This 4-day course in ‘advanced survey design’ takes students beyond the introductory courses offered in BA and MA programmes, and will discuss the state of the art in both the design of surveys, and the analysis of survey data. We explicitly discuss new ways to analyse text data and sensor data generated by modern surveys. Course participants must be proficient in working with a package for statistical software (Stata, SPSS or R). Course materials have been developed in R, but most exercises can be done in SPSS.
This 4-day course in survey design takes students beyond the introductory courses offered in BA and MA programmes, and discusses current issues in one of the most important data collection methods: surveys. Specifically, it focuses on doing surveys in the Internet-era. It will discuss how to collect data in online surveys, using smartphones, and mixing surveys with Big data. It combines short 1-hour lectures with exercises on most of the topics discussed. Course participants must be proficient in working with a package for statistical software (Stata, SPSS or R). Course materials are prepared for working with R, while most of the exercises will also work with SPSS.
The course is interesting for social scientists or statisticians at the PhD level or beyond, working on academic research projects. Two courses are offered in the Utrecht Summer school that slightly overlap with this course: S15 (Survey Research: Design, Implementation and Data Processing) and S16 (Survey Research: Statistical Analysis and Estimation). The current course is however more advanced and more focused on survey research within the academic (university) setting, as well as focused on current issues related to mobile surveys and Big Data.
We expect students to have quite extensive knowledge of survey research (for example by using survey data extensively) and have knowledge of statistics at the MSc level for social scientists (the general linear model). Note that you have to upload a brief motivation letter (up to 300 words) with your application, in which you write a short list of your experience with surveys and statistical analyses. Please include in the motivation letter what software you are proficient in, and what courses related to survey design you have taken. Note that IOPS students do not have to do this.
Note: Participants need to bring a laptop computer to the course, with SPSS or alternatively with R (https://www.r-project.org/) installed
Day 1: Total Survey Error Paradigm. We will discuss the Total Survey Error Framework in the context of ‘designed Big data’ (where surveys are complemented with organic data). The principles of probability sampling designs given different survey modes (simple random, cluster, stratified and multistage sampling). We will discuss sampling from geo-data or digital trace data.
Day 2: Questionnaire design, with specific attention to mixed-mode and mobile-device surveys. We will discuss ways in which questionnaires can be designed to account for respondents accessing them on mobile devices and when extensive information about the sample is available (e.g., through records), or when surveys include the collection of Big data.
Day 3: We will discuss the collection of process data or Paradata (what are they and how to use them). Further, we discuss nonresponse and weighting (both design and nonresponse weights) to correct for biases in surveys
Day 4: Surveys and big data. Working with geo-location data to enrich survey data. Analyzing text data (both manual coding and through machine learning), and other sensor data (pictures)
Tuition fee for PhD students from the Faculty of Social and Behavioural Sciences from Utrecht University will be funded by the Graduate School of Social and Behavioural Sciences.
PhD candidates from IOPS will be covered by IOPS.