The course is intended as an introduction to the field, taught at the graduate level. It introduces a set of principles of survey design that are the basis of standard practices in the field. The course examines research literatures that use both observational and experimental methods to test key hypotheses about the nature of human behavior that affect the quality of survey data. It also presents statistical concepts and techniques in sample design, execution, and estimation, and models of behavior describing errors in responding to survey questions. The course uses total survey error as a framework to discuss coverage properties of sampling frames; alternative sample designs and their impacts on standard errors of survey statistics; alternative modes of data collection; field administration operations; the role of the survey interviewer; impacts of nonresponse on survey statistics; the effect of question structure, wording and context on respondent behavior; models of measurement error; postsurvey processing; and estimation in surveys.
This is a hands-on course in which small groups of students will develop and deploy a web survey, collect and analyze data from actual respondents, and write a report suitable for a client. Lectures and class discussion will focus on the scientific literature in web survey methodology with a focus on practical implications.
This course is shared between the University of Michigan and the University of Maryland. Students from the two universities will sit in different classrooms but attend the same video-mediated class/lecture.
This is an integrated undergraduate/graduate course. Students taking the course for graduate credit are assigned additional readings that emphasize the conceptual and theoretical aspects of online survey data collection and are given longer versions of the course quizzes to assess their understanding of these concepts. They will also serve as the leaders of their project groups.
Methods and Theory of Sample Design is concerned with the theory underlying the methods of survey sampling widely used in practice. It covers the basic techniques of simple random sampling, stratification, systematic sampling, cluster and multi-stage sampling, and probability proportional to size sampling. It also examines methods of variance estimation for complex sample designs, including the Taylor series expansion method, balanced repeated replications, and jackknife methods.
This course is crosslisted with the Biostatistics Department.
This course is the first semester of a two-semester sequence that provides a broad overview of the processes that generate data for use in social science research. Students will gain an understanding of different types of data and how they are created, as well as their relative strengths and weaknesses. A key distinction is drawn between data that are designed, primarily survey data, and those that are found, such as administrative records, remnants of online transactions, and social media content. The course combines lectures, supplemented with assigned readings, and practical exercises. In the first semester, the focus will be on the error that is inherent in data, specifically errors of representation and errors of measurement, whether the data are designed or found. The psychological origins of survey responses are examined as a way to understand the measurement error that is inherent in answers. The effects of the mode of data collection (e.g., mobile web versus telephone interview) on survey responses also are examined.
This is the second course in a two-semester sequence that provides a broad overview of the processes that generate data for use in social science research. Students will gain an understanding of different types of data and how they are created, as well as their relative strengths and weaknesses. A key distinction is drawn between data that are designed, primarily survey data, and those that are found, such as administrative records, remnants of online transactions, and social media content. The course combines lectures, supplemented with assigned readings, and practical exercises. The second semester builds on the discussion of survey mode during the first semester, considering the role played by interviewers in telephone and in-person surveys and their effects on the data collected. Students next are introduced to the methods for extracting and repurposing found data for social science research. Methods for the classification of text, with an emphasis on automated coding methods, are introduced and selected applications considered (e.g., coding of open-ended survey responses, classification of the sentiments expressed in social media posts). Issues in using survey data and administrative records to measure change over time (longitudinal comparisons) are explored. The term concludes with an examination of methods for evaluating the quality of both designed and found data.
Methods of Survey Sampling is a moderately advanced course in applied statistics, with an emphasis on the practical problems of sample design, which provides students with an understanding of principles and practice in skills required to select subjects and analyze sample data. Topics covered include stratified, clustered, systematic multi-stage sample designs; unequal probabilities and probabilities proportional to size, area, and telephone sampling; ratio means; sampling errors; frame problems; cost factors; and practical designs and procedures.
This course focuses on the development of the survey instrument, the questionnaire. Topics include wording of questions (strategies for factual and non-factual questions), cognitive aspects, order of response alternatives, open versus closed questions, handling sensitive topics, combining individual questions into a meaningful questionnaire, issues related to question order and context, and aspects of a questionnaire other than questions. Questionnaire design is shown as a function of the mode of data collection such as face-to-face interviewing, telephone interviewing, mail surveys, diary surveys, and computer-assisted interviewing.
Survey data are only as meaningful as the answers that respondents provide. Hence, the processes that underlie respondents' answers are of crucial importance. This course draws on current theorizing in cognitive and social psychology pertaining to issues such as language comprehension, information storage and retrieval, autobiographical memory, social judgment, and the communicative dynamics of survey interviewing, to understand how respondents deal with the questions asked and how they arrive at an answer.
This is a wide-ranging graduate seminar in which several program faculty members join with the students in attempting to solve design issues presented to the seminar by clients from the private, government, or academic sectors of research. Readings are selected from literatures not treated in other classes, and practical consulting problems are addressed.
This is the first in a two term sequence in applied statistical methods covering topics such as regression, analysis of variance, categorical data, and survival analysis.
This builds on the introduction to linear models and data analysis provided in Statistical Methods I. Topics include: Multivariate analysis techniques (Hotelling's T-square, Principal Components, Factor Analysis, Profile Analysis, MANOVA); Categorical Data Analysis (contingency tables, measurement of association, log-linear models for counts, logistics and polytomous regression, GEE); and lifetime Data Analysis (Kaplan-Meier plots, logrank test, Cox regression).
Applications of Statistical Modeling, designed and required for students on all three tracks of the two programs in survey methodology, will provide students with exposure to applications of more advanced statistical modeling tools for both substantive and methodological investigations that are not fully covered in other MPSM or JPSM courses. Modeling techniques to be covered include multilevel and marginal modeling techniques for clustered or longitudinal data (with applications to methodological studies of interviewer effects and modeling trends in the Health and Retirement Study), structural equation modeling (with an application of latent class models to methodological studies of measurement error), and classification trees (with an application to prediction of response propensity). Discussions and examples of each modeling technique will be supplemented with methods for appropriately handling complex sample designs when fitting the models. The class will focus on essential concepts, practical applications, and software, rather than extensive theoretical discussions.
Each course will cover a different area of survey methodology. Titles and descriptions will be listed each term in the University Time Schedule.
Directed research on a topic of the student's choice. An individual instructor must agree to direct such research, and the requirements are specified when approval is granted.
This introductory course on the analysis of data from complex sample designs covers the development and handling of selection and other compensatory weights; methods for handling missing data; the effect of stratification and clustering on estimation and inference; alternative variance estimation procedures; methods for incorporating weights, stratification, clustering, and imputed values in estimation and inference procedures for complex sample survey data; and generalized design effects and variance functions.
SurvMeth 720 (Fall term) and 721 (Winter term)
These courses review the total error structure of sample survey data, reviewing current research findings on the magnitudes of different error sources, design features that affect their magnitudes, and interrelationships among the errors. Coverage, nonresponse, sampling, measurement errors, interviewer effects, questionnaire effects, and mode of data collection effects are reviewed. Statistical and social science approaches to the error sources are compared.
Empirical social scientists are often confronted with a variety of data sources and formats that extend beyond structured and handleable survey data. With the emergence of BigData, especially data from web sources play an increasingly important role in scientificresearch. However, the potential of new data sources comes with the need for comprehensive computational skills in order to deal with loads of potentially unstructured information. Against this background, the first part of this course provides an introduction to web scraping and APIs for gathering data from the weband then discusses how to store and manage (big) data from diverse sources efficiently. The second part of the course demonstrates techniques for exploring an dfinding patterns in (non-standard) data, with a focus on data visualization. Tools for reproducible research will be introduced to facilitate transparent and collaborative programming. The course focuses on R as the primary computing environment, with excursus into SQL and Big Data processing tools.
This course is one of the fundamental 3 courses required by all students in the Master’s Program in Survey Methodology, and focuses on the fundamentals of statistical inference in the finite population setting.
The course is design to overview and review fundamental ideas of making inferences about populations. It will emphasize the basic principles of probability sampling; focus on differences between making predictions and making inferences; explore the differences between randomized study designs and observational studies; consider model-based vs. design-based analytic approaches; review techniques designed to improve efficiency using auxiliary information; and consider non-probability sampling and related inferential techniques.
Inference from complex sample survey data covers the theoretical and empirical properties of various variance estimation strategies (e.g., Taylor series approximation, replicated methods, and bootstrap methods for complex sample designs) and how to incorporate those methods into inference for complex sample survey data. Variance estimation procedures are applied to descriptive estimators and to analysis techniques such as regression, analysis of variance, and analysis of categorical data. Generalized variances and design effects are presented. Methods of model-based inference for complex sample surveys are also examined, and the results are contrasted with the design-based type of inference used as the standard in the course. The course will use real survey data to illustrate the methods discussed in class. Students will learn the use of computer software that takes account of the sample design in estimation. Students will carry out a research and analysis project, using techniques and skills learned during the course. A paper describing the student's research will be submitted at the end of the course, and each student will give a short presentation of his/her findings.
This course is an advanced course in selected topics in survey sampling. Topics to be covered include: estimation and imputation approaches, small area estimation, and sampling methods for rare populations. A selection of additional topics, chosen by the instructor, will also be covered. Examples of such additional topics include: sample designs for time and space, panel and rotating panel survey designs, maximizing overlap between samples, controlled selection and lattice sampling, sampling with probabilities proportionate to size without replacement, multiple frame sampling, adaptive cluster sampling, capture-recapture sampling, sampling for telephone surveys, sampling for establishment surveys, and measurement error models. Both applied and theoretical aspects of the topics will be examined.
This course is a statistical methods class appropriate for second year Master’s students and PhD students. The course will be a combination of hands-on applications and general review of the theory behind different approaches to sampling and weighting. Topics covered include:
- Sample size calculations using estimation targets based on relative standard error, margin of error, and power requirements;
- Use of mathematical programming to determine sample sizes needed to achieve estimation goals for a series of subgroups and analysis variables;
- Resources for designing area probability samples;
- Methods of sample allocation for multistage samples;
- Steps in weighting, including computation of base weights, nonresponse adjustments, and uses of auxiliary data;
- Nonresponse adjustment alternatives, including weighting cell adjustments, formation of cells using regression trees, and propensity score adjustments;
- Weighting via poststratification, raking, general regression estimation, and other types of calibration.
This course describes modern practices in the administration of large-scale surveys. It reviews alternative management structures for large field organizations, supervisory and training regimens, handling of turnover, and multiple surveys with the same staff. Practical issues in budgeting of surveys are reviewed with examples from actual surveys. Scheduling of sequential activities in the design, data collection, and processing of data is described.
This is the first course in a two-term introduction to the integration of social science and statistical science approaches to the design, collection, and analysis of surveys. The seminar will focus on six to eight areas of statistical and methodological literature that have benefited from alternative approaches. Students demonstrate mastery of those literatures through critical review papers, ideas for extensions of the literature, and empirical projects related to research reviewed.
This is the second course in a two-term seminar designed to develop skills in the identification of research problems, specification of hypotheses/theorems to extend current understanding of the field, and planning for original research. A common set of readings in four to six advanced research activities of the faculty are studied, with the faculty engaged in research discussing areas of potential innovation. Students present and critique oral and written proposals for research.
These special seminar courses address specific research problems currently under study by faculty members. The topics will be announced in the University Time Schedule each academic term.
Directed research on a topic of the student's choice. An individual instructor must agree to direct such research, and the requirements are specified when approval is granted.
Laboratory for study of survey methodology practices and principles.
Election for dissertation work by doctoral student not yet admitted as a candidate. Students doing dissertation work prior to achieving candidacy should register for SurvMeth 990 for that portion of their schedule spent on dissertation work.
Graduate School authorization for admission as a doctoral candidate. The defense of the dissertation (the final oral examination) must be held under a full term candidacy enrollment period. Students who have advanced to candidacy for the Ph.D. are required to register for SurvMeth 995 in any term when they are consulting with members of their dissertation committee or using the library or other facilities of the University. If the student is to be engaged in a period of study away from the University, the student should file a Certification for Detached Study in advance. Students doing dissertation work prior to achieving candidacy should register for SurvMeth 990 for that portion of their schedule spent on dissertation work.