The article explores how working together in school can affect students' grades and well-being. The researchers used a method called propensity score weighting-regression to figure out if collaboration networks have a causal impact on students. They found that this method is helpful when other models have trouble with sparse networks. By comparing students who work together with those who don't, they were able to see how collaboration affects outcomes. This approach helps control for factors like how connected students are in the network.