About the Survey Research Laboratory
SRL is a research and service unit established in 1964. It is a division of the University of Illinois at Chicago's College of Urban Planning and Public Affairs that provides survey research services to the faculty, staff, and students of the University of Illinois campuses; other academic institutions; local, state, and federal agencies; and others working in the public interest.
SRL has offices on the Chicago and Urbana-Champaign campuses of the university and can undertake complete survey projects from initial study design through data analysis. It also can conduct partial survey work and provide consultation on survey problems on a fee-for-service basis.
SRL engages in highly customized work in the following areas:
(Click on each topic for more information.)
Address-based sampling (ABS)
Sampling of residential homes from a database with nearly universal coverage. The approach uses an address database rather than a traditional block-listing approach. The United States Postal Service’s Delivery Sequence File is a frequently used frame for ABS.
Area probability sampling
A sampling technique in which geographic areas are sampled, typically as part of a multistage design. In the first stage, primary sampling units (PSUs) are selected. These units may be counties or Census tracts. In the second stage, secondary sampling units (SSUs) are selected. These may be Census tracts within sampled counties or Census blocks within sampled tracts. The next stage is often a residential housing unit, and the ultimate sampling unit is an individual in the housing unit.
Area probability surveys
A face-to-face survey in which the sample is selected using area probability sampling. SRL has significant ongoing experience conducting area probability surveys in Chicago.
Behavior coding involves the systematic coding of recordings of respondent-interviewer interactions. Over the past several decades, it has been used for questionnaire testing and to evaluate interviewer performance. SRL staff have employed this methodology to better understand racial/ethnic variability in the cognitive processing of survey questionnaires.
Case control studies
A case control study is an observational study in which two groups of subjects with different outcome conditions are compared on characteristics thought to be related to the outcome. For example, a sample of cancer patients may be matched to a sample of healthy subjects with similar characteristics (age, race, income, education, geography, etc).
A survey focusing on the social environment or climate of the respondents. They are often used in business or university settings to assess characteristics like collegiality, institutional support, hostility, opportunities for advancement, etc. SRL has experience conducting higher education climate surveys in institutions of varying size.
Cognitive assessment of questionnaires
The use of cognitive interviewing to evaluate survey questionnaires was pioneered in the 1980’s by then-SRL Deputy Director Seymour Sudman. It has since been widely established as an important tool for instrument development. Although several variations of this methodology exist, all involve in-depth interviewing and probing to learn how respondents interpret survey questions, retrieve the information necessary to answer them, format judgments, and report their answers.
Complex sample design analysis
Analysis of complex data refers to the incorporation of sample design elements (e.g. stratification, clustering, unequal sampling probabilities) into the analysis to insure that estimates and variances are computed properly. Several statistical analysis programs include modules or procedures that allow the user to input variables identifying strata and clusters as well as sample weights, number of stages of sampling, and finite population correction factors.
Computer-assisted programming and/or interviewing
Computer-assisted interviewing (CAI) can take many forms: computer-assisted telephone interviewing (CATI); computer-assisted personal interviewing (CAPI); computer-assisted self interviewing (CASI), with or without an audio (ACASI) component. SRL has experience programming and collecting data with all of these formats. The advantages of CAI data collection are significantly improved data quality, as the programming ensures that illegal or out-of-range answers to closed-ended questions cannot be entered, contingency checks can be programmed flagging any contradictory answers to previous questions, question wording can be tailored according to the respondents prior answers, and complicated skip patterns can be programmed. SRL programmers are proficient in CASES and BLAISE interviewing software. Programming support services can be obtained with or without SRL involvement in data collection.
Cross-cultural survey methodology
When collecting survey data across respondents from multiple cultural groups, typically defined as different racial/ethnic groups or as residents of different nations, the comparability of responses must be considered, in addition to their reliability and validity. Over the past 20 years, considerable progress has been made in addressing the comparability of cross-cultural survey data. SRL Director Timothy Johnson has edited an award-winning book (Survey Methods in Multinational, Multiregional and Multicultural Contexts) and a special issue of Public Opinion Quarterly concerned with cross-cultural survey methodologies.
Data entry and validation
SRL uses the University of California-Berkeley CASES software for data entry and data set construction. The CASES instrument is programmed to match the questionnaire and displays the valid numeric responses at each item. The system limits the coder to only those answers. Any input not conforming to the programmed range will not be accepted, and the coder is prompted to enter a valid response. During data entry, any necessary coding decisions on ineligible responses are made and recorded by the coder. SRL CASES validation utilizes double-data entry verification. Double-data entry verification involves the data entry of a given questionnaire for a second time by a different coder. CASES compares the second data input of each question against the original data, and informs the coder if there is a discrepancy. The coder must then correct the data input before proceeding. During validation, 20% of the questionnaires on a given study are selected at random. Validation is usually completed by supervisory staff. If an individual coder’s work is found to have an error rate of more than 0.5%, an additional 20% of that coder’s work is validated.
Face-to-face data collection
For information regarding the specifics and cost of focus groups in our Chicago office, download Focus Groups at SRL.
Hard-to-reach population sampling
Sampling of hard-to-reach populations includes methods that allow researchers to efficiently gather a sample of respondents who are difficult to sample, identify, contact, or interview. These populations typically make up a small percentage of the population, making standard screening procedures typically used with telephone or face-to-face surveys prohibitively expensive. Methods for sampling hard-to-reach populations include snowball sampling, respondent driven sampling, network sampling, venue sampling, and adaptive cluster sampling, among others.
Health survey research methodology
A unique set of methods have been developed to meet the special needs of surveys concerned with health-related topics. For almost 50 years, SRL has been a leader in the development of these methodologies. This includes the coordination of the last five national conferences concerned with health survey research methods (links to the proceedings from those conferences can be found here: www.srl.uic.edu/links/proceedings.html) and the publication of the recent Handbook of Health Survey Methods by SRL Director Timothy Johnson.
An intercept survey is an alternate research method for reaching respondents that would be difficult to identify via more traditional methods. Intercept surveys are needed when no frame exists of the target population and the population is congregated in specific areas. SRL has experience with intercept surveys at various locations, including train stations, bus stops, airports, youth drop-in centers, and clinic waiting rooms.
SRL can be retained to train your field staff on proper interviewing techniques. We place a strong emphasis on the integrity of our field staff procedures. This emphasis extends from the recruitment and training of interviewers through data collection, quality control, refusal conversion, and data security. New interviewer hires receive general training on basic interviewing principles and procedures before the more specific training required of each individual study. SRL validates approximately 10% of each interviewers’ work throughout data collection.
Longitudinal (panel) surveys
Panel studies involve the collection of data over time from the same sample of respondents. Unlike other forms of longitudinal studies, panels allow for the study of individual behavior change over time. SRL has considerable experience with panel studies, particularly the important task of tracking respondents between waves of data collection.
Many studies require mixed modes of data collection to achieve the desired results, and when time and budget allow, are often recommended to boost response and minimize nonresponse bias. Examples of mixed-mode surveys include mail surveys with telephone or Web options to nonrespondents.
The analysis of the effects of frame or auxiliary data on the likelihood that respondents complete a survey. The purpose is to ascertain the degree to which nonresponse is affecting outcome variables of interest.
Evaluation of health and social programs typically requires the use of both survey and related social science methodologies, including qualitative interviewing, analyses of various administrative databases, record abstractions, and content analyses, to name a few. SRL has considerable experience in the conduct of program evaluations, including recent evaluations of the UIC Center for Clinical and Translational Research and Center for Population Health and Health Disparities.
Whether or not you engage SRL to assist with your data collection, SRL can assist you with questionnaire development. Designing a questionnaire is an important aspect of conducting a survey, as it affects the quality of data collected using it. The current literature offers a clear science to designing a good questionnaire that will yield data that are reliable and valid. To this end, each questionnaire is reviewed by members of the SRL Questionnaire Review Committee (QRC) to ensure that it follows principles of good questionnaire construction. The QRC review will bring scientific principles to bear on your specific questionnaire during an iterative process in which we consult with you at various stages to revise and refine the data collection instrument. The committee is composed of academicians and practitioners at SRL who have extensive experience in both questionnaire construction and survey administration. This composition ensures that the review covers both theoretical and practical aspects of optimal design.
SRL has significant experience recruiting subjects, whether from your lists or targeted population groups. We can help you recruit subjects for cognitive interviews, focus groups, or in-depth interviews. We create a customized approach to recruitment that best fits your needs.
Sample frame development
A sample frame is the source list from which a survey sample is drawn. In practice, sample frames often do not exist or are limited in their coverage, requiring the use of multiple sources to create a single usable frame. For example, a sample of communities based on characteristics of the community (e.g., income and crime rates) may require acquisition and the merging of data from multiple sources (Census data and FBI crime data) before a sample can be drawn.
Survey data analysis (from basic to complex)
Analysis of survey data can cover everything from descriptive univariate statistics of the variables measured to complex multivariate and/or multilevel analyses such as structural equation modeling or survival analysis.
Are you unsure of how to design your study? Let SRL’s senior methodology experts help you design the most optimal design that fits your needs – including your available budget and schedule. No two studies SRL conducts are alike; let us customize the right design for your needs.
The conduct of survey research often requires making decisions that have important ethical ramifications. For example, it is important to recognize that many strategies that researchers often undertake to increase response rates may be viewed as coercive by some respondents. SRL staff have considerable experience confronting these ethical issues and adapting methodologies to ensure that our work always conforms to the highest ethical standards. Several members of SRL’s staff have broad experience serving on UIC’s Institutional Review Boards (IRBs), which are responsible for the ethical oversight of social research at our university.
Over the last 50 years, SRL has incorporated many experiments into our surveys that have contributed to the advancement of the science of survey research. Examples of survey method experiments include question or response wording, question order randomization, and the effects of incentives on respondent cooperation. If you are interested in survey experiments, we can help tailor the right test to your study.
Survey News Bulletin
As a service to the research community at the University of Illinois, SRL regularly e-mails brief news bulletins that provide tips to researchers regarding best practices in the conduct of survey research. These are available free to anyone interested in registering to receive them, which can be done at www.srl.uic.edu/Publist/bulletin.html.
Total survey error assessment
Total survey error assessment is the assessment of the difference between a population parameter and the estimate of that parameter from a survey or census. It incorporates both sampling error and nonsampling error. Sampling error is the difference between a sample statistic and the true population value. It is due to the fact the estimate is based on a sample rather than the entire population. Nonsampling error includes frame under- or overcoverage, errors in sample selection, errors due to poor question wording, and errors in data processing. Total survey error assessment takes into account the contribution of each of these to the accuracy of survey estimates.
Survey research in urban environments can be particularly challenging for a variety of important reasons, including respondent inaccessibility due to long commutes, the prevalence of restricted access buildings, and residential fear of crime. SRL has decades of direct experience confronting these challenges and has developed a set of best practices for conducting research in urban environments.
Webinars in survey research methodology
SRL offers a series of free lunch-hour webinars concerned with specific aspects of survey methodology during the Spring and Fall semesters each year. To learn more about upcoming webinars and to register to participate, visit www.srl.uic.edu/seminars.htm.