Early UD proponents: defining necessary knowledge

Universal Design emerges alongside evidence-based practice in the late 1960's and early 1970's, when a liberalizing social and political climate began affecting the design professions and their emerging focus on users. The dedication of many early Universal Design proponents to the goals of the disability rights movement is well-documented (Ostroff, Limont, & Hunter 2002; Steinfeld & Maisel 2012, 15-18). The architect, Ron Mace, coined the term "Universal Design", in the early 1970's to describe budding socially inclusive design practices that privileged users in the design process rather than making them secondary to aesthetic or formal considerations (Kawuchi 2009, 7; Mace 1985; Steinfeld 2011). The term first appeared in print in an article by Mace in 1985 (Mace 1985).

While UD's emergence correlates with the civil rights movement era, and is an example of "value-explicit" design and research (Moore et al. 1985, 21), causality for its appearance on the scene of architecture and industrial design cannot be drawn from civil rights ideologies alone. Instead, it is useful to consider the entangled network of actors, epistemologies, and professional commitments that materialized UD. UD proponents did not only demand the protection of people with disabilities by a civil rights regime, but also worked on strategies for challenging the design profession's attitudes, norms, and values regarding the normate template (Dolmage 2005; Williamson 2011). Likewise, while the disability rights movement may have won legislative battles culminating in federal disability rights laws, it did not directly propose strategies for challenging the professional practices through which the design professions had adopted the normate as an ideal type. 5 Efforts to educate designers and challenge professional practices by UD advocates can be understood as a parallel movement to the direct action and legislative demands of the disability rights movement.

In the mid-1990's, a group of practitioners working on UD in the fields of architecture, industrial design, assistive technology, and environmental design research convened at North Carolina State University's Center for Universal Design to draft the "Seven Principles of Universal Design" (Center for Universal Design, 1997; Story 1988). The Principles' authors were researchers and practitioners in evidence-based fields such as ergonomics and human factors research, industrial design, gerontology, and rehabilitative medicine. 6 They drafted principles guiding the UD method toward designs that are flexible, equitable, simple, easily usable, and tolerant of error. While often cited as the basis of the UD design method and ideology, the Principles also reflected their authors' commitment to applied research as a design tool. For example, flexibility obliges designs to account for a range of user body sizes and functions. To understand the range of possible variations, designers can rely upon anthropometric data about bodily movement, reach, or size. This data influences the width of doorways, the height of cabinets, and the shape of door handles. Although not explicitly evidence-focused, the Principles made legible the need for bodies as evidence in UD.

UD as evidence-based design

Feminist science and technology studies scholar, Michelle Murphy (2001), has usefully borrowed from Michel Foucault the concept of the regime of perceptibility or intelligibility in order to elaborate upon the overlapping histories of design, science, and spatial users in the emergence of the phenomenon of sick building syndrome. Regimes of perceptibility are "the regular and sedimented contours of perception and imperception produced within a disciplinary or epistemological tradition" (Murphy 2001, 24). According to Foucault, new regimes of perceptibility do not replace but rather "supplant" or "superimpose" themselves upon previous regimes, creating space for the intelligibility of new objects and practices that were previously imperceptible (Foucault 1980, 106).

What kind of regime of intelligibility is evidence-based design and what is its epistemological significance for UD? The significance of knowledge production activity in UD can be traced back to a discourse developing between designers and researchers interested in evidence-based and user-centered practices in the late 1960's (Steinfeld & Maisel 2012, 34-35). Evidence-based design, like its counterpart, evidence-based medicine, relies upon the production of knowledge to address and solve practical problems (Moore and Geboy 2010, 106). Changes occurring in the civil rights era in architectural standards of evidence, as well as in the field of environmental design research, anticipated UD's emergence as a knowledge-producing counterpoint to mainstream design. This section charts two new directions within design epistemology and methodology: one advocated in order to encourage designers to use evidence and data rather than intuition, and another advocated to change the focus of the research that produces evidence for designers to include a user focus.

Early UD designers and researchers theorized that the lack of usable data about human body variation was in part responsible for the inaccessibility of existing designs (Mace et al. 1966; Preiser 1973; Preiser & Vischer, 1991; Salmen 2001, 181-182; Sandu 2001). The changing demographics of disability created uncertainty about the range of sizes and functions of bodies, serving as a barrier to meaningful accessibility standards (Story, Mueller, & Mace 1998, 6-7; Steinfeld & Tauke 2002, 166-167). Making evidence the basis of practice would, in turn, address uncertainties associated with the diversity of potential users by expanding the scope of what designers could know about disability.

The change in focus was not only about including both normates and mis-fits in research, but also about person-environment relations. Before evidence-based design, knowledge used by designers from outside the profession tended to be about the environment and not its interaction with users (Moore & Geboy 2010). Whereas mathematics and engineering knowledge previously informed architects about the materiality of buildings, and while biomedical knowledge independently located disability and medical diagnosis within bodies rather than their environments, evidence-based design focused on the moments of chance interaction between human bodies and built, social, cultural, and natural environments.

The availability of evidence-based practice was necessary to the eventual emergence of UD as a design method. The UD Principles' insistence on flexibility, equity, and user needs became perceptible as design mandates of evidence-based and user-centered design in non-disability contexts. The U.S. military, for example, pioneered human factors and ergonomics research in order to flexibly design for the range of its soldiers in the early 20th century (Gordon & Friedl 1994). While focused on non-disabled military men, this research made users' bodies legible as evidence within design, rendering issues like aging and disability relevant to the built environment. Eventually, the emphasis on health, aging, and disability came to characterize much of evidence-based design because its research base drew upon available human factors research data and methods (Moore & Geboy 2010; Zeisel 2006). The legibility of concerns with the body, in turn, enabled UD advocates to argue that designers had an ethical responsibility for accessibility because architecture and industrial design impact everyday life (Preiser & Vischer 1991).

The shift to evidence-based practice was a shift in an architectural regime of perceptibility in which epistemological standards governed what designers perceived as the scope and inclusionary imperatives of practice. This does not mean that evidence-based practice completely replaced intuitive approaches to design, or even that it became successful in the displacement of this prior regime of perceptibility. Instead, it was an epistemic and methodological overlay producing the intelligibility of users as evidence to designers. 7 Tanya Titchkosky writes that access "is an interpretive relation between bodies" (Titchkosky 2011, 3). In UD, human bodies and behaviors became legible proofs of the normate template's inadequacies for design, creating an "interpretive relation" between bodies and environments. Human factors research testified to human-environment interactions as bodies and behaviors became useful and valid evidence for research-driven design.

The truth-telling qualities of bodies and behaviors, however, were not always immediately accessible. Rather, the methods and standards by which mis-fitting bodies became legible as privileged evidence emerged through specific, historical research practices. The next section describes one research practice of import to contemporary UD: anthropometry. UD anthropometry research exemplifies the call for multidisciplinary research in evidence-based practice. Understanding its history illuminates significant and otherwise imperceptible epistemic shifts within design and scientific research on disability.

Anthropometry: rendering and challenging the normate template

Anthropometry encompasses methods of measuring samples of human bodies for the collection of data about size, ranges of motion, and other quantifiable spatial measurements. Its historical trajectory includes the use of bodies as evidence in establishing standards of positive science and quantification. Following anthropometry's history demonstrates that UD has responded to the normate template by re-conceptualizing research design methodologies and standards of data interpretation rather than simply changing its sources of evidence or the intentions of its practice.

Sociologist of science, Steven Epstein (2009), has noted the paradox that many marginalized groups since the late 20th century have demanded inclusion in scientific research that less than a century prior was used to segregate or eliminate these same populations. The same paradox is present in UD's use of anthropometry, the previous practice of which has been at odds with the goals of broad inclusion and social justice (Davis 2006, 4). Co-produced with 19th century positivism and eugenics, anthropometry provided evidence of the supposed excess or deficiency of bodies that deviated from the statistical average (such as disabled bodies and differentiated racial populations). It was entrenched in 19th century practices of measuring bodies to quantify normate and mis-fit.

Historian of science, Stephen J. Gould, attributes the biological determinism of 19th century sciences to their reduction of human bodies and anatomical features to measurements (Gould 1981, 25). Francis Galton, a key figure in the development of modern statistics, used anthropometry to measure intelligence through race and inheritance, laying the groundwork for the use of biometrics for racial differentiation and the establishment of the statistical law of normal distribution (Gould 1981, 76; Graves 2001, 92-93; Hacking 1990, 105-106; Kevles 1985, 18; Sekula 1986, 20-21). The logic of measuring and ordering bodies infused eugenics (Laughlin 1919, 55-61). This logic also enabled physical anthropologists and state officials, some working in the service of colonialism, to establish hierarchies of racial dominance through scientific measurement (Adebisi 2008; Hrdlicka 1919, 22-23; Phillip 2004, 99; Blanckaert 1988, 49).

The practice of anthropometry is co-produced with 19th century positivist epistemology. In developing statistical laws, Galton conflated norms and averages by relying upon Adolphe Quetelet's finding (via anthropometry) that larger proportions of the population are represented by the mean, making it possible to quantify the "average man" (Canguilhem 1989, 157; Davis 2006, 6). Nineteenth century anthropometry typifies what Georges Canguilhem describes as the false conflation of norms and averages. The norm is a state of equilibrium determined on an individual and qualitative basis, making it distinct from the statistical average, which is based upon a quantified data set derived from a population of individuals (Canguilhem 1989, 76, 151). When the statistical average is conflated with normalcy, deviations from the mean are characterized as both quantitatively and qualitatively deviant.

With the "average man" occupying the place of normalcy in Galton's eugenic ideology, he developed the idea of statistical correlation (Stigler 1989; Cowan 1972, 509). Correlation allows the conflation of the average with the norm without consideration of the causal events through which bodies come to appear as normal or deviant. Standardization and correlation are foundational to eugenic anthropometry's claims to objectivity because they justify the validity of claims about the correlation of skull size to intelligence or body proportions to evolutionary desirability (Gould 1981, 256).

While eugenic anthropometry focused on measuring bodies as evidence of racial and ethnic superiority, the U.S military adopted anthropometric methods to make functional changes to weapons of war. Early 20th century military anthropometry in the U.S. tested the characteristics of soldiers with high performance and presumed high intelligence (Meister 1999, 148). After the war, anthropometric data obtained from approximately 100,000 soldiers were published and distributed for general knowledge (Meister 1999, 152-3). The military continued to use anthropometrics for the ergonomic design of cockpits and weapons (Wickens and Hollands 2000).

Eventually anthropometric data and ergonomics entered industrial design and civilian product design as evidence about potential users. The normate template for mainstream architecture was institutionalized by the mid-20th century industrial designer, Henry Dreyfuss, who used anthropometry data taken from young, able-bodied, military men for to develop anthropometric figures of supposedly average bodies for designers. In their early iterations, these figures displayed no data about disabled bodies, women, children, elderly people, or any other body not compliant with the normate template (Dreyfuss 1960; Hosey 2001; Serlin 2002, 65-8).

Dreyfuss distributed anthropometric data through images appearing in texts intended for designers. The images made the military data intelligible to designers who were not trained to interpret statistical data and tables. Other figures representing ideal geometric embodiment, such as the Vitruvian Man or Le Corbusier's Modulor, had guided architectural conceptions of the body long before Dreyfuss's anthropometric images (Hosey 2001; Wittkower 1949). By using images to convey knowledge to designers, Dreyfuss continued the tradition of depicting bodies graphically. As in this tradition, he represented a seemingly neutral body that in actually represented an impossible ideal type—the normate (Hosey 2001). Anthropometry entered architectural practice when the handbook of American architecture, the Architectural Graphic Standards, adopted images from Dreyfuss and others (Ramsey & Sleeper 1947). As a result, anthropometric images often serve as the primary available information about human bodies for architects in the mainstream, but primarily display data about average bodies (Hosey 2001).

The new disability anthropometry

Disability anthropometry as evidence for design is a recent epistemic and methodological challenge to the normate template and to anthropometry as a scientific practice. By showing that the normate, constituted through exclusionary anthropometric data, is a socially and historically specific figure that is not generalizable to the whole population, disability anthropometry undermines its seeming neutrality and objectivity. Central to the UD research strategy is establishing a regime of perceptibility whereby knowledge production makes excluded bodies apparent to designers while also shifting the meaning and methods of research.

Just prior to the shift toward human factors in environmental design research, the British architect, Selwyn Goldsmith, published Designing for the Disabled (1963) as a model for incorporating anthropometric data about people with disabilities into design as a corrective to existing data sets. Over a decade later, evidence-based design advocate and UD Principles author, Edward Steinfeld, received a grant from the U.S. Department of Housing and Urban Development to conduct a study on 60 wheelchair users that became the basis of the federal government's technical accessibility standards (Steinfeld, Schroeder, & Bishop, 1979). The resulting study, published in 1979, constituted a limited challenge to the normate template but became the basis of the American National Standards Institute access codes (Seelman 2005). This mobility anthropometry data was not updated until the 2000's, when Steinfeld and a new team of researchers at SUNY-Buffalo's Center for Inclusive Design and Environmental Access (IDEA), which houses the RERC on Universal Design and the Built Environment, undertook a decade-long study with a much larger sample size, different targets and objectives of study, and improved research methods (Steinfeld et al. 2010). The final version, published in 2010, included 500 users of wheeled mobility devices, including manual and power wheelchairs and scooters (Steinfeld et al. 2010). It is to date the largest study of the anthropometry of wheeled mobility devices.

While the study's report does not mention anthropometry's problematic histories, it makes several epistemological interventions in the standards of knowledge around anthropometry and statistics, and also develops methodologies that reinterpret standards of scientific rigor. First, the study historicizes existing disability anthropometry within 20th century developments in assistive technology. It notes that the previous sample of only 60 wheelchair users was not only inadequate due to its small sample size, but also was inapplicable to technological advances in mobility technologies, such as powerchairs and scooters, since the 1970's (Steinfeld et al. 2002, 17; Steinfeld et al. 2010, 84). Recognizing technological change demonstrates a shifting understanding of possibilities for body-environment relations.

Second, the study challenges statistical methods by rejecting the notion of proportionate sampling that is central to the generalizability of quantitative research. Proportionate sampling conventionally means that if a general population includes 20% individuals of a certain category, the study sample must also include the same percentage of those individuals (Steinfeld et al. 2010, 17-18). Generalizing from a small sample to the whole population in this way is a mechanism by which privileged figures, like the normate, become a legitimate representation of the supposedly average body.

Steinfeld et al.'s study did not endeavor to achieve this level of generalization because it would not be useful for designing for broad inclusion. As the authors note,

most people with severe disabilities fall outside [the representation of the whole population] when proportionate sampling schemes are used. This means that disability will be very underrepresented in conventional anthropometric studies. Some studies purposely exclude people with disabilities and older people to keep the results unaffected by 'outlying cases' or people who have widely divergent abilities and characteristics (Steinfeld et al. 2010, 17-18).

Steinfeld et al.'s study sought data that has been controlled out of other studies via sampling. Within the 2010 sample, researchers oversampled for previously underrepresented populations of power wheelchair users. Because more data already exists on manual wheelchair users (a significant portion of the population of mobility aid users), researchers concluded that oversampling power chair users who are least represented in previous studies would not distort their results, but rather ensure some representation of this group (Steinfeld et al. 2010, 83).

Third, the study develops methods for making body-environment systems and interactions the object of research. To do this, the study measured people performing "real world tasks" (Steinfeld et al. 2002, 13). As Steinfeld et al. note in the proceedings to a 2002 conference on disability anthropometry,

…structural measurements alone cannot fully predict human performance in real world settings where the body is usually in motion or under stress…This has added another level of complexity to anthropometry because free, or unloaded, movement is not always sufficient to capture the nature of performance in real world tasks (Steinfeld et al. 2002, 13).

Previous anthropometry used static measurements (called "structural" anthropometry) of bodies at rest and divorced from environmental context; the new anthropometry uses both static and dynamic ("functional") anthropometry of bodies relating to and using environments and technologies (Steinfeld et al. 2010, 2; Steinfeld et al. 2002, 16-17).

Studying body-environment interactions in real world tasks also shifted the way that evidence about bodies becomes valued within a data set. For functional anthropometry, bodies are irreducible to their measurements, and measurements are only meaningful when they convey information about how bodies and environments act in tandem. Disabled bodies are not merely objects of medical knowledge or curiosity (Mitchell and Snyder 2006, 28), but evidence of social inclusion that becomes perceptible when put into interaction with the built environment. As evidence, inaccessibility and mis-fit become measurable and testify to the environment's fit to the body.

Fourth, whereas static anthropometry measures the aberrance of the body from a statistical average, functional anthropometry measures the environment's mis-fit with a range of bodies. 8 This range is depicted both graphically and in tables. Data about ranges makes it possible to design, for instance, desk chairs or tabletops that are adjustable to the height of multiple users. The range defies the notion of standardization, requiring flexible solutions to the built environment that accommodate multiple abilities and embodiments within a single design. Thus, the norm of representation within disability anthropometry is the range rather than the average (Steinfeld et al. 2002, 33-34).

Valuing ranges as evidence requires adopting methodological reflexivity about the reliability of standard measurements. Steinfeld et al. (2002) found that the diversity in bodies, mobility devices, and environments in their study made it difficult to standardize measurements. Conventional anthropometry methods rely upon able-bodiedness as a norm, requiring subjects to remain upright in difficult poses during research and de facto excluding those who do not fit the experimental design (Steinfeld et al. 2002, 24). Also, the inability to standardize body landmarks for measurement across diverse bodies creates "barriers in translating research findings to standards development" (Steinfeld 2010, 4). The researchers adopted reflexivity by questioning the objectivity of quantified measurements based on the inadequacies of current methods to yield usable data. This reflexivity requires new accessibility standards to be flexible and reflect ranges of interactions between bodies and environments.

Finally, in Steinfeld et al.'s study, multidisciplinary research required anthropometry to be conducted alongside other methods. The process of defining "measurement variables and procedures" entailed, in Steinfeld et al.'s study, consultation with "experts in anthropometry and ergonomics, human modelers, architects and designers, and clinicians" (Steinfeld et al. 2010, 2). This meant that epistemic standards and methodologies from all of these fields entered the design of research.

Multidisciplinarity shifts the terms of expertise and the value of sources providing valued knowledge for accessible design. Steinfeld et al. used surveys, focus groups, and ethnography alongside anthropometry to identify qualitative factors and preferences that are inaccessible to quantification alone, rendering legible the people whose bodies are being used as sources of evidence (Steinfeld et al. 2002, 21, 29). These qualitative studies rest upon the validity of the experiences and perspectives of spatial inhabitants as valuable knowledge and follow models for user-centered design created by people with disabilities specifically for architectural interventions (Lifchez 1987). For example, Steinfeld et al. conducted focus groups with potential anthropometry research subjects in order to define non-invasive research methods of contact with bodies (Steinfeld et al. 2010, 2). In this way, the bodies of people measured by anthropometry also participated in designing research methods and defining their own consensual limits. This participation departs dramatically from the invasive and (often) coerced anthropometry conducted in eugenic, military, or physical anthropology contexts.