Constructing a Mixed Methods Research Design: Exploration of an Architectural Intervention

Introduction

A mixed methods research (MMR) design can be used for design and planning studies. Therefore, we demonstrate how MMR was employed to answer a complex inquiry requiring the use of both qualitative and quantitative methods in an architectural intervention. Creswell and Plano-Clark (2011) noted that despite the developments in applying this form of research within different disciplinary contexts, many questions remained unanswered. Thus, the aim of this methodological description is to illustrate the application of MMR in architectural design using a hybrid model consisting of a taxonomy development model and an embedded quasi-experimental model. We describe the purpose and procedure of our MMR design to bring awareness to the use of mixed methods in a “discipline unique” context.

Hitherto, the application of MMR design in the field of design and planning has not been fully explored. In their book, Architectural Research Methods, Groat and Wang (2002) used the term “combined strategies” (p. 361) to describe the incorporation of multiple methods from diverse traditions into one study. They noted that the mixed methods design signifies the most comprehensive integration of two or more research designs. Furthermore, they indicated that “architectural research that combines strategies represents an important and necessary frontier” (p. 370) in the field of architecture. In any case, despite the numerous strengths and opportunities offered by MMR design, it has been adopted by few architectural researchers (Yeap, Yaacob, Rao, & Hashim, 2012; Youmans, 2011). Distinct research designs are frequently taken for granted as the preferred method for architectural research (Groat & Wang, 2002). For example, the case study method is frequently used as an important way to study precedents in architectural design; this method allows for categorizing and identifying the type and describing the function of a particular building, enabling the designer to better visualize the new design. Traditionally, architectural design was established from disciplines in the arts and humanities and only recently have the sciences and social sciences shown more prominence in influencing architectural design. Many researchers in the field of architecture have only been exposed to certain research strategies without realizing the enormous opportunity offered by MMR design.

We aim to shed some light on this situation by offering an insight into how a research team consisting of research architects overcomes the methodological challenges in formulating an MMR design. By publishing the procedure in constructing the MMR design, we hope to promote the development of awareness of and the skills required to use these methods within the architectural disciplinary context (Vrkljan, 2009). Consequently, this methodological discussion should be viewed as a catalyst that advocates the use of MMR design in the field of design and planning, furthermore demonstrating how MMR design could be adopted to create meaningful architecture capable of guiding primary school students toward ecological literacy. The role and application of both qualitative and quantitative approaches is highlighted in our article, exploring the complex environment–behavior relationship.

The following discussion consists of six sections. First, an overview of the study, including the research problem and question, is presented to provide an appropriate background for this methodological discussion. Second, the rationale for using the MMR design is discussed. Third, the research design is presented to illustrate how we arrive at a hybrid mixed methods design. Subsequently, we briefly summarize the data collection and analysis. After that, the discussion outlines the challenges involved in conducting a hybrid MMR design, including the key strategies to overcome these challenges to enlighten future research. Finally, important findings and suggestions are summarized in the conclusion.

Overview of the Study

Physical settings are essential in shaping the perception, performance, and behavior of students (Boeve-de-Pauw & Van-Petegem, 2011; Dresner & Gill, 1994; Leeming, Dwyer, Porter, & Cobern, 1993). Beyond the functional backdrop of activities, the design of a learning environment can be harnessed as a tool for teaching and imparting values. Architects therefore have the opportunity to use creative design to promote pro-environmental behavior among students (Dietz, Mulford, & Case, 2009; Duerden & Witt, 2010). A three-dimensional (3-D) textbook (Orr, 1993; Taylor & Enggass, 2009) has been designed to encourage and inspire students to act in a more sustainable manner. The 3-D textbook is the result of multidisciplinary collaboration between architecture, psychology, and education and has been developed significantly in recent years, particularly by environmental psychologists (Abrahamse, Steg, Vlek, & Rothengatter, 2007; Arbogast, Kane, Kirwan, & Hertel, 2009; Halpenny, 2010; Kaiser, Oerke, & Bogner, 2007; KasalI & Dogan, 2010; Ozdemir & Yilmaz, 2008).

A 3-D textbook is defined as the physical environment and the real-world objects within it that can be used as teaching and learning tools for understanding phenomena usually studied from textbooks (Taylor & Enggass, 2009). Although the notion of a 3-D textbook promises many potential benefits for environmental education (EE; Higgs & McMillan, 2006), research on this subject is rather limited (Taylor & Enggass, 2009). To date, there has been no systematic effort to explore the attributes that characterize a 3-D textbook from students’ point of view. Furthermore, Mitchell (2005) and Taylor and Enggass (2009) noted that empirical evidence is required to identify learning outcomes specific to the 3-D textbook. Further studies are needed to scrutinize the contributions of the 3-D textbook in promoting pro-environmental knowledge (EK), attitudes (EA), and behavior (EB), particularly in how the child–environment interaction could be a possible source of environmental learning (Dresner, 1990; Dresner & Gill, 1994; Linn, Vining, & Feeley, 1994; McNeill & Wilkie, 1979; Padua & Jacobson, 1993; Ramsey, 1993; Ramsey, Hungerford, & Tomera, 1981).

From the background information, we highlighted two potential areas of research. First, “What are the design features of a 3-D textbook?” The first inquiry was made due to the lack of information and investigation in this area of study (Orr, 1993; Tanner, 2008). Although previous researchers have illustrated some of the content and benefits of a 3-D textbook, scholars and designers have not explicitly described all the design features (Ford, 2007; Orr, 1997; Taylor & Enggass, 2009). From the review of the literature, passive design and green technology are relevant factors, but other criteria and strategies may also shape the 3-D textbook (Higgs & McMillan, 2006). A response to the first inquiry would provide a series of design features for further testing and validation, leading to the second inquiry: “Can a 3-D textbook enhance EE outcomes?” Although some scholars have claimed that a 3-D textbook is a useful educational tool, the effectiveness of architectural intervention has not been evaluated in the literature (Tanner, 1974; Taylor, 1993). Thus, the impact of a 3-D textbook on students’ EK, EA, and EB is unknown.

Rationale for Using a Mixed Methods Research Design

In light of the literature review, we identified a theoretical framework that could provide direction for the present study (Figure 1). This theoretical framework consists of an a priori component based on the existing literature and an a posteriori component informed by a qualitative study. We assumed that forming a complete theoretical framework preceding all data collection is not possible for several reasons. First, this field of study is relatively new and unexplored (Taylor & Enggass, 2009). Thus, there is insufficient information to develop the design features from the literature. Second, Creswell (2008) noted that any exploratory or inductive research should not be restricted or directed by the literature. Instead, the literature should be used as an aid once patterns or themes have emerged from the data. The present study was designed with Creswell’s suggestion, addressing the complexity of design problems and issues by combining qualitative and quantitative research in parts (Buck, Cook, Quigley, Eastwood, & Lucas, 2009; Fielding, 2012). MMR design was identified as the preferred methodology for the present study due to the following reasons.

OPEN IN VIEWER

Figure 1.

Theoretical framework.

First, qualitative information is required before quantitative research can be performed for the intervention to occur. Creswell (2007) stated that MMR design is the most appropriate method for the type of problem requiring qualitative exploration at the first stage of research. Nastasi et al. (2007) proposed that the best research procedure for developing an intervention should involve qualitative data collection preceding quantitative research. In this process, the intervention development is guided by qualitative methods to generate formative data, followed by a quantitative evaluation that tests the effectiveness of the intervention. Therefore, the present research process begins with a collection of qualitative data to ascertain and describe the design features specific to a 3-D textbook. The collection and analysis of qualitative data may help uncover physical and environmental conditions that can support and enhance EE.

Second, we intend to generate and test the design features of a 3-D textbook by combining the qualitative and quantitative approaches. One type of evidence from one method is not comprehensive enough to answer the main research question sufficiently. By combining the qualitative and quantitative data gathering techniques, we can gain clarity on the subtleties of the data, cross-validate the findings, and justify the planning, as well as implement and assess the design strategies. In particular, the MMR design enabled us to assimilate and present the qualitative and quantitative findings effectively (Kington, Sammons, Day, & Regan, 2011). This is in line with the methodological literature that suggests adopting a “fitness for purpose” approach in the selection of research methods (Bryman, 2007). Combining different methods enables the amalgamation and synthesis of data to provide convincing evidence for the research (Johnson & Onwuegbuzie, 2004). It also reveals the key factors that influence the design of a 3-D textbook and the relationship between architecture and EE.

Finally, we attempt to offset the weaknesses of both quantitative and qualitative research in the present study by adopting an MMR design. We used an initial qualitative approach to generate design criteria for the 3-D textbook. We acknowledged that it is difficult to generalize the qualitative findings to a larger group with a limited number of interviews in a single case study; therefore, we used a quantitative approach in Phase 2 to assess the generalizability of the findings. Additionally, we embedded the qualitative data method within a quasi-experiment in Phase 2 to examine the intervention qualitatively in addition to the quantitative results. Qualitative data were collected during and after the intervention trial to clarify the results of the intervention. By examining participants’ perspectives during their intervention trial, we can better realize the findings of the intervention study. In short, the amalgamation of qualitative and quantitative data offers greater and more comprehensive evidence for studying the 3-D textbook. This is in accordance with methodology studies that promote a mixture of qualitative and quantitative methods to minimize the risk of overreliance on one method of data collection (Pintrich & Schunk, 2002).

Mixed Methods Research Design

Research Design

The present study began with an exploratory qualitative approach followed by a confirmatory quantitative study (Figure 2). Christ (2007) stated that this sequence is a logical choice when researchers wish to generate and test a theory or intervention. Thus, the present study was spread across two phases. Phase 1 was conducted using a qualitative case study that involved interviews and observations to establish the design features of the 3-D textbook. A key outcome of this phase was the description of the 3-D textbook, defining the variables and abilities in developing the intervention for further testing and validation. We envisioned that one or more design features are possible, thus remaining open to possibilities during this exploratory phase of the entire research process. McNabb (2010) defined Phase 1 as the exploration for the preliminary identification of potential variables. The initiation stage is characterized by an inductive theoretical drive with the purpose of logically guiding critical components, including the hypotheses, operating in the construction, and analysis of the subsequent quantitative study. Adopting multiple phases of an exploratory research project is a way of developing and testing a theory (Oliver, 2004). Oliver proposed that an initial problem will stimulate research from identified themes to be further tested in an MMR design.

OPEN IN VIEWER

Figure 2.

Research design.

Phase 2 is the stage identified by converging qualitative data within a quasi-experimental design. We placed great importance on a methodology that allows manipulation and a design that permits qualitative data to be used in the experiment. The data collected in Phase 2 that contributed to the effectiveness of the 3-D textbook were revealed, and the value of the qualitative findings was determined. The use of quantitative data concurred with the sequential exploratory design. The quantitative study facilitated the analysis of Phase 1 findings. Merging qualitative data sources in Phase 2 enabled us to better comprehend participants’ experience during the intervention trial, underlining the benefits of qualitative inquiry to fathom the child–place interaction and gain a more holistic understanding of the differences between the treatment and control groups. Robinson and Mendelson (2012) stated that the type of “qualitative experiment” (p. 333) is capable of merging two methods within a single study. In the present study, the “qualitative experiment” offers a way to investigate the effectiveness of the 3-D textbook and participants’ responses to various physical models of architectural design.

In summary, the methodology adopted was chosen based on logical reasons and was best suited to answer the research question, depicting how qualitative and quantitative methods can be adopted at different phases of a study to practically explore architectural design. The sequential processes highlighted can be used to analyze data that guide the research, enabling us to make constructive modifications in subsequent phases of the design referred to in the initial findings.

A Brief Overview of Data Collection and Analysis

Phase 1: A Qualitative Case Study

The Green School in Bali, Indonesia, has been selected as the case study for the present research. The school was selected because the objectives of the school’s establishment reflected recommendations by Orr (1993, 1997) and Taylor and Enggass (2009) for a 3-D textbook by using its physical environment (i.e., buildings and landscape) as an instrument for teaching and learning EE. The case study focuses on exploring the attributes that characterize a 3-D textbook from the students’ point of view, identifying and examining the contributions of the 3-D textbook in promoting EK, EA, and EB. Child–environment interaction was also explored as a possible source of contribution to environmental learning. The case study was used to collect data specific to the context of the Green School, rather than to collect information that can be generalized to a larger population (Zeisel, 1981). Thus, the specific context of the case becomes inseparable from the definition of the 3-D textbook (Yin, 2004). The case study provided multiple sources of evidence for the development of the architectural intervention. This evidence comes from two main sources, namely, observation and interview techniques (Figure 3). The use of the case study is based on Yin (2004), who stated that when any finding or conclusion in a case study is based on several different sources of evidence, it is likely to be much more convincing. Text and quotes obtained from the participants’ interviews were instrumental research data used in subsequent stages of the study to guide the design and construction of a physical model. Thus, the use of the qualitative method was justified, as the texts contained decisive connotations on the phenomenon of architectural design.

OPEN IN VIEWER

Figure 3.

Phase 1: A qualitative case study.

An on-site observation technique was carried out to record physical evidence and environmental behavior. The indoor and outdoor spaces in the Green School were examined to understand the design decisions, in particular the relationship of a designed space constructed by the builder, to meet the needs of users. Observations were carried out in the outdoor spaces (garden, farming plot, etc.), while nonstationary observations were conducted in indoor spaces (science lab, classroom, etc.; KasalI & Dogan, 2010). We distinctively looked at how a physical setting supported or countered the EE activities taking place to generate data about the behavioral opportunities and constraints of the environment.

In-depth interviews were conducted with selected fifth-grade students. Discussions and brief interviews were also carried out with selected teachers, administration staff, and the project architect of the Green School to support and clarify the evidence collected from the students. Prior to the interview, selected fifth-grade students were asked to draw two pictures illustrating their perceptions of the school and how they would like the school to be changed. The interview questions were designed to be open ended to encourage students to express their experience as users. Whenever required, we would ask follow-up questions to achieve greater clarity on the topic discussed. An example of the interview questions below will explain how the interviews were conducted: “Please list all the places that you use in your school. What activities do you usually do at these places?”

All interview data were transcribed from the audio format to written text and eventually stored in a computer database (Creswell, 2008). Verbatim transcription was done manually without the aid of software. We analyzed the qualitative data by adopting the method of constant comparison (Glaser & Strauss, 1967) and the grounded theory approach (Miles & Huberman, 1994; Morrow & Smith, 1995). First, open coding was conducted to examine sections of text made up of individual words and sentences (Creswell, 2008). An overall grouping system of the interviewees’ responses was developed. Second, the initial scheme was classified into fixed categories and subcategories to identify basic themes. The categorization represented similarity of responses (with regard to the design of educational settings) and frequency of responses. The transcripts and field notes were analyzed to search for regularly occurring terms or phrases and unanticipated counterintuitive material. Third, the basic themes were reviewed in relation to the existing design theory. Codes and categories were continuously sorted, compared, and contrasted until saturation occurred (Morrow, 2007; Munhall & Chenail, 2008). Consequently, the basic themes were combined and retitled into four main themes.

Although the findings of Phase 1 cannot be generalized too broadly, this case study is nevertheless an initial attempt to better comprehend the qualitative dimension of the child–environment relationship and how the physical environment can be a source of environmental learning. Phase 1 expanded our understanding on the use of architecture as a setting for EE. The emerging themes provide a distinct and important guide to define the design features of the 3-D textbook.

Phase 2: A Quasi-Experiment

The quasi-experiment phase was designed to test the effectiveness of the 3-D textbook by gathering numerical information and figures with statistical support to strengthen the findings from Phase 1. A pretest–posttest nonequivalent comparison group design (Babbie, 1992) was used to address the hypotheses and research question (Figure 4). The approach to examining the quasi-experiment was to use the physical model evolved from the findings of Phase 1. We assumed from the hypothesis that students who interacted with the physical model would demonstrate an improvement in their EK, EA, and EB compared with their peers who had not interacted, after controlling for the effect of pretest scores. The treatment group consisting of 42 fifth-grade students was allowed to design, build, and operate the physical model. On the other hand, 42 fifth-grade students who were not involved with the intervention trial and with characteristics, for example, gender, ethnicity, and academic performance, as similar as possible to the treatment group, were selected to serve as the control group.

OPEN IN VIEWER

Figure 4.

Phase 2: A quasi experiment.

Due to the lack of existing instrumentation, the authors designed the instruments in this study to assess students’ EK, EA, and EB. We administered the instruments to the participant and comparison groups as a pretest before the intervention and as a posttest after the intervention. To test the hypotheses, repeated measures using ANCOVA were conducted to compare the scores of the treatment and control groups regarding knowledge, attitudes, and behavior. The ANCOVA was also used to control for differences in pretest scores. Additionally, qualitative data were collected so that we could better interpret the quantitative data. The findings were analyzed to ascertain what participants actually experienced during the intervention trials. The sources of qualitative data were based on on-site observations, personal journal entries, and follow-up interviews. All interviews were recorded and transcribed. Field notes and personal journal entries were used as an aid in the analysis of the transcripts, guided by grounded theory methodology as outlined by Strauss and Corbin (1990) and the study’s research questions.

The results from the ANCOVA supported the hypotheses that the treatment group would experience a significant growth in knowledge, attitudes, and behavior with respect to the control group. Additionally, Phase 2 participants’ perceptions of what setting influenced their knowledge, attitudes, and behavior were consistent with the themes in Phase 1. Therefore, Phase 2 helped validate Phase 1 findings by providing empirical evidence that links Phase 1 findings with improved EE outcomes.

Overcoming Challenges in a Mixed Methods Research Design

A hybrid of the taxonomy development model and the embedded quasi-experimental model was adopted for the present study due to the complexity of the factors involved in exploring the design features of the 3-D textbook and the associated implications of using this architectural intervention on EE outcomes. A few challenges associated with the use of the mixed methods have arisen. Nevertheless, certain strategies were adopted to resolve them, as described in this section.

As stated by Creswell and Plano-Clark (2011), a sequential exploratory design with the objective of developing and testing an intervention would reveal what information is most useful in designing a treatment. The insertion of an Intermediate Phase into the process allowed for the two data collections to address those issues. The inserted middle phase between the two main phases enabled researchers to determine and select the most useful information for the last phase. In this present study, the qualitative data analysis yielded specific quotes from individuals, qualitative findings that were essential in the design and development of the quasi-experiment in Phase 2. For instance, the quotes from individuals were incorporated into the instruments to assess participants’ EK, EA, and EB. Additionally, the codes and themes were incorporated into the design of the physical model, completing the architectural intervention for the quasi-experiment.

Concurrently, the theoretical framework from Phase 1 was further developed and strengthened during the Intermediate Phase. Phase 1 findings helped uncover the undefined theoretical components, that is, design features, to direct the follow-up procedures in Phase 2. Additionally, the findings from Phase 1 were used as a guide to refine the research questions and hypotheses as part of the exploratory and pragmatic approaches. As argued by Creswell and Plano-Clark (2011), “the initial qualitative phase produces specific categories or relationships. These categories or relationships are then used to direct the research questions and data collection used in the second, quantitative phase” (p. 77). Therefore, the Intermediate Phase was incorporated into the overall procedures in this mixed methods study (Figure 2). This figure depicts the intervention development stage and how this Intermediate Phase connected the qualitative and quantitative procedures in this study.

A review of published MMR studies informed us of the procedures to be used to integrate the qualitative and quantitative data. Examples of these designs included a study by Nastasi et al. (2007) and Robinson and Mendelson (2012). The present study concurs with the suggestion of Nastasi et al. (2007) by adopting a Qual → Quan sequence in the intervention development, as well as with the recommendation of Robinson and Mendelson (2012) to integrate the participants’ interviews within an experimental study. Following a review of the methodological literature, it was decided that a hybrid of the suggestions of Nastasi et al. (2007) and Robinson and Mendelson (2012) could be used in the present study. An initial qualitative phase was employed to identify important design features of the 3-D textbook. A subsequent quantitative phase was used to test the qualitatively identified variables and examine the prevalence within a larger sample. Data collected from both phases helped confirm and enrich the content analysis by using MMR sequentially resulted in a fuller and richer picture of the environmental literacy in our subjects.

It is essential to include various validation strategies during data collection and analysis in an MMR design. Tashakkori and Teddlie (2003) argued that validity is the most important feature of a study. On the other hand, Creswell and Plano-Clark (2011) stated that researchers who adopt an MMR design need to report and discuss validity within the setting of both qualitative and quantitative methods. Thus, several procedures were implemented in the present study to validate the qualitative findings and the quantitative results. Phase 1 used a qualitative case study in which the data were collected through interviews and on-site observations. The text analysis was verified and interpreted by conducting peer review with interested colleagues and experts, including an architect and an EE practitioner. These reviewers examined the coding and data analysis during the “peer debriefing session” to confirm emerging themes and findings (Lincoln & Guba, 1985). Additionally, an external audit was conducted by a professor of education to review the research procedure and product periodically (Strauss & Corbin, 1990). Other validity approaches adopted for the present study included prolonged engagement (a research technique conducted over a longer time frame), persistent observations (a research method conducted in a series or repeatedly), triangulation, and referential adequacy using electronic technology such as camera and audio recordings (James & Bixler, 2008). Furthermore, the reliability of the coding was established by keeping a code book and using multiple coders to analyze transcribed data (Creswell, 2007).

In Phase 2, the content validity of the researcher-designed instrument was established through a review by two professors of architecture education from a university (Martin, 2003). Furthermore, the contrasted-group validity for the scale was determined by requesting the teachers to identify students with high and low degrees of environmental consciousness (Cronin-Jones, 2000; Ozdemir & Yilmaz, 2008). Significant differences were found by comparing the scores of these two groups of students. Additionally, a pilot study was carried out with a convenience sample to assess the reliability of the instrument (Duerden & Witt, 2010). We administered the test to the pilot group twice, with no treatment and with an interval of 30 days, to determine the test–retest reliability. The test score reliability was established using the Cronbach’s coefficient alpha (Boeve-de-Pauw & Van-Petegem, 2011). Creswell (2008) stated that researchers need to design their experimental studies to reduce threats to internal validity and external validity. The quasi-experiment in the present study took place in a natural setting as a way to strongly control threats to external validity. However, the self-selection bias posed a potential threat to the internal validity. We adopted several strategies to address this problem. First, matching through cohort controls was implemented by using different classes of fifth-grade students. The authors used cohorts as the comparison group because they were considered to be less nonequivalent than most other nonmatched groups would be. Second, cohort control was improved by adding a pretest. The use of the pretest scores as a covariate in the ANCOVA allowed error variance to be reduced and elimination of systematic bias (Culen & Volk, 2000). Finally, internal control was implemented by drawing samples from the same pool of the population, that is, from students in the same school.

We were previously involved with qualitative and quantitative research projects in the fields of architecture and sustainability. Thus, we had acquired the underlying foundations of each type of research and had basic skills in collecting and analyzing qualitative and quantitative data. Additionally, we had prior experience in designing and conducting a triangulation type of mixed methods study. This concurred with the recommendation by Creswell and Plano-Clark (2011) that a researcher should first develop his or her skills in both qualitative research and quantitative research separately prior to carrying out an MMR design. We also attended methodological workshops and sought guidance from three experts in qualitative, quantitative, and MMR design. For instance, before the commencement of the qualitative phase, we discussed with experts the central phenomenon of our case study and the techniques to pose qualitative and meaning-orientated research questions. We also learned the methods of collecting quantitative data and developed the basic skill in analyzing statistical data, including ANCOVA, t tests, Pearson correlation, and Cronbach’s coefficient alpha. Additionally, we attended workshops to study data collection, SPSS software analysis and common validation strategies in MMR.

In addition, syntheses of MMR studies were carried out. These methodological studies included journal articles, book chapters, or entire books that provided discussions on qualitative and quantitative research. They emphasized the synthesis of methods instead of a review of the results. We learned about the organization of the writing by reading examples of how other researchers had reported qualitative and quantitative data in their MMR. These examples also served as important references for how to design an MMR study, how to analyze and combine the data, and how to explain the details of the studies. We analyzed numerous examples of MMR studies that related to triangulation (e.g., Fielding, 2012; Hesse-Biber, 2012), embedded (e.g., Robinson & Mendelson, 2012), explanatory (e.g., Buck et al., 2009; Igo, Kiewra, & Bruning, 2008), and exploratory (e.g., Christ, 2007; Gilbert, 2010) designs. On reviewing all these studies, we identified the best design type that suited the research questions.

Time and resources were considered a crucial issue in the present study because the MMR design was complex and required extensive effort on our part (Tashakkori & Teddlie, 2003). As a result, a list of expenses and a timetable were developed to assess the feasibility of the research. The list of expenses provided the cost estimation for the entire project, including transportation, building materials, printing, and software programs. The timetable, on the other hand, offered a systemic timeline for data collection and analysis. Considering the amount of effort and expenses required to implement an MMR design, we applied for financial support at the start of the study. The availability of funding was an essential factor that determined the success of the present study because sufficient funds were necessary to support the construction of a full-scaled model in Phase 2. We contacted several government and private agencies to seek and apply for funding opportunities. The research project was subsequently awarded a grant by the Ministry of Higher Education in Malaysia, under the Exploratory Research Grant Scheme. We also used an unequal weighting in the research design so that it was more feasible to implement. The quantitative approach was given lesser priority. More time was allocated for qualitative data collection and analysis, in which the research was implemented using a two-phase approach with only one phase implemented at a time. This was more viable for those of us with limited resources and budget.

Conclusion

This article reports on an MMR design for an architectural study. The complexity of the problems and issues called for qualitative and quantitative data to be combined. A hybrid of the taxonomy development model and the embedded quasi-experimental model was identified as the most appropriate research design because it matches the timing, weighting, and mixing criteria. The qualitative and quantitative research techniques combined created access to data that we could understand, interpret, and implement for this study. Multiple sources of evidence were distinct to the MMR design objectified for the development and testing of an architectural intervention.

The present study began with an exploratory qualitative component followed by a confirmatory quantitative study to generate and test an architectural intervention. Within this design, the practical and logical approach best suits the requirements for the research to be flexible and developed at different stages. Phase 1 focused on a qualitative case study that involved interviews and observations to establish the design features of the 3-D textbook. A key outcome of this phase was the description of the 3-D textbook in terms of its defining variables and the ability to develop the intervention for further testing and validation. Phase 2 involved the embedded qualitative data within a quasi-experimental design. The data collected in Phase 2 enabled the effectiveness of the 3-D textbook to be verified and the prevalence of Phase 1 findings to be determined.

This article highlights the use of MMR design to create highly robust research consistent with the purposes of the study. Adopting the qualitative and quantitative methods to investigate multifaceted architectural issues, namely, the design and testing of the 3-D textbook, can promote innovations within the design and planning professions. The current study provides an important insight into the use of MMR to enhance the level of architectural design quality. Qualitative and quantitative research methods are used to reduce errors in design decisions, thus creating quality places for human occupation. We noted that credible research enables architects to evaluate, predict, and track the consequences of their design implementation. This is in contrast to the traditional thinking that design is an act of intuitive creativity established in the context of uncertainty.

MMR also promotes the conceptualization, evaluation, and testing of design concepts through an organized research procedure. It enables architects to gather data of different types from different sources and then analyze them. Therefore, architecture becomes an interactive process. MMR creates access to data that the architect can understand, interpret, and act on to inform his design. Architects can develop a design, experiment with it, learn from the results, revisit their design, and experiment with the design until the best solution is obtained. This is particularly useful when the architects intend to achieve a particular objective in their buildings. Examples include a hospital that promotes physical healing, an office that enhances productivity, or a school that encourages recycling. MMR is appropriate for this type of project because these buildings require the exploration and development of design solutions to improve the built environment and occupant performance. Additionally, the qualitative and quantitative findings provide a greater opportunity for design creativity by uncovering more evidence of human–environment interaction. Zeisel (1981) noted that a designer can be more creative if he or she collects evidence as a framework for added information to promote innovation.

Although MMR can be useful to architects, its application has been rather limited among the professions. Many architects might not have the expertise to conduct MMR. Others might face difficulties in analyzing the qualitative and quantitative data collected, especially in terms of whether data are strong or weak and how it could affect the design decision. As the concept of evidence-based design develops, architectural education ought to better prepare the architects to use MMR as a design generator. MMR can coexist with design creativity to expand the spatial, geometric, and aesthetic stating point of architecture to embrace building and human performance (Brandt, Chong, & Martin, 2010).