How to motivate homeowners to invest in sustainable renovation? Abstract: It is important to induce homeowners to invest in sustainable renovation. It has been shown that homeowners are willing to renovate for many other motives than energy and cost savings. It seems therefore important to integrate homeowners’ individual values in renovation decision making processes. A critical review of decision-aid methods, which are susceptible to be integrated in a process of home renovation, is presented in this paper. Methods have been primarily classified on the basis on whether or not they allow incorporating truthfully decision-makers’ individual values while still improving the process effectiveness. Concepts such as constructivism, interactive and iterative process, quantitative attributes and values based on qualitative judgments, simplicity and accuracy of the methods, time use and modularity are discussed. It is concluded that multi-criteria decision making approaches using interactive and constructivist methods are the most relevant methods to use in a renovation process dealing with building nonexperts as decision-makers.

Keywords: sustainable home renovation, multi-criteria decision making, decision-aid, homeowners’ individual values.

Introduction At the start of any home renovation project, homeowners considering carrying out a major renovation work are faced with the problem of deciding which renovation strategy they should go for. At this stage, many factors influence their decisions and homeowners, who are building nonexperts or laymen, find inspiration, support and guidance from various parties such as architects, energy consultants, building engineers, material suppliers (building experts) or friends. During that phase of the renovation project, they see and meet a myriad of barriers which can easily lead them towards the selection of a poor renovation strategy or a strategy targeting only one or two goals. Though, homeowners’ preferences, wishes and behaviours are generally much more complex than that. Studies have shown that energy conservation measures as well as production of renewable energy can actually be costeffective until a certain extent [1-3]. Some methods and tools, available on the market, allow determining the most cost-effective energy conservation measures and renewable energy production systems for a specific application [4,5]. Studies have shown that an optimised energy auditing process can considerably improve the level of energy efficiency of a renovated building and can produce reasonable estimates of energy savings [6-9]. To estimate energy savings reasonably, it is important to assess the energy consumption of buildings to a certain level of accuracy. Yet, investigations reveal significant uncertainties in the estimation of energy consumption in buildings, the most sensitive parameters being directly related to occupants’ behaviours [10-12]. Although studies have shown the link between energy efficiency of buildings and behaviour of the building occupants, indoor environment quality, architecture and impact on the environment, it seems that most approaches to “energy renovation”, methodologies and tools are still focusing predominantly on the resulting energy and cost savings. It appears that such a narrow focus is not justified. In the case of homes, it has been shown in reviews [6] that homeowners are willing to renovate for many other motives than energy and cost savings. In order to overcome barriers to sustainable renovation, it seems therefore important to approach the issue in a more holistic manner, to support 1

homeowners’ decision making and to integrate in the process the occupants’ behaviours as well as any other factors which might directly influence them. Therefore, we have reviewed decision-aid methods with the purpose of selecting the most adapted methods to support and induce homeowners to invest in sustainable renovation. Decision-aid methods The presence of several independent and conflicting evaluation criteria, either qualitative or quantitative, makes the problem of selection of a sustainable renovation scenario a candidate to be solved by multi-criteria decision-making (MCDM) methods. The purpose of MCDM methods is to identify the most appropriate solution out of a set of alternatives which are characterised by multiple and possibly conflicting attributes. Interesting reviews of multicriteria decision making methods have been done by Hwang and Yoon [13], Chen and Hwang [14], Yoon and Hwang [15], Salminen et al. [16], Andresen [17] and Hopfe [18]. Most multicriteria approaches have advantages, disadvantages and pertinent fields of applications. Those methods vary especially according to several parameters: type of problem to resolve (structured, semi-structured, and unstructured), type of decision maker and surrounding (evaluator expertise), types and number of criteria to be taken into account, types of scales of preferences to be used, how close to the human behaviour we want the prescriptive method to be. We have sorted the approaches within two families: the decision-aid expressed by a single synthetizing criterion (SSC) and the decision-aid based on preference or human behaviour models. Decision-aid expressed by a single synthetizing criterion (SSC) The first type of deterministic SSC approach is the multi-attribute decision making approach (MADM). In a single synthetizing criterion or a value function approach, each of the scenarios is evaluated with respect to the criteria in order to obtain scores also called value functions. This is done by using a rating scale. Scaling factors or weights give the relative importance of each of the criteria. Once the information obtained, it is then synthesised and aggregated to give an initial overall evaluation for each scenario. The simplest deterministic method is the weighted sum method (WSM); it is generally applied in single dimensional problems. Even though green or sustainable building design is not a single dimensional problem, green building rating schemes such as LEED (LEED for Homes for new homes or green residential remodelling guidelines for existing homes), BREEAM (Code for Sustainable Home for new homes and BREEAM Domestic Refurbishment for existing homes) or DGNB (small residential buildings for new homes) have all applied the weighted sum method. Another deterministic method with simple implementation is the simple multi-attribute rating technique (SMART). It was developed by Edwards [19] based on the Multi-attribute Utility Theory (MAUT). MAUT uses utility functions, whereas SMART uses value functions. A utility function differs from a value function for the reason that it takes into account the attitude to risk. As a consequence, the value functions are much simpler than the utility functions; furthermore the evaluation process is less complex. The most widely applied SSC method is the analytical hierarchy process (AHP) or its generic form; the analytical network process (ANP), developed by Saaty [20]. In this theory, the problem is formulated in the form 2

of a hierarchy and the decision-makers judgement in the form of pairwise comparisons. Other examples of SSC methods are COPRAS [21] and TOPSIS [13]. While multi-attribute decision making SSC methods can be used in an iterative and constructivist manner (i.e. constructing progressively, actively and collaboratively knowledge), and even if the weighted sum method (WSM) would be the simplest approach to be understood by nonexpert decisionmakers; these methods do not incite towards an improvement of the communication around value systems and preferences. They do not use either a language in phase with the decision maker. Indeed, it has been demonstrated in multiple occasions that decision-makers tend to have difficulty in ascribing numerical values to their subjective interpretations or preferences. In addition, hesitation and knowledge acquisition are not easily introducible. SSC methods accept quantitative attributes but values based on qualitative difference judgements need to be directly quantified to relate with a synthetized function. Such a function models decisions in a very rigid and rational thinking approach. Most decision makers are very likely to be only partly rational, and be irrational in the remaining part of their actions. Furthermore, the construction of such a function is a difficult problem and requires a lot of information from the decision maker in a very short lap of time. The second approach of methods within the deterministic SSC family is the multi-objective decision making (MODM). While all the previously cited methods are multi-attribute decision making methods (MADM) and focus on the decision process itself in order to create something which is viewed as liable to help an actor taking part in a decision process either to shape, and/or to argue, and/or to transform the decision maker’s preferences [22], MODM methods focus on identifying a preferred alternative from a theoretically infinite set of alternatives. The alternatives are therefore not defined explicitly, but rather implicitly by a set of constraints and indirectly by the pursuit of the objectives. However, multi-objective decision making methods (MODM) are difficultly constructivivist and computations are normally fully automatic. Interactive MODM methods exist but lack of simplicity, they are quite fastidious, time consuming and require lots of expertise knowledge. It is therefore not adapted to nonexpert decision makers. Another type is the non-deterministic SSC methods. They include game theoretical models, multi-attribute utility models (MAUT) [23] , methods based on the fuzzy and grey systems theories: fuzzy AHP [24], fuzzy TOPSIS [25], fuzzy COPRAS [26] , fuzzy multiple-criteria complex proportional evaluation [27] as well as the fuzzy linguistic methods. While in deterministic methods, change of parameter values are not considered during the decision making process (possibly after the search via a sensitivity or robustness analysis), in the non-deterministic decision making approach, the parameter values are not fixed, instead, sources of uncertainty need to be identified and parameters are treated with a risk factor. Thanks to these methods, experts can deal with uncertainties and imprecision they identify. However, most cons of deterministic SSC methods persist. Decision-aid based on preference or human behaviour models The outranking theory developers did not accept the fact that all alternatives are comparable. They believed that in some circumstances decision-makers do not want or are not capable of comparing all the alternatives. They refused a total compensation between criteria. Instead 3

they used the concept of incompatibility to avoid arbitrary or fragile judgements. It is therefore why they presented the output not as a value like in the single synthetizing criterion approach but rather as an outranking graph or a partial aggregation. The partial aggregation allows the possibility of ranking, sorting and ordering the possible actions while this is not feasible with the aggregate value function approach. There are two prominent types of outranking methods [28]; ELECTRE (I to IV, IS & TRI) and PROMETHEE (I & II). An outranking relation is built in order to enable the comparison of an alternative to a pre-defined profile. This approach differs from standard classification approach because the categories considered are defined a priori and do not result from the analysis. Claude-Alain Roulet et al. [29] were inspired by ELECTRE IV [22] during the development of the ORME method. To observe how the weights modify the resulting modifications of ranking, a sensitivity analysis module embedded in the tool needs to be used. Those methods have succeeded to overcome values previously forgotten by introducing the notion of incomparability and the refusal for total compensation between criteria. Furthermore, outranking methods generally require less information from the decision maker. While most of those methods have evolved to solve recurring issues, the last versions of those methods have become very complex. This means that the decision maker may feel like having a loss of control and has most of the times no choice than trusting the decision aiding expert. The success of those methods goes therefore through the need of having a high trust between the decision maker and a highly skilled expert. Another approach is the knowledge based approach. It is applied on rules informed by past experience. An example is the application of rough set theory by Roman Slowinski [30]. Results are particularly promising when a database of past decisions is available. However, such a database and a highly skilled expert are not always available. In an ideal world, descriptive and prescriptive approaches to decision making would follow a harmonious relationship: human behaviour would lead the rules of decision-making, and decision theorists, analysts or experts would use those descriptions of behaviour and a natural and adapted language to construct or select a decision support or aid system the most adapted to the situation [31]. Based on a qualitative model of the human decision maker, Larichev hence proposed to adapt decision methods to human behaviour in ZAPROS III method [31-33]. However, this human behaviour based qualitative approach does not exactly suit the needs for the solvation of our problem. Indeed, it is a purely qualitative approach which has been developed for unstructured problems. Our problem would be more defined as a semistructured problem with both quantitative attributes and values based on qualitative verbal difference judgements. The second issue is that the approach is quite complex and time consuming. The number of incomparable alternatives is often too high and the approach not robust enough. The last approach is the interactive and constructivist approach. It includes the MACBETH method. It is based on the additive value model but requires only qualitative judgements about differences of value. In this method, the quantitative model of values is based on qualitative verbal difference judgements. This minimizes the complexity for the decision maker. The MACBETH method has very strong mathematical foundations, is interactive, transparent and constructivist. Moreover, the method avoids the need of having the decision-maker answering very complex judgmental questions. The method complies with 4

most of requirements sensible for homeowners: it is an iteratively constructivist, interactive and humanistic approach and still mathematically accurate; it accepts conflicting balanced criteria of importance with different types of value scales; it accepts both quantitative attributes and values based on qualitative verbal difference judgements and it allows estimating or evaluating modularly the impact of an action. However, as for most of the outranking methods, the method is quite complex and requires the participation of an expert with good knowledge about the method. This approach also includes the Hermione method. It is a qualitative multi-criteria method [34]. The particularity of Hermione is that, instead of attempting to model the decision maker preferences and then substitute his or her judgement in a model, the method focuses on structuring the problem so the decision maker can directly interact with it and aggregate in his or her mind the various influencing factors. This method can deal equally with both quantitative and qualitative aspects. The most important strength of this method is it has been developed towards avoiding this feeling of loss of control from the decision maker. The main idea is to escape a delegation of the ability to perform a global judgement or preferences to a complex mathematical model, and a need of having a high skilled and trustworthy expert. This, however, requires that the decision-maker or the decision-maker plus some aiders is or are able to structure his or their problem, as well as that they are informed enough or have enough knowledge to evaluate all the criteria. The construction of a new decision criteria structure is work intensive and must comply with a set of requirements. If the requirements are not respected, the method will not be used correctly. In practice, this can lead to confusion and to mistake results if the decision-maker or any aider does not respect coherently the method requirements. Nevertheless, the method has found a good balance between simplicity and limited loss of information. Such a method will work particularly well when there is a fair balance between quantitative and qualitative information. Discussion Homeowners who have been living in their home for some time are usually more aware of what make them feel satisfied or dissatisfied in the existing conditions of their home. They are often sentimental about the building because they have experienced the place in different manners. The situations of building experts making decisions and of a building expert dealing with building nonexperts as main decision makers are very different. Indeed, in the second situation, the decision making process becomes much more personal, emotional and therefore the decision makers have less tendencies of being rational in their actions and decisions. When dealing with homeowners as decision makers, it is therefore even more important to ensure that the decision aid approach used follows a harmonious relationship with the homeowners’ behaviours, values and preferences. Still, the method also needs to be accurate. The qualitative and constructivist Hermione method has shown to be interactive and humanistic, quite simple to understand, relatively accurate, accepting conflicting and balanced criteria with different types of value scales, accepting both quantitative attributes and values based on qualitative judgements and time-effective. Even though it needs the presence of a building expert (or at least of an advanced tool interface), we recommend giving a priority of use to the Hermione method. If it does not lead to the distinction of a most favourable 5

renovation scenario within the alternatives which fulfil the objectives, we suggest that the expert presents to acceptable alternatives to other experts participating to the building renovation project and discuss the issue. The use of the MACBETH method could be another option. The paper has presented decision-aid methods which could allow a human-based customization and a possible integration of nonexpert decision-makers in a sustainable home renovation process. Yet, these methods need to be adapted and combined with properly selected tools in order to be able to deal with building nonexperts as main decision makers. Acknowledgment This paper is based on research conducted in a Ph.D. project which is part of the Strategic Research Centre for Zero Energy Buildings at Aalborg University and financed by Saint Gobain Isover A/S, Aalborg University, the Danish Council for Strategic Research (DSF), and the Programme Commission for Sustainable Energy and Environment. References [1] Kragh, J., Rose, J. (2011). Energy renovation of single-family houses in Denmark utilising long-term financing based on equity. Applied Energy, 88(6), 2245-2253. [2] Amstalden, R. W., et al. (2007). Economic potential of energy-efficient retrofitting in the swiss residential building sector: The effects of policy instruments and energy price expectations. Energy Policy, 35(3), 1819-1829. doi:10.1016/j.enpol.2006.05.018. [3] Martinaitis, V., et al. (2004). Criterion to evaluate the "twofold benefit" of the renovation of buildings and their elements. Energy and Buildings, 36(1), 3-8. doi:10.1016/S03787788(03)00054-9. [4] Christensen, C., et al. (2006). BEopt™ software for building energy optimization: Features and capabilities. USD O.Energy.Golden, Colorado, National Renewable Energy Laboratory. [5] Intelligent Energy Europe. (2008). E-RETROFIT-KIT, tool-kit for "passive house retrofit". Retrieved, 2011, from http://www.energieinstitut.at/retrofit/Dateien/Startseite/E-RETROFITPopular.pdf. [6] Gram-Hanssen, K., Christensen, T. H. (2011). Improving the energy labelling scheme. ( No. SBi 2011:23).Statens Byggeforskningsinstitut, SBI. [7] Laustsen, J. (2000). Energy labelling of buildings in denmark. Danish Energy Agency, 1-20. [8] SQUARE. (2010). A system for quality assurance when retrofitting existing buildings to energy efficient buildings – WP6 – national pilot project in Finland – final report. [9] Wierzba, A. L., et al. (2011). A study to optimize the potential impact of residential building energy audits. Energy Efficiency, 4(4), 587-597. doi:10.1007/s12053-011-9106-x. [10] Brohus, H., et al. (2010). Influence of occupants’ behaviour on the energy consumption of domestic buildings. Antalya: Clima 2010 : 10th Rehva World Congress. [11] Koiv, T., et al. (2010). Indoor climate and energy consumption in residential buildings in estonian climatic conditions. WSEAS Transactions on Environment and Development, 6(4), 247256. [12] Larsen, T. S., et al. (2010). Occupants influence on the energy consumption of Danish domestic buildings. ().Aalborg University. Department of Civil Engineering. [13] Hwang, C. L., Yoon, K. (1981). Multiple attribute decision making: Methods and applications: A state-of-the-art survey Springer-Verlag New York.

6

[14] Chen, S. J. J., et al. (1992). Fuzzy multiple attribute decision making: Methods and applications Springer-Verlag New York, Inc. [15] Yoon, K., Hwang, C. L. (1995). Multiple attribute decision making: An introduction Sage Publications Thousand Oaks, CA. [16] Salminen, P., et al. (1998). Comparing multicriteria methods in the context of environmental problems. European Journal of Operational Research, 104(3), 485-496. [17] Andresen, I. (2000). A multi-criteria decision-making method for solar building design. Norwegian University of Science and Technology, Faculty of Architecture, Planning and Fine Arts, Department of Building Technology. [18] Hopfe, C. J. (2009). Uncertainty and sensitivity analysis in building performance simulation for decision support and design optimization. Technische Universiteit Eindhoven. [19] Edwards, W. (1971). Social utilities. Engineering Economist, Summer Symposium Series, 6 119-129. [20] Saaty, T. L. (1990). How to make a decision: The analytic hierarchy process. European Journal of Operational Research, 48(1), 9-26. [21] Zavadskas, E. K., et al. (1994). The new method of multicriteria complex proportional assessment of projects. Technological and Economic Development of Economy, 1(3), 131-139. [22] Roy, B., Bouyssou, D. (Eds.). (1993). Aide multicritère à la décision: Méthodes et cas. [23] Keeney, R. L., et al. (1979). Decisions with multiple objectives: Preferences and value tradeoffs. Systems, Man and Cybernetics, IEEE Transactions On, 9(7), 403-403. [24] Van Laarhoven, P., Pedrycz, W. (1983). A fuzzy extension of saaty's priority theory. Fuzzy Sets and Systems, 11(1), 199-227. [25] Wang, Y. M., Elhag, T. (2006). Fuzzy TOPSIS method based on alpha level sets with an application to bridge risk assessment. Expert Systems with Applications, 31(2), 309-319. [26] Zavadskas, E. K., et al. (2009). Multi-attribute decision-making model by applying grey numbers. Informatica, 20(2), 305-320. [27] Zavadskas, E. K., Antucheviciene, J. (2007). Multiple criteria evaluation of rural building's regeneration alternatives. Building and Environment, 42(1), 436-451. [28] Belton, V. (1990). Multiple criteria decision analysis–practically the only way to choose. ( No. 1990).Operational Research Society Birmingham. [29] Roulet, C., et al. (2002). ORME: A multicriteria rating methodology for buildings. Building and Environment, 37(6), 579-586. doi:10.1016/S0360-1323(02)00005-7. [30] Slowinski, R. (1992). A generalization of the indiscernibility relation for rough set analysis of quantitative information. Decisions in Economics and Finance, 15(1), 65-78. [31] Larichev, O. (1999). Multicriteria decision making: Advances in MCDM models, algorithms, theory and applications. In T. Gal, T. J. Stewart & T. Hanne (Eds.), (pp. 5-1-5-24) Springer. [32] Larichev, O. (1992). Cognitive validity in design of decision‐aiding techniques. Journal of Multi‐Criteria Decision Analysis, 1(3), 127-138. [33] Larichev, O., & Moshkovich, H. M. (1997). Verbal decision analysis for unstructured problems Kluwer Academic Publishers Boston. [34] Flourentzou, F., et al. (2003). Hermione, une nouvelle méthode d'agrégation qualitative basée sur des règles. 58èmes Journées du groupe de travail Européen d'aide multicritère à la décision, Moscow.

7