How do we know that sulfuric acid does not cure scurvy? Is it our wonderful intuitive sense about sulfuric acid’s properties? That can’t be it. Vitriol (sulfuric acid) used to be the Royal College of Physicians’ favored choice. The British Admiralty preferred vinegar. Still others favored a variety of remedies including the seemingly ridiculous notion of fresh fruit. This mess wasn’t sorted out until a young Scottish naval surgeon named James Lind did something revolutionary. In 1747 Lind conducted world’s first controlled clinical trial. Fresh fruit won, sulfuric acid and vinegar lost.
Similarly, there are various untested claims about which architectural interventions are beneficial for individuals with autism (see: here, here, here, and here). For the most part, these claims are mired in anecdote and conjecture. This makes it impossible to decipher which ones are sulfuric acid, vinegar or fresh fruit. Fortunately there are a few architects that have started to embrace the Lind spirit. This is the most important and necessary step architects need to take. If architects do not try to verify their claims through fair tests then they run the risk of undermining the public’s trust or worse, unintentionally doing harm to a vulnerable population.
When autism diagnoses took off in the 1990s, cottage industries promising cures for the disorder sprung up from every corner. They arose with the best of intentions, but many failed and still fail to realize that they are taking advantage of parents and caregivers who are willing to try anything to help their child. Parents will spend small fortunes for unproven interventions claiming benefits without risks. Only later do the financially strapped and disillusioned parents realize how scant the evidence is for the promised cure.  This is the context architects step into when they propose a building for individuals with autism.
To avoid becoming another source of distrust, architects must test their claims. If a design truly affects behavior, sensory regulation, and/or social integration architects can measure the effects. Is this easy? No, especially not with the diversity along the autism spectrum. Are there difficulties in controlling for confounding factors? Yes, but the level of difficulty is not an argument against studying the effects of a design decision. If architects claim there is a discernable difference between their design and any other design we should be able to measure it. If not then they cannot be able to claim it is true. If architects claim there is an effect, but it cannot be measured then all is valid based on conjecture. That is an idea the profession cannot accept if architects wish to be taken more seriously than snake oil salesmen.
Architects need to demonstrate that their design decisions are not only statistically significant, but also statistically meaningful. For example, a particular acoustical intervention might be statistically significant, but if it comes with a price tag that negatively influences other budget items, like teacher salaries, it could drive down the retention rates of a field that already suffers from high attrition rates. This could essentially wipe out the positive gains from the acoustical intervention. It is like a pill that cures a slight headache, but increases the risk for heart failure; the costs aren’t worth the benefits. If architects cannot demonstrate the statistical meaningfulness of their design, and the design costs substantially more than a non-autistic school/home of a similar size, then we should be highly skeptical and cautious in accepting the proposal. Furthermore, how we measure a design’s effectiveness matters.
Until recently, the only data about an autism design’s success came from post-occupancy interviews with teachers, staff, and parents. These types of evaluations are highly unreliable because of their susceptibility to selection and cognitive biases. (I have had to begrudgingly admit that much of the information I gathered by this method while visiting autism schools in the U.S., U.K., and Denmark is terribly suspect and next to useless when it comes to deciding what, if any, improvements resulted from the various design decisions, a costly methodological oversight.) A few architects, like Professor Magda Mostafa from American University in Cairo, are trying to change this hapless state of affairs.
In 2008 Professor Mostafa published a small study that used a control group to measure the effects of changing a classroom’s spatial characteristics and a speech room’s acoustics. Mostafa’s study is far from perfect, but it represents a paradigm shift in how architects have been studying autism design. The study has methodological weaknesses, such as a small sample size, treating the children as individual data points when they weren’t, and struggling to control for confounding factors and cognitive biases. These weaknesses make the results far less reliable and convincing than the study’s final conclusions might have you think (The conclusiveness of the study’s results is an issue Mostafa and I still respectfully debate). That aside, disregarding this study because the results might be inconclusive would be to miss the importance of it. The study is among the first autism design studies to be prospective not retrospective, have a control group, and measure quantifiable factors in a systematic way. This approach is leagues above what most other architects have been doing. Additionally, Mostafa is currently improving on the study’s weaknesses by putting together a larger more robust study.
In the end, the impact of the architectural environment on individuals with autism might be nominal for many or most and substantial for only a few or none. There is simply too little evidence to make definitive claims either way. However, since other environmental factors lead to vastly different outcomes not knowing the effects of designed environments is not an option. Environmental factors such as cultural acceptance, early intervention and educational programs have already been shown to substantially improve the quality of life for those on the autism spectrum.
Architects are still a long way off from discovering which design decisions are fresh fruit and which ones are sulfuric acid, but it is heartening to see architects moving in the right direction.
 Evans, Imogen, Hazel Thornton, Iain Chalmers, Paul Glasziou and Ben Goldacre. Testing Treatments: Better Research for Better Healthcare. Pinter & Martin; 2nd edition 2011. Kindle Location: 346-49.
 The idea of controlled trials predates James Lind’s study. For example, in Daniel 1:8 of the Bible, Daniel ordered a 10-day trial to determine whether the royal diet was deleterious to his health. Whether or not this actually occurred is unknown to me.
Another quasi example comes from King Frederick II of Sicily in the thirteenth century. Frederick wanted to know whether food was digested faster if you rested after eating or if you exercised. He took two prisoners, fed them, and sent one to bed and one to hunt. Later he disemboweled them and examined their innards. From Frederick’s point of view the sleeper won; from subjects point of view they both lost. (This might not be considered a controlled trial in the Lind sense, as it didn’t test the effectiveness of a treatment, unless you wanted to twist it into some kind of indigestion treatment. Importantly, the thinking was similar to Lind, just in a more twisted Medieval kind of way. See Sapolsky, Robert. Why Zebra’s Don’t Get Ulcers. Third Edition: The Acclaimed Guide to Stress, Stress-Related Diseases, and Coping - Now Revised and Updated. Holt Paperbacks, 2004. Kindle Location 2309-20.
 It should be noted that as early as the fifth century the Chinese “were growing fresh ginger in pots on board their ships.” I am unaware of how wide spared this practice was and if it was one among many as seen in Europe. The importance of Lind’s study is that it was able to systematically rule out incorrect theories based on unreliable evidence. See Le Couteur, Penny and Jay Burreson. Napoleon’s Buttons: How 17 Molecules Changed History. Jeremy P Tarcher 2004, Kindle Location: 520-33.
In Lind’s experiment he took twelve sailors with similar stages of scurvy, divided them into six pairs, and gave the first pair a daily dose of cider (1 quart), the second pair vitriol (25 drops daily), the third pair vinegar (2 spoonfuls 3 times daily), the fourth pair seawater (half a pint daily), the fifth pair the medicinal paste of garlic, mustard radish root and gum myrrh, and the sixth pair two oranges and a lemon. There were other suggested remedies including bloodletting, mercury paste, hydrochloric acid, bury the sick individual up to his neck in sand, and hard labor.
Lind’s experiment did not immediately change peoples mind for various reasons. This had terrible consequences as needless death continued to plague sailors. Over a decade after Lind’s discovery the British lost 1,512 sailors in action during the Seven Years War, 100,000 had succumbed to scurvy.
For nice short summaries of Lind’s experiment see:
Singh, Simon and Edzard Ernst M.D. Trick or Treatment: The Undeniable Facts about Alternative Medicine, W. W. Norton & Company; 1 edition October 19, 2009.
Le Couteur, Penny and Jay Burreson. Napoleon’s Buttons: How 17 Molecules Changed History. Jeremy P Tarcher 2004.
Evans, Imogen, Hazel Thornton, Iain Chalmers, Paul Glasziou and Ben Goldacre. Testing Treatments: Better Research for Better Healthcare. Pinter & Martin; 2nd edition 2011.
 Grinker, Roy Richard. Unstrange Minds: Remapping the World of Autism. Cambridge, Basic Books 2007.
Offit, Paul A. Autism’s False Prophets: Bad science, risky medicine, and the search for a cure. Columbia University Press New York 2008.
 Billingsley, Bonnie S. “Special Education Teacher Retention and Attrition: A Critical Analysis of the Research Literature,” The Journal of Special Education Vol. 38/No. 1, 2004, p. 39–55.
 Frith, Uta, Autism: Explaining the Enigma 2nd edition, Oxford, (Blackwell) 2003.
Grinker, Roy Richard. Unstrange Minds: Remapping the World of Autism. Cambridge, Basic Books 2007.
Tammet, Daniel. Born on a Blue Day. New York: Free Press, 2007.