New Technologies Including Hearing-Impaired Respondents in CATI Surveys This article is written by Kara L. Sumner and Courtney Nordhagen of Clearwater Research, Inc. It is based on a paper presented at the Centers for Disease Control and Prevention's 15th Annual Behavioral Risk Factor Surveillance System Conference held in Atlanta in May, 1998. Hearing-impaired respondents can now be included in Ci3 CATI telephone surveys with the addition of some specialized hardware and software. Adding this option to a telephone survey allows inclusion of the 28 million hearing-impaired persons formerly unable to participate in telephone surveys. In February 1998, Clearwater Research, Inc., a full-service, custom marketing research firm in Boise, Idaho, began conducting telephone interviews with hearing-impaired respondents for the Texas Department of Health (TDH). The TDH received funding through a grant to include questions about adults with disabilities as part of the Behavioral Risk Factor Surveillance System (BRFSS), an on-going, state-based telephone surveillance system supported by the Centers for Disease Control and Prevention. The grant funding prompted the TDH to seek ways to further enhance the data collected about people with disabilities in the BRFSS. Clearwater Research reviewed the options available for allowing hearing-impaired people to communicate in real-time over a telephone line, including relay services, teletypewriter hardware, and software options. We selected a NexCom 300 modem and NexTalk software, which uses the respondent's personal computer. The modem supports telecommunications devices (TDD) calls and the software allows users to program and store survey questions so they do not need to be typed during interviews. Questions are retrieved and transmitted using only a few keystrokes. The figure below shows the menu from which an interviewer retrieves a survey question to transmit to a hearing-impaired respondent. (image) Using CATI for TDD Interviews Because the NexTalk software cannot control skip patterns, our TDD-trained interviewers use Ci3 CATI during TDD interviews, multi-tasking between NexTalk and Ci3 CATI. The interviewer sees, in Ci3 CATI, which question needs to be transmitted to the hearing-impaired respondent, retrieves that question in NexTalk, and then transmits it. When the respondent replies, the interviewer multi-tasks back to Ci3 CATI and enters the response. This procedure allows data entry to occur simultaneously with the interview; and, the interviewer knows which questions to transmit to the respondent, because the CATI survey automatically follows skip patterns. The figure below displays the interviewer's screen when using both NexTalk and Ci3 CATI to conduct the interview. The right side of the figure shows Ci3 CATI. For dynamic questions that use text unique to each survey (for example, using a contact name), NexTalk has the capability to let the interviewer add the unique text to the programmed question prior to transmission. (image) Limitations of the Process Conducting interviews with hearing-impaired persons has limitations. Many questions require modification because they can not be easily translated or understood by respondents using American Sign Language (ASL), which uses different syntax and has a limited vocabulary. Another limitation relates to the respondent's TDD capabilities. Not all hearing-impaired persons have the software or hardware to allow them to communicate by telephone. Many who use TDD hardware (e.g., teletypewriters) are limited to viewing only 24 characters at a time. Thus, questions need to be brief enough that respondents can remember and understand them. Another limitation is time. A 10-minute telephone survey could take anywhere from 20 minutes to an hour to administer via TDD. Despite these limitations, our experience demonstrates that it is possible to conduct telephone interviews with hearing-impaired respondents. We expect to see these capabilities improve as the prevalence of personal computers and Internet use increases. Back to Sawtooth News Index