real estate investors and buyers), which adds more variability to the data collected. This makes the results more acceptable to use in any region or real estate market worldwide. 2.1. The housing market in Jordan Apartment or home financing in Jordan comes from three sources: (1) individual financing through savings, money transfer from abroad, or property sale; (2) regular financing through loans from banks or other financial institutions; and (3) irregular financing such as family assistance or family member loans (Al-Homoud, Al-Oun, & Al-Hindawi, 2009). According to statistics in 2011, about 60% of Jordanians own their homes (OXFORD Business Group, 2011). In Jordan, people who want to rent a home or an apartment are between 25 and 40% of the Jordanian population (OXFORD Business Group, 2011).

JOURNAL OF PROPERTY RESEARCH 5 Considering land tenure in Jordan, people of Jordan and even foreign investors have the ability to buy land or apartments. Considering the role of mortgages in Jordan, there are a large number of banks and financial institutions that support customers in buying their apartments or homes by giving loans with competitive interest rates associated with comfortable payback periods. This enhances the process of home and apartment and even land buying in Jordan. Regarding the types of dwellings, there are several options for housing including villas, which are of interest to rich people, apartments, regular homes, etc. The price of residential properties in Jordan remained stable at a high level throughout 2016; however, there has been a decline in residential property transactions (Delmendo, 2017). Based on the Jordan-based news agency Petra, the demand for property remained weak in the first quarter of 2017, with a 6% decline in property transactions compared to the same period of 2016 (Delmendo, 2017). Apartment prices in Jordan vary between cities. In addition, apartment prices in the same city also vary between districts or neighbourhoods. For example, in the capital of Jordan, Amman, prices are between $550 and $1940 per square metre. Irbid city is the largest city in the north of Jordan, and apartment prices in this city range between $450 and $850 per square metre. In Zarqa city, the largest industrial city and the second largest city based on population in Jordan, apartment prices are the same as those of Irbid city. According to the Central Bank of Jordan, loans to the construction industry are increasing (Delmendo, 2017). In March 2017, these loans were $8.66 billion with about a 16% increase from the previous year (Delmendo, 2017). Currently, the Jordanian government waives registration fees for the first 150 m 2 of any home smaller than 180 m 2. As a result, the trend is that consumers prefer buying smaller units with competitive prices. Usually, apartments in this range consist of three bedrooms, one living room, and one kitchen; larger sizes (those close to 180 m 2 ) may include an additional maid or storage room. 3. Methodology The methodology of this study is twofold. The first considers real estate investors who construct buildings and sell apartments to customers (the investors), and the second is focused on those people who are searching for an apartment (the customers). Data in this study were collected in Jordan from five real estate investors and 305 customers. Feedback from five investors in the real estate sector in Jordan related to apartment specifications that customers consider when buying an apartment was considered. Several apartment buyers were interviewed and asked about their needs for the apartment they wanted to buy to enhance the feedback obtained from the five investors. Based on the feedback collected from the five investors and the interviewed customers, the following were obtained: (1) Ten apartment alternatives (named A1, A2,, A10) were chosen for comparison. Details of these apartment alternatives are shown in Table 1. (2) Next, a pairwise comparison questionnaire was designed to prioritise the 10 apartment alternatives. The questionnaires were distributed using social media and targeted those people interested in buying an apartment. The buyers were divided into first time buyers and repeat buyers. (3) The data obtained were analysed according to the AHP multi-criteria decision approach using Expert Choice v.11 software to obtain the apartment priorities. The 10 apartment alternatives were evaluated based on the following main criteria related to the apartment: location, design, building design, and financials or economics. Table 1

6 M. S. OBEIDAT ET AL. Table 1. The 10 apartments specifications. Main criteria Sub-criteria Sub-subcriteria Sub-sub- Sub-criteria Apartments A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 Location Close to villas Yes No No Yes No No No No No Yes Worship places Yes Yes Yes No No Yes Yes No No No Close to work Close to markets No No No No No No No No No No Close to schools No No No No No Yes No No No No Transportation No Yes Yes No No Yes Yes Yes Yes Yes Side with street Main Local Local Local Local Main Local Local Local Local Population Low Low Low Low Moderate Moderate Moderate Low Low Low Apartment design Storey number 4 0 2 1 0 2 4 3 4 2 Ready kitchen No No No No No Yes Yes No Yes No Direction North North West North North North North East East West No. of balconies 1 1 1 1 2 3 3 2 2 2 Finishings Super Deluxe Super Normal Super Deluxe Super Deluxe Deluxe Deluxe Apartment style Normal American American American Normal American American American American American Water tank volume Apartment area Apartment area Number of rooms 2 2 2 2 4 2 11 5 5 2 203 140 150 150 180 200 240 222 150 165 Kitchen area Moderate Small Moderate Small Large Moderate Large Large Small Small 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 No. bathrooms. No. living rooms 2 2 2 2 2 3 2 2 2 2 Maid room Yes No No No No No Yes No No No Storage room Yes No No No Yes No No No No No Bedroom Master room Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Regular room No No No No No No No No No No

JOURNAL OF PROPERTY RESEARCH 7 Main criteria Sub-criteria Sub-subcriteria Sub-sub- Sub-criteria Apartments A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 Building age New New New New Old New New New New New No Yes Yes Yes Yes Yes Yes Yes Yes Yes Sewage connection Elevator Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes 3 3 3 3 2 2 3 3 3 3 No. of apartments per storey Entrance 1 2 1 1 2 1 1 1 1 1 Parking garage Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Interface Stone Stone Stone Stone Stone Stone Stone Stone Stone Stone Economics Price (Jordan 90,000 53,000 50,000 52,000 72,000 70,000 92,000 85,000 50,000 61,000 Dinar) Method of payment Cash Cash Cash Cash Cash Cash Cash Cash Cash Cash

8 M. S. OBEIDAT ET AL. provides the specifications of the different apartments that were considered, and Figure 1 shows the AHP model for this study. The structure of the questionnaire supported the AHP approach. It was constructed of comparison tables (matrices) of similar rows and column number, and simultaneously, this number was equal to the number of criteria/sub-criteria considered in each comparison table at the same level. For example, under the financials criteria, the price and the method of payment were compared. The comparison table of these two sub-criteria was constructed with two columns and two rows. Participants were asked to assign a weight in the form of a numeric value between 1 and 9 according to their preference that measured the relative importance of the row criteria over the column criteria in the same comparison table. As mentioned previously, higher values indicate that the row factor is of higher importance. The 1 9 scale was explained to the participants at the beginning of the questionnaire. The data obtained for all of the pairwise comparison tables in the questionnaire were averaged Level 1: Goal Level 2: Criteria Level 3: Sub-criteria Level 4: Sub-Sub-criteria Selecting an apartment Location Apartment Design Building Design Finances Close to villas Close to worship places Close to work Close to markets Close to schools Available transportation Street type Population density in area Bathrooms Living room Storey Number Ready kitchen Direction Number of Balconies Finishings Apartment style Water tank volume Apartment Area Number of Rooms Apartment area Kitchen area Building age Sewage connection Elevator No. apartments per storey Entrance Parking garage Outside Interface Price Method of payment Maid room Storage room Bedroom Level 5: Sub-sub-sub-criteria Master room Level 6: Alternatives Regular room A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 Figure 1. The AHP model for selecting the best apartment.

JOURNAL OF PROPERTY RESEARCH 9 to obtain the average comparison tables for all criteria/sub-criteria in the decision model, and these average tables were used in the analysis. 4. Respondents The total number of respondents to the questionnaire was 305 participants. About 60% of respondents were female. The respondents were divided into two categories: first time home buyers and repeat buyers. The average age for first time buyers was 28 years, while it was 48 years for repeat buyers. Respondents income was classified as follows (among all respondents): about 40% of the responders have a monthly income over $1400, about 21% a monthly income between $850 and $1400, about 21% a monthly income between $570 and $850, about 13.5% a monthly income between $285 and $570, and about 4.5% low income of less than $285 per month. 5. Results and discussion The ten apartment alternatives studied in the Jordan real estate market were compared using the AHP decision approach. The apartments were compared based on four main criteria: location, design, building design, and financials, as shown in Figure 1. The location criterion, divided into eight sub-criteria, formed the third level of the hierarchy; these sub-criteria are whether the apartment is located close to villas, close to worship places, close to work, close to markets, close to schools, has available transportation, the type of street close to the apartment location in the building (main or local street), and the population density in the area. Considering the design criterion, nine sub-criteria formed the third level of the hierarchy; these are the apartment stories, whether the apartment has a ready kitchen, the apartment side in the building (direction), the number of balconies, finishings, the apartment style, the volume of the additional water tank, the apartment area, and the number of rooms. The apartment area and the number of rooms sub-criteria are composed of additional sub-criteria that formed level 4 of the hierarchy. The area sub-criterion is divided into two components: the apartment area and the kitchen area; the number of rooms sub-criterion is composed of five components: the number of bathrooms, the number of living rooms, the availability of a maid room, the availability of a storage room, and the number of bedrooms. Of these components, the number of bedrooms is also composed of two additional components forming level 5 of the hierarchy that includes a master room and regular room. The building design criterion is divided into seven sub-criteria forming the third level of the hierarchy. These criteria include building age, sewage connection, presence of an elevator, number of apartments in each storey, building entrance, whether the building has a parking garage, and the building interface from outside. Financials is the last main criterion, which is divided into 2 sub-criteria forming level 3 of the hierarchy: the apartment price and the method of payment. Relative weights obtained based on the results of pairwise comparisons of the different criteria/sub-criteria in the AHP model are shown in Figure 2. These weights are obtained using the Expert Choice v.11 software that is used to perform the AHP analysis. Out of the four main criteria, relative weights indicate that apartment location is the most important criterion with a relative weight of 39.5%, followed by apartment design criterion with a relative weight of 32.3%, financials criterion comes next with a relative weight of 18.4%,

10 M. S. OBEIDAT ET AL. Location (0.395) Close to villas (0.068) Close to worship places (0.167) Close to work (0.092) Close to markets (0.116) Close to schools (0.121) Available transportation (0.190) Street type (0.108) Population density in area (0.138) A1.118 A2.118 A3.118 A4.118 A5.059 A6.059.038.042.091.053.175.115.125.091.105.175.115.125.091.105.175.115.042.091.105.025.115.042.091.105.025.038.125.182.105.175 A7 A8 A9 A10.059.115.125.091.105.175.118.115.125.091.105.025.118.115.125.091.105.025.118.115.125.091.105.025 Figure 2. The AHP model relative weights..130.087.087.130.087.087.087.087.087.130 Selecting an apartment Apartment Design (0.323) Building Design (0.098) Finances (0.184) Storey Number (0.134) Building age (0.161) Price (0.667) Ready kitchen (0.068) Sewage connection (0.270) Method of payment (0.333) Direction (0.093) Elevator (0.145) Number of Balconies (0.035) No. apartments per storey (0.108) Finishings (0.140) Entrance (0.065) Apartment style (0.063) Parking garage (0.162) Water tank volume (0.148) Outside Interface (0.089) Apartment area (0.163) Number of Rooms (0.156) Bathrooms (0.261) Apartment area (0.667) Living room (0.149) Kitchen area (0.333) Maid room (0.079) Storage room (0.147) Bedroom (0.363) Master room (1.00) Regular room (0.00).000.214.167.087.091.119.054.167.146.055.031.063.30.083.045.000.071.083.087.045.060.054.083.075.055.031.063.110.167.045.000.071.083.087.091.119.054.083.146.055.285.063.152.083.045.000.071.083.087.045.085.054.083.042.055.031.063.204.083.045.000.214.083.087.182.174.167.146.103.031.063.110.167.318.000.071.083.217.091.122.054.083.075.184.031.188.152.083.318.000.071.167.087.182.031.311.083.146.184.031.188.030.083.045.000.071.083.087.182.059.132.083.075.103.122.063.030.083.045.000.071.083.087.045.085.132.083.075.103.122.188.030.083.045.000.071.083.087.045.146.054.083.075.103.285.063.152.083.045.016.022.020.148.207.165.050.060.017.027.207.097

JOURNAL OF PROPERTY RESEARCH 11 and finally the building design criterion with a relative weight of 9.8%. The weights of the main criteria imply that the apartment location occupies the highest priority for customers, followed by apartment design, the financials, and lastly by the building design. Higher weight means higher customer preference. Considering the location sub-criteria, the availability of transportation occupies the highest priority of 19%, followed by closeness to worship places with a relative weight of 16.7%, population density in the area with a relative weight of 13.8%, closeness to schools with a relative weight of 12.1%, closeness to markets with a relative weight of 11.6%, type of street with a relative weight of 10.8%, closeness to work with a relative weight of 9.2%, and closeness to villas with a relative weight of 6.8%. Regarding the design of the apartment, the area of the apartment has the highest priority of 16.3%, followed by the number of rooms with a relative weight of 15.6%, the volume of the additional water tank with a relative weight of 14.8%, the apartment finishings with a relative weight of 14%, the apartment storey number with a relative weight of 13.4%, the side of the apartment (direction) with a relative weight of 9.3%, whether the apartment contains a ready kitchen with a relative weight of 6.8%, the apartment style with a relative weight of 6.3%, and the number of balconies with a relative weight of 3.5%. The financials sub-criteria indicate that the apartment price occupies the highest priority with a weight of 66.7%, followed by the method of payment with a relative weight of 33.3%. The sub-criteria of the building design show that the sewage connection is the most important for customers and occupies the highest priority of 27%, followed by approximately a tie between the presence of a parking garage and the age of the building with relative weights around 16%, the presence of an elevator with a relative weight of 14.5%, the number of apartments in each storey with a relative weight of 10.8%, building interface with a relative weight of 8.9%, and building entrance with a relative weight of 6.5%. For the apartment area, two components formed level 4 of the hierarchy: The apartment area with a relative weight of 66.7% and kitchen area that is the rest of the relative weight. The number of rooms sub-criteria is divided into 5 additional components forming level 4 of the hierarchy. The highest relative weight is for the number of bedrooms of 36.3%, followed by number of bathrooms with a relative weight of 26.1%, number of living rooms with a relative weight of 14.9%, presence of a storage room with a relative weight of 14.7%, and the presence of a maid room with a relative weight of 7.9%. The bedroom component in level 4 is also decomposed into two additional components forming level 5 of the hierarchy; these are a master room and a regular room. The relative weight of the master room is 100%, which means all participants prefer having master rooms in the apartment. Considering the location criterion and its 8 sub-criteria in the model hierarchy as shown in Figure 2, the relative weights of the apartment closeness to villas sub-criterion shows that apartments 1, 4, and 10 occupy the highest weights of 13% each, while the remaining apartments have equal relative weights of 8.7% each. The highest relative weight means highest priority or highest customer preference. Regarding the apartment closeness to worship places sub-criterion, the relative weights show that apartments 1, 2, 3, 6, and 7 have the highest priority of 17.5% each, while the remaining apartments have the next priority of 2.5% each. Considering the apartment closeness to work sub-criterion, the relative weights show that apartments 2 10 share the highest priority of 10.5% each, and apartment 1 has the lowest priority of 5.3%. Comparing the apartment closeness to market sub-criterion, relative weights show that all apartment alternatives (1 10) have the same priority of 10% each. Apartment 6 occupies the highest priority of 18.2% when relative weights of the

12 M. S. OBEIDAT ET AL. apartment closeness to school sub-criterion are compared, and the remaining apartments share the same priority of 9.1%. Regarding transportation availability in the area sub-criterion relative weights, apartments 2, 3, 6, 7, 8, 9, and 10 have the highest priority of 12.5% each, and the remaining 3 apartments have the lowest priority of 4.2% each. Based on the relative weights of whether the apartment has a side with a main or local street sub-criterion, apartments 2, 3, 4, 5, 7, 8, 9, and 10 share the highest priority of 11.5% each, and the remaining 2 apartments share a priority of 3.8% each. Considering the population density in area sub-criterion, the relative weights show that apartments 1, 2, 3, 4, 8, 9, and 10 have the highest priority of 11.8% each, while the remaining apartments have a priority of 5.9% each. Considering the apartment design criterion, which is divided into 9 sub-criteria as shown in Figure 2, the relative weights of the apartment storey number sub-criterion show that apartment 1 occupies the highest priority of 30%, followed by apartment 4 with a priority of 20.4%, apartments 3, 6 and 10 with a priority of 15.2% each, apartments 2 and 5 are tied with a priority of 11%, and the remaining apartments have the lowest priority of 3% each. Comparing the relative weights of whether the apartment has a ready kitchen sub-criterion, apartments 6, 7 and 9 have the highest priority of 18.8% each, and the remaining apartments have a priority of 6.3% each. The relative weights of the apartment direction sub-criterion show that apartments 3 and 10 have the highest priority of 28.5%, followed by apartments 8 and 9, which also have the same weights of 12.2%, and the remaining apartments have the lowest priority of 3.1% each. Regarding the number of balconies sub-criterion, the highest priority is tied between apartments 6 and 7 with 18.4%, followed by apartments 5, 8, 9 and 10 with priority of 10.3% each, and then apartments 1 4 with a priority of 5.5% each. Considering the finishings sub-criterion, apartments 1, 3, 5 and 7 have the highest priority of 14.6% each, followed by apartments 2, 6, 8, 9 and 10 with a priority of 7.5%, and then apartment 4 with a priority of 4.2%. Comparing the relative weights of the apartment style sub-criterion, both apartments 1 and 5 have the highest priority of 16.7%, and the rest of the apartments have the lowest priority of 8.3%. The relative weights for the additional water tank volume sub-criterion show that apartment 7 has the highest priority of 31.1%, followed by apartments 8 and 9 that are tied with a priority of 13.2%, apartment 5 of 10% priority, and the remaining apartments have a priority of 5.4% each. As shown in Figure 2, the apartment area factor is divided into two components: the apartment area and the kitchen area. Regarding the apartment area, the relative weights show that apartment 5 has the highest priority of 17.4%, followed by apartment 10 with a priority of 14.6%, apartment 6 with 12.2% priority, apartments 1 and 3 with a priority of 11.9% each, apartments 4 and 9 with a priority of 8.5% each, apartment 2 with a priority of 6%, apartment 8 with 5.9% priority, and then apartment 7 with 3.1% priority. The highest priority of the kitchen component is for apartments 5, 7 and 8 with a priority of 18.2% each, followed by apartments 1, 3, and 6 with a priority of 9.1% each, and the remaining apartments have the lowest priority of 4.5% each. The number of rooms factor is divided into 5 components as shown in Figure 2. Considering the number of bathrooms component, all of the apartments have equal priority of 10% each. Regarding the number of living rooms component, the highest priority is for apartment 6 with 21.7%, and the remaining apartments have a similar priority of 8.7% each. Comparing the relative weights of the maid room component, apartments 1 and 7 have the highest priority of 16.7%, and the remaining apartments have the lowest priority of 8.3% each. For the storage room, the relative weights show that apartments 1 and 5 have the highest priority of 21.7%, and the remaining apartments have the next priority of 7.1%

JOURNAL OF PROPERTY RESEARCH 13 each. The bedroom factor is decomposed into a master and regular room components. The relative weights of the master room component show that all of the available apartments have the same priority of 10% each; however, the relative weights of the regular bedrooms are all zeros, meaning that any apartment that has only regular rooms, with no master room(s), is not preferred by customers. The financials criterion is divided into two sub-criteria as shown in Figure 2. Comparing the relative weights of the price, apartments 3 and 9 have the highest priority of 20.7% each, followed by apartment 4 with a priority of 16.5%, apartment 2 with 14.8% priority, apartment 10 with 9.7% priority, apartment 6 with a priority of 6%, apartment 5 with 5% priority, apartment 8 with 2.7% priority, apartment 1 with a priority of 2%, and apartment 7 with a priority of 1.7%. For the method of payment sub-criteria, the relative weights are equal to 10% for all alternatives. The last main criterion is the building design, which is divided into 6 sub-criteria. The relative weights for the building age show that apartment 5 has the lowest priority of 2.2%, and the remaining apartments have the highest priority of 10.9% each. The relative weights of the sewage connection sub-criterion show that apartment 1 has the lowest priority of 1.6%, and the reaming apartments have the highest priority of 10.9% each. Comparing the relative weights of the presence of an elevator in the building sub-criterion, all apartments have an equal priority of 10%. This equivalence in the relative weights is because all buildings have elevators. Considering the relative weights of the number of apartments per each storey sub-criterion, the highest priority is given to apartments 5 and 6 with 31.8% each, and the remaining apartments have the lowest priority of 4.5% each. The relative weights of the building entrance sub-criterion show that apartments 2 and 5 have the highest priority of 16.7% each, and the remaining apartments have the lowest priority of 8.3% each. The relative weights of the parking garage availability sub-criterion show that all apartment alternatives have the same priority of 10% each. This equivalence in the relative weights is because all apartments have a parking garage. Finally, the relative weights of the building outside interface show that all apartment alternatives have the same priority of 10% each. This is due to the fact that all of the buildings have the same natural stone interface from outside. The AHP model shown in Figure 2 combines the weights of each alternative. Table 2 shows the total weight obtained for each alternative. According to the calculations in Table 2, apartment 3 is the most preferred by customers since it has the maximum total weight of 12.7%. Apartment 9 is the next alternative with a total weight of 10.9%. Apartment 1 has the minimum total weight according to customer rating of 8.1%, meaning that it is the least preferred. Table 2. Total weights of the considered apartments based on the AHP model. Apartment alternative Total weight A1.081 A2.104 A3.127 A4.095 A5.090 A6.104 A7.103 A8.085 A9 A10.102

14 M. S. OBEIDAT ET AL. In summary, 10 apartments are compared in this study based on an AHP approach according to customers rating of criteria. These criteria are the apartment location, the design, the building design and the financials. Based on pairwise comparisons between these criteria, the relative weights for each criterion and each apartment are determined. Based on these relative weights, the total weight of apartment alternatives is calculated and the results are shown in Table 2. The location of the apartment is the greatest factor of interest of customers, followed by the design of the apartment, the financials and the building design. Considering the location sub-criteria, the availability of transportation is the most preferred criterion by customers. For the apartment design sub-criteria, the area of the apartment has great interest of customers. Number of rooms has a very close relative weight value to that of the apartment area, making it the second preference of customers. Two sub-criteria are considered for financials, the price and the method of payment, with the price overwhelming the method of payment factor in importance. For the building design, sewage connection is the most important factor to customers. Taking into account all the results for criteria and sub-criteria and other factors being considered, apartment 3 has the greatest total relative weight and is considered the best choice for customers. The second best choice is apartment 9, followed by a tie between apartments 2 and 6, then apartment 7, apartment 10, apartment 4, apartment 5, and apartment 8. The least favourable apartment is apartment 1. Table 1 provides the apartment descriptions. Apartment 3 s description is summarised as follows. The location is away from villas, schools and markets. It is close to worship places, has a side to a local street, transportation is available in the area, and the population density in the area is low. Regarding the apartment design, apartment 3 is a second storey apartment, located on the west side (direction) of the building, does not have a ready kitchen, has one balcony, has super deluxe finishings in an American style, the area of the apartment is 150 m 2, and the kitchen has a moderate area. It has 3 bedrooms and one of them at least is a master, two bathrooms, two living rooms, does not have storage or maid rooms, and the additional water tank has a volume of 2 m 3. The building specification is as follows: it is a new building, connected to sewage, has an elevator, each storey has three apartments, one entrance with stone outside interface, and parking is provided. The price of this apartment is JD 50,000 for cash buyers, which is about $70,500. Details of other apartments are shown in Table 1. Changes in the relative weights of the criteria used may impact the decision process of selecting an apartment to purchase. For this reason, the sensitivity of the decision (recommending apartment 3) to changes in the priorities must be considered. Expert choice software is used to obtain a performance sensitivity graph for the baseline decision in the implemented AHP model as shown in Figure 3. The performance sensitivity graph is a dynamic graph consisting of two axes. The horizontal axis is the alternatives axis that measures the total weight of each alternative as the priorities given to criteria change; the objective axis on the left (vertical) of the graph depicts the relative importance of each main criterion considered. Figure 3 shows that apartment 3 has the highest priority for the location and financials criteria, the second priority for the apartment design criterion, and the fourth priority for the building design criterion. To perform a sensitivity analysis, priorities of the selection criteria are randomly changed, and then changes in the total weights of the apartment alternatives are monitored. Figure 4 shows an example of changing the priorities to 70, 10, 10, and 10% for location, apartment

JOURNAL OF PROPERTY RESEARCH 15 Figure 3. Performance sensitivity graph for the AHP model. Figure 4. Performance sensitivity graph for the AHP model with relative weights of 70, 10, 10, and 10% for the location, the apartment design, the financials and the building design. design, financials and building design. As shown in Figure 4, apartment 3 remains the best choice under these new priorities. Changing priorities for criteria do not change the decision unless a much higher priority is given to apartment design over the three other criteria. This situation is depicted in Figure 5 where the priorities are divided into 70, 10, 10, and 10% for building design, location, apartment design, and financials. Figure 5 shows that apartment 6 is the best selected alternative according to this new set of priorities. In practical, in this research, it is not common to assign a relatively high priority (exceeding 70%) to the building design; hence, apartment 3 is the most preferred apartment by customers.

16 M. S. OBEIDAT ET AL. Figure 5. Performance sensitivity graph for the AHP model with relative weights of 70, 10, 10, and 10% for the building design, location, apartment design, and financials. The main assumption of the AHP approach is that the criteria are independent at each level of the hierarchy. This means that criteria or sub-criteria are independent of each other. Independency of criteria at each level of the hierarchy is assumed throughout the results and the discussion. The objective of this paper is to consider comparison criteria that are independent of each other while building the model that is shown in Figure 1. This assumption of independency seems to be valid for the criteria and sub-criteria used throughout the model constructed for selecting an apartment to purchase. There is another tool called the Analytic Network Process (ANP) that can be used with models of clear dependency. The ANP is a general class of decision-making tool that deals with interdependency between the criteria, the sub-criteria, and the alternatives (Abu Qdais & Alshraideh, 2016). Readers may refer to (Saaty & Vargas, 2006) for additional details of the ANP decision tool. The utility maximisation problem in microeconomics is faced by a consumer of any product. This theory deals with how consumers must spend their money to maximise the utility. Several constraints are considered as major factors in this theory based on the case. Examples of these constraints include but are not limited to income amount, budget, commodity prices, and others. As discussed earlier, the AHP is a structured technique for organising and analysing complex decisions, considering psychology and mathematics (Venkataraman et al., 2014). In our case, the AHP dealt with four main criteria related to apartment selection: location, apartment design, apartment building, and financials. The AHP technique is used to choose the most suitable apartment considering the main criteria and related sub-criteria. Applying the AHP in this paper resulted in recommending apartment 3 to customers. In this paper, the utility maximisation problem is considered to be one internal component of the AHP process in that it focuses indirectly on the financial criteria included in the AHP.

JOURNAL OF PROPERTY RESEARCH 17 6. Conclusions This paper discusses the process of selecting an appropriate apartment for interested customers by adopting a case study in the Jordanian real estate market. Ten apartments are compared based on buyers preferences using the Analytical Hierarchy Process (AHP), which is considered one of the most powerful multi-criteria decision-making approaches. The compared apartments are chosen after investigating the real estate market by interviewing investors in this sector and by considering the needs of people. Data used to obtain the priorities were collected from 305 customers who were looking for an apartment in Jordan. The overall weights of the ten alternatives are calculated, and apartment 3 is found to be the best alternative. The results show that location and finance are the highest priorities of customers, followed by apartment design; the least is building design. In addition, for this decision, sensitivity analysis is conducted because priorities may change with time or for a new real estate market. Some factors such as road conditions and classifications are not addressed in this study, but the authors believe that such factors may have a stronger influence on home (villa) buyers than apartment customers. This paper provides a systematic approach that helps customers who are searching for an apartment select the appropriate apartment that satisfies their requirements. The main criteria in this paper (location, apartment design, apartment building, and financials) are the same for apartment customers around the globe. This paper will assist individuals who are searching for an apartment in the region or the world in selecting the appropriate apartment. This will help customers improve their decision-making capabilities by choosing the best apartment based on their requirements and preferences. Disclosure statement No potential conflict of interest was reported by the authors. Notes on contributors Mohammed Said Obeidat is an assistant professor in the Department of Industrial Engineering at Jordan University of Science and Technology. He obtained his B.S. (2004) and M.S. (2008) degrees from Jordan University of Science and Technology, Irbid, Jordan, and Ph.D. from Kansas State University; all in industrial engineering. Tarek Qasim is an Assistant Professor in the Department of Industrial Engineering at Jordan University of Science and Technology. Dr. Qasim has combined academic, research, administration and industrial experiences. He worked in the industry for more than ten years and also achieved several research projects at the University of Western Australia (UWA) and at Jordan University of Science and Technology (JUST). Aseel Khanfar is a graduate student in the industrial engineering department at Jordan University of science and technology.