Too many students leave school without even the essential skills (ACT, 2011), and many others are so drained by the experience they lack a desire to continue on to a post-secondary education. Academic engagement has emerged as a construct representing students' personal investment in school (Greenwood, Delquadri, & Hall, 1984), and may be a psychological variable which can be intervened on. However, interventions must occur as quickly as possible to maximize their efficiency (Heckman, 2007). Students' peer groups may be a particularly potent venue of intervention, however several options exist for how to go about measuring their social networks.;In this thesis, social networking data of the only middle school of a small town in the north-eastern United States is analyzed to determine the properties of two collection methods (self-reported networks and participant observations) and four network identification methods (probability scores, reciprocal nominations, factor-analyses, and rule-based). Analyses overwhelmingly supported participant observations as a more inclusive, less biased data collection method than self-reports. Meanwhile, hypothesis tests were somewhat mixed on the most inclusive, least biased network identification method, but after a consideration of the findings and the structural properties of each network, the probability score method was deemed the most useful network. Implications, future research, strengths, and limitations are discussed.
展开▼
