Investigators reported an overall primary school dropout prevalence of 1.68% in the abstract of this analysis of UNICEF Multiple Indicator Cluster Survey (MICS) data. The survey-based estimate comes from children in the primary age range. The Methods section of the source also lists a slightly different sample count (14,305 children aged 6–12 years), while the abstract reports a total sample of 14,278; the source text contains both figures and does not reconcile the discrepancy. The authors used bivariate comparisons and multivariable log-binomial regression to derive risk estimates for demographic groups.
Male sex emerged as a clear correlate in adjusted models. The authors reported an adjusted risk ratio (ARR) of 2.44 (95% CI: 1.63–3.64) for boys compared with girls. Short sentence: boys were at higher risk. The adjusted models thus indicated sex-specific vulnerability within the primary cohort.
Reading exposure in the household was the single strongest association reported. Data indicated that children who did not read books and were not read to at home had an ARR of 14.47 (95% CI: 9.30–22.51) for primary school dropout compared with those with home reading activity. That effect size is large. The authors framed household reading and book access as central elements of the home learning environment, and modeled these variables alongside demographic and regional covariates.
The survey instruments from MICS capture caregiver-reported household stimulation activities; investigators treated absence of reading activities as a marker of limited home learning stimulation. The magnitude reported suggests that household literacy practices may act as a proximal factor linked with school continuity in this context.
Regional differences were apparent in the adjusted analyses. The authors used Barishal as a reference division and reported that Mymensingh had higher dropout risk (ARR = 3.82, 95% CI: 1.59–9.18). The source presented division-level contrasts across Bangladesh’s eight administrative divisions, with Mymensingh singled out for the highest adjusted rate relative to Barishal.
Age also correlated with dropout status. The study team observed that children aged 9–12 years at the beginning of the school year had an increased risk of dropout (ARR = 1.83, 95% CI: 1.15–2.93) compared with younger children in the 6–8 year band. That pattern suggests dropout risk rises as children move through primary grades, a finding the authors linked to grade repetition, late entry, and household role demands in the broader literature cited.
The investigators analyzed the 2019 UNICEF Multiple Indicator Cluster Survey (MICS), a nationally representative, two-stage cluster sample with standardized questionnaires covering child development, education, parental involvement, and household characteristics. The source states high household-level response rates and explains that primary sampling units were selected with probability proportional to size. The Methods section specifies that 31,196 children admitted to school last year were considered, with subsequent inclusion criteria narrowing the analytic sample to children aged 6–12 years; the abstract, however, reports a final analytic sample of 14,278 while the Methods cites 14,305.
Analytic methods included bivariate testing and crude and adjusted log-binomial regression models to estimate risk ratios for dropout across covariates. The source text did not provide every modeling parameter or step-by-step variable-selection strategy in the abridged content provided here. The data are publicly available from the UNICEF MICS repository, as the authors note.
The authors proposed several remedial strategies tied to their findings. They recommended strengthening early childhood education programs and promoting home-based literacy activities to increase retention. They also recommended targeted interventions for older primary-age children who face higher dropout risk, and policy attention to regional disparities, with a focus on divisions showing elevated adjusted risks.
Short, pragmatic actions were suggested: low-cost learning materials to increase book access; parent-education initiatives to encourage reading and caregiver involvement; and community-led tutoring support for children at risk of leaving school. The authors framed these interventions as household-focused complements to school-level policies and as ways to align monitoring with Sustainable Development Goal 4 on inclusive, equitable quality education.
The authors included a strengths and limitations section in the full article; the abridged source text available here did not present a full list of methodological limitations or sensitivity analyses. The Methods describe cross-sectional analysis of survey data, which limits causal inference. The source did not report in this extract whether the models adjusted for all potential confounders such as parental education, household wealth, or child labor status in full detail, though the Discussion and referenced literature address socioeconomic and contextual drivers of dropout.
Policy makers and program designers may want further detail on temporal sequencing (for example, whether weak home reading preceded dropout) and on subgroup analyses by wealth, parental education, and school factors; such specifics were not included in the shortened content provided. The article does state data availability and declares no specific funding or competing interests.
The paper ends with concrete recommendations for interventions and monitoring but leaves several empirical questions open in the source excerpt: the reconciliation of discrepant sample counts, the full covariate-adjustment list used in final models, and whether region-specific program trials could reduce the elevated dropout risks documented in divisions such as Mymensingh.