Background: Secondary school attendance is a critical factor for overall educational attainment and adolescent wellbeing. Since the outbreak of the Coronavirus-19 (COVID-19) pandemic, elevated absenteeism has been a pervasive issue across the world. Understanding patterns of attendance and the factors that may lead to students following different patterns provides insight useful for improving school engagement, student wellbeing and post-secondary outcomes. Purpose: This study aimed to identify latent patterns of secondary school attendance throughout a southern region of England (Hampshire and Southampton) in order to establish factors influencing the likelihood that students follow different attendance patterns across one full academic year using self-reported psychosocial and linked socio-demographic data. Participants: Participants were recruited as part of the #BeeWell Programme. The #BeeWell dataset (https://beewellprogramme.org/) is a rich data source containing information on a wide range of adolescent wellbeing domains (e.g., internalising symptoms, self-esteem, psychological wellbeing), drivers (e.g., social/school connection, bullying, physical health), and characteristics (e.g., school year group, ethnicity, gender/sexual identity). #BeeWell follows a hybrid population cohort study design comprising: a truncated longitudinal study in which participants are tracked with annual data points from ages 12-15 (i.e., from Year 8 to 9 to 10 of secondary school; Sample 1); and, a serial cross-sectional study comprising annual data points for participants aged 14-15 (i.e., those in Year 10 of secondary school at a given time point; Sample 2). The current study used the first annual wave of data from Sample 1 and Sample 2, collected from Year 8 and Year 10 students in Autumn 2023 (N=20,241). Attendance data for the 2023-2024 academic year were provided by Hampshire and Southampton local authorities and linked to the #BeeWell dataset. Overall attendance was measured across the Autumn (T1), Spring (T2), and Summer (T3) terms, quantified as the percentage of possible sessions attended (where one session equates to 0.5 school days) after removing those missed due to authorised or unauthorised absence. The analytical sample comprised all adolescents attending mainstream schools for whom linked attendance data were available (regardless of whether they also returned a #BeeWell survey to reduce response bias), provided the maximum number of possible sessions for a student was ±20 (i.e., two school weeks). This approach was enlisted to remove erroneous data and limit the possibility that special cases (e.g., dual-registered students) biased results. The final analytical sample comprised 28,298 students. Analysis: Latent Class Growth Modelling (LCGM) was conducted to identify latent school attendance patterns across the 2023-2024 academic year. Thereafter, #BeeWell survey data (collected Autumn 2023) were used to identify a broad range of individual, social, and school-based predictors of attendance. The predictors analysed included year group, special educational need (SEN) status, free school meal (FSM) eligibility, ethnicity, gender and sexuality, socioeconomic deprivation, school belonging, happiness with attainment, staff relationships, migration status, physical health, internalising symptoms, school pressure, peer pressure, bullying, substance abuse, and experiences of discrimination. Analyses were carried out using Mplus version 8.9 following a maximum likelihood threestep method whereby: (1) latent trajectory classes were enumerated without the inclusion of covariates; (2) latent classes were saved alongside a measure of classification error to be used in subsequent analysis; and (3) determinants of latent trajectory class membership were included as auxiliary variables in a secondary model to test associations with each of the latent classes. Clustering of responses by school was handled using a sandwich estimator. Effects were estimated in the presence of missing data using Full Information Maximum Likelihood Estimation (FIML). Findings: Using LCGM, a 4-class solution provided the best fit to the data (Entropy >.90). As expected, the large majority of students followed the Consistently High attendance pattern (n = 25,968; 91.8%). Additionally, Rising (n = 661; 2.3%), Declining (n = 1,087; 3.8%), and Persistently Low (n = 577; 2.1%) patterns emerged. Relative to Consistently High attenders, students' age, SEN status, FSM eligibility, internalising symptoms, bullying, gender/sexual minority status, socio-economic deprivation, substance use, and discrimination all increased the odds of following sub-optimal patterns. Conversely, a strong sense of school belonging, happiness with attainment, good relationships with staff, physical health and being of a minority ethnic background all reduced odds of following sub-optimal patterns. Conclusions: The study identifies distinct patterns of school attendance throughout one academic year, highlighting key predictors of multiple patterns. Evidence presented strengthens the literature on school engagement signposting significant predictors of poor attendance useful to educators, policymakers, and public health officials alike. Findings highlight systemic inequalities and myriad factors influencing non-attendance, underscoring the need for targeted intervention dedicated to improving school environments and mental health support, to achieve more equitable educational outcomes for all young people.