Appendix

Appendices included in this section are:

• Appendix A: Methodology
• Appendix B: About Forrester Research, Inc.

Appendix A: Methodology

Forrester conducted a nationwide survey in May through July 2003 to assess the incidence rates and the degree of severity of difficulties and impairments among US working-age adults and computer users. The survey was conducted by phone and mail, yielding a total of 15,477 respondents. Respondents were asked a range of questions about difficulties, impairments, computer use, and attitudes toward technology. Resulting data is representative of the adult US population.

Forrester created a survey, in conjunction with Microsoft's team, to be fielded to a representative selection of households within the US to assess incidence rates and the degree of severity for several difficulty/impairment types. The difficulty/impairment types are: visual, dexterity, hearing, speech, and cognition. To measure a range of potential demand for accessible technology, a three-pronged approach was developed asking questions about the respondents' level of functioning within a difficulty/ impairment type; assessing task-based limitations, self-assessed limitations, and difficulties/impairments effecting employment.

Survey questions: The survey was designed to identify individuals who self-identify as having a difficulty or impairment, as well as those who do not consider themselves to have an impairment but do report difficulty with tasks.

The survey contained the following three types of questions:

1. Task-based questions: Designed to understand difficulty with ordinary daily tasks in each of the difficulty/impairment types. The types of questions asked include: "Please indicate how often, if ever, you have difficulty seeing the words and letters in ordinary newspaper print because of your eyesight. If you usually wear glasses or contacts, please indicate whether you have difficulty while wearing glasses or contacts," and "Please indicate how often, if ever, you have difficulty using a keypad on a phone/dialing the phone because of physical difficulties with your arms, hands, wrists or fingers."
2. Direct questions about impairments: Designed to assess the proportion of the population who believe that they have an impairment. Individuals were also asked to assess the degree of severity of their impairment. The purpose of asking people for their own assessment was to understand the role of self-identification in seeking out solutions to difficulties with computer tasks. These questions were more direct, for example, "Do you have a visual impairment?"
3. Direct questions about impact on employment: Designed to allow individuals to communicate their assessment of the limitations imposed by their impairment. For example, people identified themselves to have a visual difficulty/impairment were asked directly: "Do you have a visual impairment that limits the kind or amount of work you can do?"

In addition to a rigorous assessment of respondents' ability within a difficulty/impairment type, the survey assessed computer usage; technology, health, and life attitudes; employment; accessible technology use; and a range of demographic characteristics.

Sample size: A sample size of 15,000 ensured that the study would capture a large sample of computer users with various impairments. Additionally, this ensured that some low-incidence difficulty/impairment types (like severe hearing impairment) would be sufficiently represented.

Phone and mail surveys: Respondents were contacted by either mail or phone from May to July 2003; 10,464 respondents are members of the mail panel managed by National Family Opinion (NFO); the remaining 5,013 respondents were contacted by phone through random digital dial.

Fielding the surveys by phone and mail allowed us to take advantage of the benefits of each method, accurately capturing a representative sample of the US population while minimizing bias against specific difficulties/impairments that would have challenged answering phone or mail surveys alone. The benefits of the mail survey include that it was more accessible for people with hearing difficulties and impairments and that the panelists know the source of the questionnaire, making them feel more secure revealing private information. Additionally, the mail survey was supplemented with rich background data on panelists (this background data also allows consistency checks on answers like income, age, and gender). The benefits of the phone survey include that the RDD method is more likely to include less traditional household structures, it is more accessible for people with visual difficulties/impairments, and the survey is guided by an interviewer, which reduces errors in the answers.

US focus. Impairments will likely affect people similarly in other countries and our findings will be noteworthy outside of the US. As such, Microsoft made the decision to focus the survey in the US because of the high cost of a multi-country survey and because the survey couldn't be translated simply. The survey questions would need to be reconsidered to field the survey outside of the US, rather than directly translated, because of the sensitive nature of the research topic. Forrester Research believes that the survey design, and to some extent the research methodology, would need to be customized for each country. A key benefit of a US-only study is that the lessons from this experience can help ensure better execution of non-US versions, if Microsoft decides to invest in research on this topic in other countries.

Weighting the data. Forrester Research uses weights to ensure that the final group of survey respondents gathered through each method is representative of the general US population. Mail surveys were sent to a representative group of households, but the final data was weighted due to differing response rates. Similarly, phone surveying, while providing a random selection of households, creates a self-selected bias in the sample that requires weighting to correct. Weights were created by identifying target characteristics of US households from the June 2003 Current Population Statistics. Sampling was done at the household level, so the weights are designed to ensure that the data is representative of US households and the individuals within them, rather than all individuals.

Although phone-based and mail-based data was weighted along similar grounds and attributes, weights were developed separately to ensure representation across both groups. Each survey was weighted so that, when combined, it represented 50% of respondents. While the mail survey actually captured twice as many respondents as the phone survey, mail survey respondents were more likely to identify limitations/disabilities in many cases. As such, the decision to weight them equally provided more conservative estimates of difficulty/impairment rates.

How this study compared with the US Census SIPP. The US Census provides data on disability based on three primary sources: the Survey of Income and Program Participation (SIPP), the decennial census, and the Current Population Statistics (CPS). The SIPP is the only one of these with an extensive set of disability questions and is therefore identified by the Census as the preferred source for examining most disability issues. The following table shows data collected in the SIPP compared with data collected in our research study. Additionally, a question was asked to be comparable to the CPS and Census ("Do you have a physical, mental, or other health condition which prevents you from working or limits the kind or amount of work you can do?").

Comparison of Question Types Asked in the US Census SIPP vs. This Study

Accuracy of results:  For results based on a randomly chosen sample of this size (N=15,477), there is 95% confidence that the results have a statistical precision of plus or minus 1% of what they would be if the entire adult population of US households had been polled. The phone survey is randomly sampled, but the mail sample is not a random sample; while individuals have been randomly sampled from NFO's panel for this survey, they have previously chosen to take part in the NFO mail panel.

Appendix B: About Forrester Research, Inc.

About Forrester

Forrester Research identifies and analyzes trends in technology and their impact on business. It provides companies with practical ideas, rigorous research, and objective guidance to help them thrive on technology change.

In February 2003, Giga Information Group became a wholly owned subsidiary of Forrester Research, Inc. Giga provides objective research, pragmatic advice, and personalized consulting to global IT professionals.

Together, Forrester and Giga enable companies to make better strategic decisions that maximize technology investments and achieve identifiable business results.

The Forrester Project Team

Forrester's Custom Consumer Research team helps companies make strategic business decisions by providing analysis of consumer attitudes and behavior and helping to formulate market strategies. To gain a thorough understanding of consumers, the CCR designs, implements, and analyzes proprietary consumer surveys. Furthermore, the CCR may work with data proprietary to clients to provide the best possible assessment of current customer demand issues. In addition to data analysis and consumer research, senior research staff at Forrester combines the consumer analysis with relevant industry analysis. Senior research staff works together to find the best solution for clients' business problems.

The project team included:

• Betsey Stevenson, Ph.D. Custom Consumer Research Advisor, who was the lead researcher. Betsey has experience leading market research for established and nascent technologies and leads Forrester's Custom Consumer Research team.
• James McQuivey, Group Director of Consumer Data® in North America, supervised the execution of the program.

Betsey Stevenson, Ph.D., Custom Consumer Research Advisor

Betsey leads Forrester's Custom Consumer Research team, helping clients make strategic business decisions by designing, implementing, and analyzing proprietary consumer surveys. She has served as an analyst in Consumer Technographics®, covering a wide range of industries.

Before coming to Forrester, Betsey taught and conducted research in economics at Harvard University. Her work there focused on statistical analysis of individuals to predict behavior in a wide range of consumer, health, and family areas. Betsey also taught courses in applied econometrics at Harvard University. Her writings have appeared in a variety of publications, including The Wall Street Journal. Prior to her work at Harvard, Betsey was part of the International Finance group of the Federal Reserve Board, analyzing the financial situation of developing countries and debt restructuring packages.

Betsey graduated from Wellesley College and received an M.A. and Ph.D. in economics at Harvard University with an emphasis on econometric analysis and quantitative research design.

James L. McQuivey, Group Director

As the leader of Consumer Technographics® in North America, James directs a team of analysts and associates who track and analyze consumer use of technology products and services. His team conducts original research—surveying hundreds of thousands of households each year—on topics ranging from the spread of broadband to the adoption of wireless devices. This primary research is the largest and longest-running technology research effort in the world and provides Forrester and its clients with the most accurate view of emerging as well as established technologies.

James is regularly consulted for comment on the role of technology in our lives by The Wall Street Journal, The New York Times, London Financial Times, Wired, and USA Today. He has appeared on CNN and CNBC and has been interviewed for National Public Radio's "All Things Considered" and "Morning Edition."

Prior to joining Forrester, James conducted new media research and taught advertising at Syracuse University's prestigious S.I. Newhouse School of Public Communications. During his academic career, James has published articles in journals such as The Journal of Media Economics and The European Journal of Communication. He has presented research at the annual conferences of the International Communications Association, the American Association of Public Opinion Researchers, and the Association for Education in Journalism and Mass Communication.

James holds a Ph.D. (ABD) in mass communications research from Syracuse University and received a master of business administration degree from the University of South Carolina.

Forrester Oval program, Forrester Wave, WholeView, Technographics, TechStrategy, and TechRankings are trademarks of Forrester Research, Inc. Forrester Research, Inc., 400 Technology Square, Cambridge, MA 02139 USA +1 617/613-6000 Fax: +1 617/613-5000 Amsterdam, Austin, Frankfurt, London, San Francisco, Sydney, Tokyo; Web site: www.forrester.com; Email: forrester@forrester.com

(A Research Report Commissioned by Microsoft Corporation and Conducted by Forrester Research, Inc., in 2003)