Reading self-concept plays a significant role in academic achievement. Considering increasing numbers of English learners (ELs) in the United States, there is an urgent need to investigate self-perceptions of ELs in comparison to those of native English speakers (NES). We applied Elastic Net analysis (ENET), a machine learning approach, to PISA 2018 data to identify the proximal and distal predictors of EL and NES students' reading self-concept. Unlike in earlier work, the ENET in the current study was separately employed for ELs and NESs after splitting the dataset for those subgroups. Contributions of ENET-selected predictors of EL and NES students' reading self-concept will be investigated in the full paper by conducting three-level multilevel modeling analyses, separately for each student population.