| Many researchers have identified the myriad of concerns that frequently affect people who drop out of school prior to high school graduation. These include increased risks of lower income, need for welfare support, unemployment, and criminal activity (Alexander, Entwisle, & Horsey, 1997; Christenson & Thurlow, 2004; Gleason & Dynarski, 2002; Suh, Suh, & Houston, 2007). School leaders have a keen interest in helping all students successfully complete school, thereby reducing the risk of these issues occurring later in life.;In an effort to help students avoid these potential risks, school leaders have tried to identify students at risk of dropping out of school so they can intervene and help the students persist to graduation. Efforts to identify potential dropouts generally began at the high school level, but Bowers (2010) noted interventions used that late in a student's education are rarely effective. Suh et al. (2007) discussed how interventions are more effective when put into place early in students' educational careers, while Entwisle and Alexander (1993) argued that most students start elementary school with a clean slate before their pattern of performance tends to lead them toward eventual success or failure at completing high school.;This quantitative study examined a sample of 222 students who entered high school in the same cohort in an effort to identify predictors of high school dropouts. The study attempted to develop a practical model built from elementary school data that would predict the number of high school credits earned for each student, thereby giving school leaders a measure they could analyze to identify which students were at risk of eventually dropping out. Through this process, the study aimed to provide school leaders with a tool to identify potential dropouts during elementary school, giving them a better chance of applying interventions that were more likely to be successful than those applied later in students' educational careers.;This study also focused on data easily recovered from typical school records. The intention was to build a practical model that could easily be developed from school data and applied to all students in a grade level. The variables considered for this study were attendance, core grades, discipline events, standardized test scores, socioeconomic status, grade retention, and reading level.;The findings of the study were that the earlier the model was developed, the less accurately it predicted high school credits earned. This was consistent with prior research noting that earlier identification efforts tend to be less reliable than later efforts (Bowers, 2010). In addition, the models developed in this study were not very accurate for any of the grade levels considered, from fourth through eighth grades. One positive outcome of the prediction models was the relative success with which they predicted dropouts for students predicted to earn extremely low numbers of credits.;Further study can be conducted on more complete data sets to determine if these models could be improved. In addition, for any attempts to identify students at risk of dropping out of school, appropriate interventions must be developed if school leaders are to try to keep these students in school through high school graduation. This study has offered some insights to aid in this further research. |
