Abstract:

Dropout is a complex phenomenon involving early school leaving, resulting in significant negative impacts at both the individual and community levels. These impacts include reduced job opportunities, a greater risk of unemployment, and poverty. It is a multifaceted problem, rooted in students' socioeconomic conditions, family cultural background, and individual characteristics.

This paper, coming from the PRIN 2022 Evidence 4 Preventing Early School Dropout (E4PED) project involving three Italian universities (Cagliari, Palermo, and Salerno), presents an action related to a specific work package (WP): the analysis of perceptions among final-year middle school students in Campania regarding the causes of school dropout. The convenience sample was drawn from schools that voluntarily participated in the survey.

The methodology employed involved a regression model, particularly suited to analysing questionnaire data and highly effective in an educational context. Participants were asked to express their level of agreement or disagreement regarding the possible causes of school dropout using a 5-point Likert scale. The relationship between ordinal responses and explanatory variables was estimated using an ordered probit model, adjusted for ordinal variables collected via questionnaires. Analyses showed that student opinions vary significantly based on sociodemographic and behavioural characteristics, including gender, age, school district, parents' educational qualifications, and motivation to attend school.

These results prompt us to reflect on how intervention strategies to reduce early school leaving should be differentiated and adapted to these characteristics to more effectively meet the students' expectations and needs, the primary stakeholders in the education system. The informative value of student perceptions is also highlighted, an element that should not be underestimated in guiding local policies and targeted programs, while taking into account the limitations associated with the non-probabilistic nature of the sample.

Keywords:

School dropout, student voices, regression analysis.