led by Maddi Bunker, PhD student at the University of Edinburgh
Many popular social survey data sets — including: Growing Up in Scotland, Millennium Cohort Study, Understanding Society, and more — use complex sampling designs which require bespoke techniques to accommodate non-random probabilities of selection and/or to establish nationally representative estimates. Approaching these data can be challenging at first, especially when navigating the expansive array of statistical packages in the R programming language.
Designed for quantitative and quants-curious social scientists, this workshop provides an introduction to the “Survey” package in R, including: setting a survey design object, calculating survey weighted descriptive statistics, regression and other common quantitative techniques. We will also make use of the “tidyverse” packages and commands for survey data manipulation.
Learning outcomes:
Participants should leave the workshop with the tools and confidence required to approach statistical analysis in R when using data which are derived from a complex sampling design. The tutorial will provide an example of a reproducible workflow for survey data analysis using R Markdown documents and can be a “jumping off” point for further statistical techniques in R.
Prerequisites:
Prior knowledge of R and RStudio environments will be beneficial. Alternatively, researchers confident with survey weighting in Stata’s “svy” suite of commands, who have some familiarity with R, may find this workshop beneficial as an introduction to the software.
Participants should bring their own lunch to this full-day workshop.