Pre-Conference Workshops 2025

Dozent: Matthew G. R. Courtney (Nazarbayev University, Kazakhstan) & Kaidar Nurumov (University of Michigan, USA)

Datum: 29.09.2025 (10 bis 18 Uhr)

Ort: GESIS Mannheim

Abstract:

Large Scale Assessment (LSA) studies including PIAAC have complex methodologies, typically combining variance due to complex survey designs and variance due to multiply imputed plausible values (PVs) reflecting a population’s respondents’ level of proficiency. Analysis of such data is often undertaken with the assistance of point-and-click or programming statistical software that is limited to pre-defined functions. While such software provides a reliable toolkit, relying solely on it can restrict the flexibility and types of analyses researchers can conduct. In this workshop we will cover the use of base R to work with PIAAC data from Cycles 1-2. We will start with a short review of weighting, variance estimation, and estimation of PVs, continue with programming the simple descriptive analyses using weights and PVs, and move to regression modeling when PVs are specified as either dependent or independent variables. The focus will be on programming efficient pipelines of recursive functions that can take into account weights and PVs. Moreover, participants will be able to work on data manipulation and the linking of PIAAC data between Cycles 1 and 2, taking into account the linking error. The workshop will emphasize practical applications and programming exercises over theoretical material. We assume that participants have some prior knowledge of basic statistics and are familiar with R programming (at least at an introductory level). By the end of the workshop, participants are expected to manipulate and join PIAAC data as well as write their own recursive functions for descriptive analyses and regression modeling. The workshop will build a solid foundation for conducting one’s own research with LSA data, without relying on the existing pre-programmed software.

Participants should have a computer preloaded with the latest versions of R, RStudio, and public 2023 PIAAC data to practice the analysis.

Daten: PIAAC Public Use Files (Cycle 2)

Software: R

Agenda: comming soon

Literaturempfehlung: Maehler, D. B. (2023). Analysis scripts in Large-Scale Assessments in Education: Methods Series of the Research Data Center PIAAC. Cologne: GESIS – Leibniz Institute for the Social Sciences. doi: https://doi.org/10.21241/ssoar.85846.v2

Dozenten: Ting Zhang & Paul Bailey (American Institutes for Research, USA)

Datum: 29.09.2025 (10 bis 18 Uhr)

Ort: GESIS Mannheim

Abstract: This course offers hands-on training in analyzing 2023 PIAAC data files using the R package EdSurvey, designed for international large-scale assessment data with complex psychometric and sampling designs. Participants will learn how to:

• Process and explore PIAAC data
• Perform descriptive statistics (e.g., plausible value means, group statistics)
• Compare groups or countries means
• Conduct linear and logistic regression with plausible values
• Generate their own plausible values

The course is intended for individuals with basic knowledge of R software and statistical techniques, including statistical inference and regressions. Familiarity with plausible values and sampling theory is beneficial but not required. The workshop begins with an overview of PIAAC studies, plausible values, sampling, weighting, and variance estimation.

Participants should have a computer preloaded with the latest versions of R, RStudio, Edsurvey, and public 2023 PIAAC data to practice the analysis.

Daten: PIAAC Public Use Files (Cycle 2)

Software: R

Agenda: comming soon

Literaturempfehlung: Maehler, D. B. (2023). Analysis Scripts in Large-Scale Assessments in Education: Methods Series of the Research Data Center PIAAC. Cologne: GESIS – Leibniz Institute for the Social Sciences. doi: https://doi.org/10.21241/ssoar.85846.v2