Survey Data Collection
RTI is an industry leader in conducting face-to-face field surveys. We have developed cost-efficient and effective procedures for all components of field surveys, including identifying sample units in the field, selecting and training field staff, developing special protocols for bilingual interviewers, supervising field interviewing, converting refusals and minimizing nonresponse, implementing effective quality control, and managing project administration and reporting.
Since 1959, we have conducted hundreds of field surveys involving hundreds of thousands of respondents. Our expertise is reflected in the strength of our national network, and the wide variety of clients we serve and surveys we conduct.
RTI uses a field force of more than 100 supervisors and 7,000 interviewers geographically distributed throughout the United States and recruited from numerous sources. A database containing information about each interviewer's level of experience, foreign language skills, availability for travel, and performance ratings on past RTI surveys enables us to staff new projects quickly and effectively.
We conduct both national and regional studies (state, city, or neighborhood) for federal, state, municipal, and private clients. We are an industry leader in performing surveys of households, businesses, and special populations (including non-English speaking populations, inmates, children, and the elderly).
RTI has decades of experience in conducting telephone interviews at varying levels of complexity on residential and establishment populations. Advancements have been made in optimizing cost efficiency and reducing survey error, such as the modeling and construction of algorithms for call scheduling, and the development of methods for interviewing individuals on cell phones -- a source of coverage error in random-digit-dialed landline telephone surveys.
Our CATI case management system facilitates the use of a variety of questionnaire software, such as Blaise, CASES, and Web browser-based instruments. RTI maintains a staff of approximately 500 telephone interviewers and institutional contactors. The call center has over 240 production stations for telephone interviewing, Web support, inbound call handling, and tracing operations, with additional capabilities available from a separate on-site training room that is equipped with 60 training stations, and the ability of interviewers to conduct telephone interviews remotely, over a secured integrated connection.
Our call center is built on the latest advances in computer-telephony integration and data networking, using VoIP technology. This infrastructure allows data and voice to be carried over a common data network, providing more efficient and innovative data collection capabilities. Some of these advances include the ability to use digital recording for quality control and training purposes, one-click autodialing, sophisticated inbound call routing, and improved Web data collection support. RTI's VoIP network is connected to the Wide Area Telecommunications Service (WATS) and Federal Telephone System (FTS) lines, providing our clients with the most cost-effective inbound and outbound calling services. The call center is located in a controlled-access environment, ensuring the strict security of all data collection, and operates 7 days a week.
RTI is experienced in designing and conducting mail surveys of persons, households, and businesses in the United States, including subpopulations such as college graduates, nurses, school teachers, school administrators, the elderly, unions, private employers, public agencies, and various industry groups. We use our experience and knowledge to design effective mail survey questionnaires and to carry out efficient mail survey procedures, which maximize response rates.
We recruit survey endorsers and specialists to legitimize surveys and motivate respondents, schedule multiple mailing sequence and timing, and include monetary and other incentives for particular surveys.
RTI uses onsite staff to prepare questionnaires and survey instruments for mailing, to maintain records, and to edit and code responses. A fully automated survey control system monitors the survey instrument from initial distribution or mailing through data collection.
RTI is an industry leader in the development and use of state-of-the-art software systems to improve the quality of survey data. We have created several high-technology systems to facilitate data collection, including
Computer-assisted personal interviewing (CAPI)
RTI is highly experienced in developing computer-assisted interviewing systems to enhance face-to-face data collection methodology. Since 1987, we have conducted over a hundred CAPI surveys both in household and institutional settings. Currently, we are conducting the National Survey on Drug Use and Health (NSDUH), which involves electronic screening and CAPI interviews with over 67,500 respondents 12 years old and older.
Computer-assisted telephone interviewing (CATI)
RTI has used CATI to conduct hundreds of large-scale, complex telephone surveys, involving interviews with hundreds of thousands of persons from households and establishments. We are experienced in conducting random-digit-dial (RDD) and list-sample surveys. Our system is not simply a software package for conducting computer-controlled interviews. Rather, it comprises several sophisticated software subsystems integrated together to handle the various key operational components of a telephone survey, including case management, call scheduling, tracing, and reporting.
Audio computer-assisted self interviewing (ACASI)
RTI pioneered the development of ACASI. Using this tool, the respondent uses headphones connected to a laptop computer to listen to questions that have been digitally recorded to an ACASI program, then keys his/her responses directly into the computer. This methodology has proven to be a highly successful means of gathering information on sensitive or personal topics, due to the added measure of privacy that results from the absence of an interviewer and the use of headphones. To date, we have completed more than 500,000 interviews using ACASI. We have used ACASI to survey a national sample of women on such topics as abortion, sexual partners, and HIV risk behaviors. The National Survey on Drug Use and Health (NSDUH), conducted by RTI, also contains an extensive ACASI component that involves the use of tobacco, alcohol, illicit drugs, and nonmedical use of prescription drugs. The National Inmate Survey, also conducted by RTI, collects sensitive data on sexual victimization from inmates housed in U.S. prisons and jails.
Telephone audio computer-assisted self interviewing (T-ACASI)
Similar to ACASI, T-ACASI allows respondents to use a touch-tone telephone to connect to a computer-based interviewing system. The respondent follows recorded instructions, answering questions and providing data by pressing buttons on his/her telephone. The data are automatically recorded by the T-ACASI system. Because respondents do not have to report to an interviewer and can self-administer the survey in privacy, they are more likely to report sensitive behaviors accurately.
Computer-assisted recorded interviewing (CARI)
RTI developed computer-assisted recorded interviewing (CARI) techniques to monitor interviewer performance, question-administration methods, and data quality in studies involving the use of laptop computers. With CARI, portions of the interview are recorded on the laptop, with the respondent's consent. The audio files are then downloaded onto a disk and made available to project staff so they can listen to the interaction between the interviewer and respondent. As with silent monitoring techniques, which are used for telephone interviews, neither the interviewer nor respondent know when they are being recorded.
As an interviewing tool, this mode is especially suited when the target population is known to be connected to the World Wide Web. As a vehicle for computer-assisted self interviewing, this medium offers greater flexibility and standardization for projects requiring records or data extraction in remote locations. With an Internet server at RTI serving as the central repository, online data collection is now possible in real time from almost any location. One example of our Web-based CADE experience is our work for the U.S. Department of Labor to develop and implement the Occupational Information Network (O*NET) Data Collection Program.
Additionally, RTI has applied Internet-based applications to the collection of survey data and computer-assisted data entry (CADE). Other computer-assisted technology RTI has successfully applied to the collection of survey data includes geographic information systems (GIS). Our GIS specialists developed a cost-efficient, automated, map production system and applied it to several large surveys.
RTI also uses handheld global positioning satellite (GPS) instruments to provide data on geographic context for subsequent analysis. We recently used this technology for a study that examined the accessibility of health care services to the citizens of 12 counties in the Appalachian mountains of western North Carolina -- the Mountain Accessibility Project (MAP).
To maximize participation, produce accurate and complete data, and make efficient use of project funds, RTI survey designers often mix different modes of collecting data into a single survey design. Our survey methodologists and research scientists work with clients to develop the best approach (e.g., a mail survey followed by telephone interviews or by personal interviews).
To minimize the effects from an individual mode, we maintain comparable wording across modes. To measure mode effects analytically, we maintain comparable data processing procedures across modes and include all appropriate variables. Our design teams attend particularly to planning and implementing mixed-mode studies directed to special populations.
Our experience ensures that each mode we select is a viable way to collect data from a given study group, and that our mixed-mode data collections will achieve the following objectives:
Enhance participation rates while controlling costs
A mixed-mode design can allow us to achieve high response rates while reducing the overall cost of data collection. In a study of Medicare, RTI developed a data collection plan that incorporated mail and telephone interviewing. Collecting the majority of the data by mail kept project costs low, while telephone follow-ups ensured a high overall response rate.
Improve respondent reporting of sensitive behaviors
RTI is a leader in collecting data on sensitive topics and behaviors. To maximize accuracy, we often adopt a mixed-mode approach to collect sensitive information privately or anonymously. To help with this, we developed technology for confidential computer interviewing known as audio computer-assisted self-interviewing (ACASI). In ACASI, the respondent uses headphones and a laptop computer to listen to digitally recorded questions and type answers directly into the computer. To date, our work indicates that the increased privacy offered by ACASI yields a higher number of reports of sensitive behaviors than other, more traditional survey modes.
Explore methodological issues pertaining to data collection
RTI's survey methodologists specialize in designing experimental research to add to the existing body of knowledge on mode effects. For example, we studied mode effects in the National Health Interview Survey to examine the difference between the data quality of face-to-face interviews and that of telephone interviews.
Focus groups are a valuable research tool to rapidly collect qualitative data by observing the interactions in a group of respondents. RTI has extensive experience conducting high-quality focus groups with targeted populations, including asthma patients, foster parents, teachers, field interviewers, high school students and their parents, and Medicare recipients.
In conducting a focus group, RTI researchers
Select participants and arrange for the meeting place
Identify and plan the topic(s) of discussion
Guide and moderate the discussion
Listen to, observe, and record the group interactions
Analyze the product of the discussion
We recruit participants specifically for each separate group. When volunteer bias is a concern, we use random sampling methods to identify and recruit focus group participants. We also have experience identifying and recruiting representative samples of individuals based on specific characteristics, such as age, income, gender, race, and degree of literacy. Careful preparation of discussion guides and conduct of formal training sessions are standard procedures at RTI.
RTI has highly skilled moderators to lead focus group discussions and facilitate interactions among group members. High-quality interactions provide relevant, useful information for researchers. Our state-of-the-art survey facilities in Research Triangle Park, NC, and the Washington, DC, metro area benefit our clients by providing greater flexibility, proximity to clients, and variation in tested populations. RTI also conducts focus groups at client sites and other locations to meet specific project needs.
Much of the research RTI has conducted over the last 10 years has focused on sensitive behaviors. During this period, we have conducted surveys concerning HIV risk behaviors, sexual activity, illicit drug use, abusive situations, and abortion. We have found that respondents are concerned about the confidentiality of their responses, about reporting their behaviors to an interviewer, and about reporting their behaviors in their homes, where other household members might overhear.
RTI has used the findings of these and other studies to improve the quality of the data we collect on sensitive topics. Since 1990, we have conducted over 500,000 interviews involving sensitive topics and have used our research to develop and implement methodologies designed to obtain complete and accurate data, including self-administered questionnaires, computer-assisted interviewing tools, and neutral-site interviews.
We have conducted a number of studies on the reporting of sensitive behaviors, including
Effect of mode of administration on the reporting of drug usage
Ability of illiterate respondents to complete self-administered questionnaires if the questions-and-answer choices are read to them by a recorded voice
Effectiveness of neutral-site interviews in eliciting reports of abortions
Impact of incentives on the reporting of sensitive behaviors
Effectiveness of audio computer-assisted self-interviewing (ACASI) in increasing reports of sensitive behaviors
In this mode, respondents complete an answer sheet handed to them by a field interviewer, place the completed answer sheet(s) in a self-addressed, stamped envelope, and return the completed answer sheet by mail to RTI. Our methodological research into this procedure shows that it results in increased reports of illicit drug use and that the "harder" the drug (e.g., cocaine relative to marijuana), the greater the increase in reporting.
Computer-Assisted Interviewing Tools
RTI pioneered the development of a number of computer-assisted interviewing tools, including:
Audio computer-assisted self-interviewing (ACASI)
Touch-screen ACASI for surveys of low literate and illiterate groups
Telephone audio computer-assisted self-interviewing (T-ACASI)
These tools offer greater privacy and confidentiality for surveys about sensitive topics and behaviors.
Another method RTI has successfully used to administer surveys on sensitive topics is neutral-site interviews, where respondents are asked to go to a prearranged neutral site to be interviewed. Neutral-site interviews have resulted in a significantly higher number of reports of sensitive behaviors than in-home interviews, suggesting the importance of complete privacy during the administration of questions of a sensitive nature.
RTI has extensive experience tracing sample members for list sample and follow-up surveys. Tracing is customized for each project and incorporates the cohort profile, known tracing identifiers, and supplemental information derived from known data. Appropriate databases are selected based on the age and source of the information, the method of data collection, and the project's schedule and budget.
Tracing is cost-effectively implemented and progressively more intensive so that expensive procedures are applied only when necessary. This multi-phased, custom approach ensures maximum success and an efficient use of resources.
Tracing operations at RTI comprise the following:
RTI performs advance tracing using nationwide databases to locate subjects who have moved or died. We access the most comprehensive and current files and then review the information to either confirm or remand it.
Our tracing experts use a computer-assisted tracing management system to keep a history of calls to subjects and contacts. This system uses roster and comment features that tell tracers who has been contacted, what information has been developed, and other information or leads unique to the case. Tracing specialists contact friends and relatives, use crisscross directories to identify neighbors, and contact directory assistance for possible updates.
RTI has access to multiple nationwide databases to locate and verify current addresses and telephone numbers, including credit bureaus, consumer and census-oriented databases, state Department of Motor Vehicle records, the U.S. Postal Service National Change of Address database, Telematch, death indices, Social Security files, ProCD, and Internet directory services. The diversity and flexibility of these systems let us efficiently tailor our tracing methods based on the information available about the particular cohort.
Field interviewers are trained to establish trust and elicit information from a subject's relatives, neighbors, schools, business associates, and government agencies. Interviewers pursue all leads from the subject's last known location. To ensure that all reasonable tracing steps are taken, we provide field interviewers with a checklist of appropriate and cost-efficient steps to follow for a given cohort.