Browsing by Subject "Machine learning application"

Now showing 1 - 1 of 1

Metropolitan, urban, and rural regions: how regional differences affect elementary school students in Germany
(2025) Schwerter, Jakob; Bleher, Johannes; Doebler, Philipp; McElvany, Nele; Schwerter, Jakob; TU Dortmund University; Bleher, Johannes; University of Hohenheim; McElvany, Nele; TU Dortmund University
This study examined how regional differences affect elementary school students using the representative German Progress in International Reading Literacy Study (PIRLS) 2016 data (N = 3,959 fourth-grade students; M_{Age} = 10.34 years; 49% girls; 71% from a nonimmigrant background) by combining bootstrapping, multiple imputations, principal component analysis, and the least absolute shrinkage and selection operator (LASSO). Grouping regions into rural, (sub-)urban, and metropolitan, we found that students from rural and metropolitan areas are 10.9% and 15.1% more likely, respectively, to receive an academic track recommendation than their urban counterparts. Similarly, rural and metropolitan students are 0.2 to 0.3 standard deviations more likely to enjoy school and be interested in reading than their urban counterparts. Aside from students’ backgrounds and skills, many of the characteristics explaining this regional difference are structural, directly affected by policy decisions. Variables directly and indirectly influenced by policy help explain regional differences, but nonpolicy variables reduce regional differences in academic track recommendations the most.