Since the early eighteenth century, the configuration of family structures in Europe has changed significantly. Initially, socioeconomic transformations have led to the conjugal family becoming the dominant type (Durkheim 1921). Then, alongside progressive secularization, the institution of matrimony lost its centrality as a necessary step for forming a new family. As a result of this shift, the term “nuclear family” has been used to include different types of families characterized by their small size and a composition consisting of only parents and unmarried children. However, compactness does not translate into stability: deaths, separations, and divorces can disrupt the tranquility of the family unit.
These disruptive events can fall under the term “dissolution” and have medium- or long-term effects since each of them causes a loss of economic, social, and psychological resources for the children involved (Amato & Keith 1991a; Amato & Keith 1991b; Amato 2000; Bernardi & Radl 2014; Cantalini et al. 2022).

The research questions investigate the impact that the dissolution of the family of origin has on children’s educational possibilities. 
The strong connection between parents’ and children's academic attainment (Blau & Duncan 1967) necessitates the examination of the intergenerational transmission of educational inequalities and its role within the studied process.

Taking into account all these elements and the existing international literature and different national studies (Amato & Keith 1991a; Amato & Keith 1991b; Bernardi & Radl 2014; Cantalini et al. 2022; Jonsson & Gähler 1997; Fischer 2007; Steele et al. 2009; Zhelenkova & Panichella 2023), this study aims to develop a general statistical model that can be implemented within different European contexts with minor adjustments.
Two hypotheses are tested. The first hypothesis maintains that a “dissolution effect” exists in relation to children’s educational attainment. The second examines the different impacts of the “dissolution effect" based on parents' educational background. 

Based on the aforementioned literature review, the dimensions selected for analysis are the educational attainment of the respondents, the status of the family of origin, gender, place of birth, year of birth, and parents' educational attainment.
Data analysis is conducted using binomial logistic regression models, with parameters interpreted by implementing post-estimation tools.
Models have been tested in the Italian context. The selected dataset, Ita.Li - Italian Lives, is a longitudinal dataset containing 8,778 cases from 4,900 nuclear families. 

Bearing in mind the Italian school system, the completion of the secondary school education cycle has been chosen as the threshold value. This level of education grants young adults, after a five-year course, the diploma necessary to access university, allowing them to potentially choose to further their education. Different nations may require slight changes to this variable based on the structure of the specific educational system, but the binary nature of the variable facilitates easier adaptation.

Results reveal a negative association between the dissolution of the family of origin and the probability of completing at least secondary school education. Data suggest that the “dissolution effect” exists and consists of a disadvantage in the educational trajectories of the respondents.
When it comes to the differences based on parents’ educational background: those born into educationally privileged conditions appear to be the most affected. This effect may be explained by the fact that individuals from more advantaged backgrounds potentially have more privilege to lose.

Statistical models presented in this study were built with the intention of making them adaptable to different national contexts. The selected dependent variable is educational attainment, coded as a binary variable that indicates whether the respondent has completed the secondary school education cycle. The independent variable is family dissolution before the age of 14. According to the literature, variables were selected to represent different characteristics of individuals, their families, and their living context. Based on these criteria, the selected confounding variables are sex, place of birth, year of birth, and parents' educational attainment (obtained by synthesizing both parents' academic degrees). Each confounding variable selected meets the statistical criteria of being antecedent to the dependent and the independent variables. Bivariate analysis was implemented to verify the relevance of confounders. Considering the binary nature of the dependent variable, binomial logistic regression models were applied. Since they allow partially relevant variables to be retained to avoid distortions, it was decided to keep variables such as sex and place of birth, which correlate with educational attainment but not with family dissolution. The dataset of Ita.Li – Italian Life was deemed appropriate to test the models because of several characteristics: it is a detailed longitudinal dataset with a substantial number of cases (8,778 interviewees nested in 4,900 families), it employs a four-stage probabilistic sampling design implemented to ensure a nation-representative sample and to minimize the sampling error (Pisati 2023, Lucchini et al. 2023) and its contains recent data (2019-2021) covering a wide and balanced range of birth years (1921-2000). Two models were developed. Model 2 includes an interaction between the independent variable (family dissolution) and the variable representing parents’ academic attainment. This interaction allows the test of the second hypothesis. Since logistic regression models produce logistic scale coefficients, post-estimation tools were used to obtain values on the probabilistic scale, which are easier to interpret: Average Marginal Effect (AME), Marginal Effects at Representative Values (MERs), and predicted probabilities. Tested models resulted statistically significant (Model 1 with a confidence interval of 95% and Model 2, the interaction model, with a confidence interval of 68%). Post-estimation tools also showed statistical significance with a 95% confidence interval. The analysis was conducted using the software STATA.

Both models confirm the first hypothesis: the “dissolution effect” on children’s educational attainment exists and it has a negative impact. The estimated AME values remain consistent even when the interaction is considered. On average, children who have experienced the dissolution of the family of origin have a probability of obtaining a secondary school diploma approximately 9 percentage points (p.p.) lower than the ones who did not (Model 1: -9.18*** p.p.; Model 2: -8.69*** p.p.). Regarding the second hypothesis, MERs derived from Model 2 confirm the existence of variations in the impact of the “dissolution effect” based on parents’ educational background. Data seem to suggest a lower impact on children from families with a low educational background (-7.66*** p.p.) compared to those from medium and high educational backgrounds (-13.02*** p.p. and -12.2*** p.p., respectively). To formulate a possible explanation for this counterintuitive finding, predicted probabilities were estimated. Even though children from medium and high educational backgrounds are more impacted by dissolution, they had and continue to have a much higher predicted probability of obtaining a diploma. Based on Model 2 children from intact families have a probability to graduate of 38.1% if their parents have a low education level, 68.38% if parents have a medium education level, and 92.46% if parents have a high education level. Looking at these data, we speculate that the greater negative impact of dissolution on advantaged children stems from a partial loss of privilege. Children already in disadvantaged situations suffer a negative effect from dissolution too, but the effect is limited by the already low probability of completing secondary school education. Testing the developed models with data from other European countries could help corroborate whether this dynamic is consistent across nations with similar and different welfare regimes and school systems. ***p-value < 0.01

Amato, P. R. (2000). The consequences of divorce for adults and children. Journal of Marriage and Family, 62(4), 1269–1287. Amato, P. R., & Keith, B. (1991a). Parental Divorce and Adult Well-Being: A Meta-Analysis. Journal of Marriage and Family, 53(1), 43–58. Amato, P. R., & Keith, B. (1991b). Parental divorce and the well-being of children: a meta-analysis. Psychological Bulletin, 110(1), 26–46. Bernardi, F., & Radl, J. (2014). The long-term consequences of parental divorce for children’s educational attainment. Demographic Research, 30, 1653–1680. Blau, P. M., & Duncan, O.T. (1967). The American Occupational Structure, New York, Wiley. Cantalini, S., Panichella, N., Guetto, R. & Ballarino G. (2022). Divorzio e stratificazione sociale in Italia. Il ruolo della separazione dei genitori sugli esiti educativi e occupazionali dei figli. Rassegna Italiana di Sociologia, 1, 119-148. Crenshaw, K. (1989). Demarginalizing the Intersection of Race and Sex: Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics. University of Chicago Legal Forum, 1989, 139-168. Durkheim, É. (1921). La Famille conjugale. Revue Philosophique de La France et de l’Étranger, 91, 1-14. Fischer, T. (2007). Parental Divorce and Children’s Socio-economic Success: Conditional Effects of Parental Resources Prior to Divorce, and Gender of the Child. Sociology, 41(3), 475-495. Guetto, R., & Panichella, N. (2019). Family arrangements and children’s educational outcomes Heterogeneous penalties in upper-secondary school. Demographic Research, 40, 1015–1046. ISTAT (2023a). Annuario Statistico Italiano 2023. Roma: Istituto Nazionale di Statistica. ISTAT (2023b). Matrimoni, unioni civili, separazioni e divorzi - anno 2022. Roma: Istituto Nazionale di Statistica. Jonsson, J. O., & Gähler, M. (1997). Family dissolution, family reconstitution, and children's educational careers: recent evidence for sweden. Demography, 34(2), 277–93. Lucchini, M., Argentin, G., Bussi, D., Consolazio, D., De Santis, G., Gerosa, T., Guidi, G., Negrelli, S., Piazzoni, C., Pisati, M., Respi, C., Riva, E., Sala, E., Scisci, D. & Terraneo, M. (2023) Quality Profile: Questionnaires, Fieldwork, and Data Preparation. Milano: Institute for Advanced Study of Social Change. Pisati, M. (2023). The Italian Lives Survey: Sample Design, Weighting, Variance Estimation, and Data Analysis. Milano: Institute for Advanced Study of Social Change. Steele, F., Sigle-Rushton, W., & Kravdal, Ø. (2009). Consequences of family disruption on children's educational outcomes in norway. Demography, 46(3), 553-574. Zhelenkova, A., & Panichella, N. (2023). Family socioeconomic status and sibling correlations in upper secondary education. An empirical analysis of educational inequalities in Italy. The British Journal of Sociology, 74(5), 808–816.