Teacher Trajectories Across Colombia During the COVID-19 Pandemic: A Hierarchical Growth Model Analysis (2017-2023)

• In Event: Teacher workforce dynamics in Latin America: Navigating system pressures and professional agency across Chile, Colombia, and Peru

Proposal

Colombia’s education system faced unprecedented challenges during 2020-2021, as the country experienced long lockdown-driven school closures while simultaneously confronting increased school shutdowns due to demographic shifts. The combination of closing for 152 instruction days (76% of the school year) between January 2020 and May 2021 (OECD, 2021; Abadía et al., 2023) and over 6,000 documented school closures between 2019 and 2024 (LEE, 2024) created intense disruption to educational provision across Colombia’s 32 departments (the country’s administrative units). This crisis affected a system already characterized by pronounced regional divisions inequalities (García et al., 2021), where students born in rural areas and departments disjointed from the infrastructure that connects the country with the global markets are disproportionately affected by the lasting imprints of decades of armed conflict, which have upended educational infrastructure, investment, and workforce stability (Bertoni et al., 2023).

The burden of navigating this disruption fell disproportionately on teachers, yet little research has examined how Colombia’s teaching workforce responded to these simultaneous shocks across different regional contexts. Indeed, there was departmental variation in pandemic exposure, demographic pressures, conflict history, and institutional response capacity, which likely produced different teacher workforce and school closure outcomes. Joining the CIES 2026 call to reexamine the divisions that affect education, this study inquires how teacher workforce trajectories varied across Colombian departments between 2017 and 2023. Specifically, I ask: How did COVID-19 affect the teacher workforce growth patterns across departments, and what structural characteristics explain variation in departmental resilience or vulnerability to these shocks?

I attempt to solve these questions by developing a hierarchical linear growth model (HLGM) (Singer & Willett, 2003) that uses comprehensive administrative data from the National Administrative Department of Statistics Formal Education database tracking teacher numbers and work conditions (DANE, 2023), complemented by the Colombian Institute for Educational Evaluation datasets on school performance and demographics (ICFES, 2023). To account for demographic changes and exposure to the pandemic, I include datasets from the DANE’s Vital Statistics and the Ministry of Health COVID-19 exposure. More importantly, because of the expected non-linear behavior of the teacher workforce growth pattern around 2020, this approach models department-level teacher counts using piecewise linear growth functions with a structural break in March 2020.

An HLGM approach is relevant in this case because this methodology allows me to overcome the nested structure of the data and helps to disentangle departmental time-trajectories. As it has been established in the HLGM literature (Heck et al., 2013; Hox et al.,2018), the model is set up with two levels: with time-varying pandemic exposure at Level 1, while Level 2 includes between-department variation using structural characteristics, including historical teacher preparation capacity (2017-2019 averages), baseline demographic composition, pre-pandemic school infrastructure from ICFES data, and geographic factors.

By documenting teacher workforce growth patterns in Colombia’s diverse departments during this critical seven-year period, this research provides empirical evidence for understanding how systemic educational shocks interact with existing regional inequalities, directly informing recovery policies and workforce allocation strategies essential for educational improvement.

Author

• Jonathan Beltran, Teachers College at Columbia University