Expanding the Pipeline
By Richard Ladner
It is startling to learn that approximately 16% of the US population of working age have disabilities. Some of these individuals are so cognitively or emotionally disabled that they cannot work, but most are capable of working and contributing to society. Within information technology (IT) fields the numbers compiled by the National Science Foundation (NSF) from various sources are interesting:
- 11% of the population in school, ages 14-21, have disabilities.
- 13% of undergraduate IT majors have disabilities.
- 5% of graduate IT majors have disabilities.
- 0.8% of IT doctorates have disabilities (e.g., there were 53 in 1999-2004).
- 5% of employed IT scientists and engineers have disabilities.
These percentages are fairly similar for all of science, technology, engineering, and mathematics fields combined. Please note that the 5% of employed IT people is not a subset of the 11% of school age because people can become disabled all through life. There are two things to note about these numbers. First, the interest in IT majors among students with disabilities is high, but their persistence to graduate education is low. Second, the number of self-reported IT doctorates with disabilities is very low. It is not known why these numbers are so low, but I suspect that part of reason is pervasive low expectations throughout elementary and high school that have led to lack of preparation.
One blind graduate student I know wanted to take calculus in high school to prepare to become a scientist. A group that included his guidance counselor and math teacher met with him to explain that no blind person in the school had ever taken calculus before, and that they would not support him if he decided to take it. Through the National Federation of the Blind he finally met blind scientists and mathematicians for the first time and attended a summer “boot camp” to prepare him for college. He felt he was very fortunate to find people who really understood his potential and raised his expectations. He will receive his Ph.D. in chemistry in the near future. Unfortunately, in our society most of us feel that if someone has a disability then we should give them a break, rather than giving them an opportunity to excel.
The argument for including people with disabilities in IT fields is not just an argument that we need more IT professionals or to promote social justice. There is a strong case that including people with disabilities improves the quality of our engineering outcomes as described by William A. Wulf 1.
“I believe that engineering is a highly creative profession. Research tells us that creativity does not spring from nothing; it is grounded in our life experiences, and hence limited by those experiences. Lacking diversity on an engineering team, we limit the set of solutions that will be considered and we may not find the best, the elegant solution.”
Recognition of Persons with Disabilities
On the day after the inauguration of Barack Obama, the White House posted a new web page stating the new administration’s commitment to persons with disabilities and urging the Senate to ratify the United Nations Convention on the Rights of Persons with Disabilities, which has already been signed by 137 countries. There are a number of laws and policies at all levels of government and policies within organizations that recognize people with disabilities and require actions to include and accommodate them. Best known are the Americans with Disabilities Act (ADA) and the US Rehabilitation Act, Section 508.
The NSF has a number of policies and programs that focus on broadening participation, including increasing the participation of persons with disabilities. The NSF has a policy in the Grant Proposal Guide: “Conferences or meetings, including the facilities in which they are held, funded in whole or in part with NSF funds, must be accessible to participants with disabilities.” Accessibility can include accommodations such as real-time captioning or sign language interpreters for deaf participants, as well as wheel chair ramps and accessible rooms in the conference hotel. As a member of the NSF Committee on Equal Opportunities in Science and Engineering (CEOSE) I have seen that NSF considers persons with disabilities in a way similar to other underrepresented groups such as women and minorities. Currently, the committee’s 13 members include African American, Native American, and Hispanic members, and two with disabilities. In 2007, the committee focused on people with disabilities and made a number of recommendations to NSF to help increase the participation of people with disabilities in science and engineering. The CISE Broadening Participation in Computing (BPC) Program funds a number of projects, including my own AccessComputing Alliance, that focus on persons with disabilities.
Generally, the computing community represented by ACM, IEEE Computer Society, and CRA have done a good job in recognizing the need to include more women and minorities in the computing fields and taking action to increase their participation, but there has been no similar recognition and actions for persons with disabilities. There are some bright spots. The ACM has a relatively new policy on the accessibility of the Internet, including its own web pages (http://www.acm.org/public-policy/accessibility). ACM SIGACCESS has as part of its mandate not only the promotion of accessibility research, but “also strives to educate the public to support careers for people with disabilities.” Some ACM conferences, such as those sponsored by SIGARCH, offer travel support for an assistant who may have to travel with an attendee with a severe disability.
I believe that a major step in recognizing the underrepresentation of persons with disabilities in computing is to expand the annual CRA Taulbee Survey to include demographics about people with disabilities as it already does for women and minorities. Only in this way can we learn where we are and where we need to go. The data reported at the beginning of this article are from a variety of sources, some publicly available and others available to universities and societies by license agreement. The Taulbee Survey polls CRA’s member departments annually for the number of students and faculty who are women or minorities, but does not ask about disabilities. All major universities provide services for students with disabilities and typically keep track of which students receive such services in their student database. Summary data can be obtained from student databases by institutional request.
One of the exciting things about computer science research is that it can actually make a difference by creating new technologies that enable persons with disabilities to participate more fully in our field, every other field, and life generally. This research is called accessibility research because the results can enable access to activities in life that otherwise would be difficult to impossible. Advances in robotics, computer vision, natural language processing, and other fields can have a direct impact on persons with disabilities. Examples include smart vehicles that can climb and descend stairs safely used in some modern wheel chairs; speech synthesis and optical character recognition used by blind people to obtain access to both electronic and printed text; and speech-to-text used to produce real-time captions for deaf people. Moreover, research in accessibility topics can lead to solutions to problems that benefit people who are not disabled. Optical character recognition was pioneered by Kurzweil so that blind people could read printed books. Now it is being used to convert all books in print into electronic form. Captioned television was originally developed for the benefit of deaf people, but now it is used by most of us as we watch TV in noisy airport waiting areas.
The excitement about working on problems related to disability is growing rapidly as evidenced by the number of papers in the major human-computer interaction conference, CHI, that have contain the word “disability.”
|CHI Years||Number with “Disability”|
|2006-08 (3 years)||71 (17%)|
Before 1986 there were no papers in CHI that used the word “disability.” The rapid growth since 2001 seems to indicate that the CHI community recognizes that accessibility problems are important and interesting. The activity in accessibility research has grown to the point that the highest quality work will appear in the new ACM Transactions on Accessible Computing (TACCESS) that complements the annual ACM Conference on Computers and Accessibility (ASSETS), now in its 11th year. There are several older conferences that focus on “assistive” technologies that attract engineers and inventors. CHI, TACCESS, and ASSETS papers typically contain empirical studies, with human subjects or otherwise, that help verify the efficacy of some new technology.
In this short article there is no way I can cover the breadth and depth of the research done in accessibility. Instead, I would like to tell you about an upcoming paper in CHI ’09 by two young researchers, Jeff Bigham and Anna Cavender, from the University of Washington.2 What is interesting about their new technology is its simplicity and immediate applicability. Most of us are familiar with CAPTCHAs which provide a visual test that can be passed only by humans before entering some web site. For example, to create a Google mail account you have to prove you are a human first by solving a visual CAPTCHA. People who are blind, have low vision, or have some kinds of color blindness may not be able to solve a visual CAPTCHA, so audio CAPTCHAs have been invented and are in common use. Through a remote study of 162 people, 89 of whom were blind, they discovered that sighted people can solve 80% of visual CAPTCHAs on the first try, while blind people could only solve 43% of audio CAPTCHAs on the first try. Audio CAPTCHAs took five times longer to solve than visual CAPTCHAs.
There are a multitude of reasons for the inaccurate and time-consuming solving of audio CAPTCHAs, but an important reason is the design of the user interface for audio CAPTCHAs. There can be interference between the screen reader used by the blind user and the playing of the audio CAPTCHA. Because the audio CAPTCHA contains ten letters that are hard to remember in one pass, the audio CAPTCHA has to be replayed multiple times. Each replay requires navigating to the play button and then navigating back to a text-entry box, which takes several steps in a screen reader. Although current audio CAPTCHA interfaces allow accessibility, they are not very usable by blind people. Bigham and Cavender’s ingenious solution is to get rid of the play button altogether. Instead, the user remains in the text entry box at the same time as listening to the CAPTCHA using non-letter characters like the comma, period, and forward slash keys to control the player to rewind, pause/play, and fast-forward. Letters that are typed go toward solving the CAPTCHA. With the new interface, blind users solved 68.5% of the audio CAPTCHAs on the first try.
Although their final solution is simple and elegant, it emerged by creative insight from a lot of hard work building the infrastructure to reach and record the actions of many users in the field. Moreover, their solution is immediately applicable. There is no compelling reason why audio CAPTCHA interfaces cannot be changed to be more usable in a very short time on all the websites that use them.
This example of accessibility research gives a quick taste of a much wider field that covers many kinds of disabilities with a variety of approaches and solutions. There is essentially an infinite variety of loss of functional abilities within the broad categories of hearing, vision, motor, cognitive, and age-related disabilities. Web accessibility is a critical area because just about anyone can be a web author, so there is no practical way to enforce web accessibility standards on everyone. As an example, only a tiny percent of video on the web is captioned. Popular devices like the iPhone, which only have a touch screen, present a research challenge to make them accessible to blind people and people with tremors or other motor disabilities. Very little new technology is built to be accessible in the first place, so there will always be new and interesting problems to solve. Trying to find ways to economically and esthetically design new technology that is accessible to the vast majority of people from the outset is the hope of universal design.
So far I have stressed accessibility research problems that have a short time horizon. Their solutions can have an immediate impact on the lives of people. Other researchers are taking a longer view, approaching science and technology problems that may eventually lead to remarkable results. For example, researchers in neuro-computing and neuro-robotics are trying to find ways for the brain to directly control a screen cursor and the brain and nervous system to control a robotic hand. Whether you enjoy doing research as pure science with the potential of application, or more applied science and engineering with more immediate impact, there are plenty of exciting things to do in accessibility research.
I hope I have helped convince the readers of this article that including persons with disabilities is important to computing fields, that more can be done to facilitate that inclusion, and that research in accessibility is an exciting and rewarding activity. For more information about AccessComputing, please visit the web site.
Richard Ladner is Boeing Professor in Computer Science and Engineering at the University of Washington, where he has been on the faculty since 1971. In addition, he is an Adjunct Professor in the Department of Electrical Engineering and in the Department of Linguistics. He was the winner of the 2008 A. Nico Habermann Award.
Acknowledgment: Thanks to Joan Burrelli, Division of Science Resources Statistics, NSF, for help with the statistics on persons with disabilities.
1 Wulf, W.A. (2000). “How Shall We Satisfy the Long-Term Educational Needs of Engineers?” Proceedings of the IEEE, 88(4), 593-596.
2 Jeffrey P. Bigham and Anna C. Cavender (2009). “Evaluating Existing Audio CAPTCHAs and an Interface Optimized for Non-Visual Use.” In Proceedings of the ACM Conference on Human Factors in Computing Systems (CHI 2009) Boston, Massachusetts, 2009. To Appear. Preview copy.