Introduction

There is a long history of armed conflict in Colombia. Since the late 1940s, there were confrontations between two political wings and tension between peasants and landowners. This period of approximately ten-year internal conflict is known as La Violencia. Despite the signing in 1958 of a coalition between the two political parties to alternate the government, violence persisted. The arisen of left-wing guerrillas and right-wing paramilitary groups brought a new armed conflict. In the early 1980s, criminal groups associated with drug production and distribution led to an increase in violence within cartels and terrorism acts to persuade the Colombian government. As a consequence of these confrontations, civilian victims increased considerably during this period (^{Ibáñez & Moya, 2010}; ^{Moya & Carter, 2019}).

Since then, the dynamics of the armed conflict in Colombia has marked by three events. In 2006, paramilitary groups started a process of disbanding and disarming. However, new armed factions began to fight for the control of some zones formerly under paramilitary control. In 2016, the largest and oldest guerrilla group in the hemisphere, the Revolutionary Armed Forces of Colombia (Farc), signed a peace agreement and started its demobilization (Internal Displacement Monitoring Centre [IDMC], 2020; ^{Moya & Carter, 2019}).

During many stages of the Colombian armed conflict, armed groups have been used violence against civilians as a deliberate strategy to gain control of disputed areas and spread terror among the population (^{Prada, 2016}). Indeed, more than 8.5 million civilians have been victims of the armed conflict since 1985, including 5.6 million internally displaced persons until 2019 (^{Gottwald, 2016}; ^{IDMC, 2020}).

Armed conflict can cause long-lasting damages. It has generated irreversible consequences on the population (^{Andrade et al., 2015}; ^{Buvinić et al., 2013}). There is scarce literature about the armed conflict impacts on household well-being (^{Kadir et al., 2019}). Part of this is due to the complex and multi-dimensional effects of the armed conflict, which becomes an obstacle to sustainable development and human security (^{United Nations Development Programme & World Health Organization, 2005}).

Civil war negatively affects educational outcomes (^{Bohada, 2010}; ^{Brück et al., 2019}). Boys and girls from households with Internally Displaced Persons (IDPs) have been the most vulnerable to the effects of armed conflict on educational achievement (^{Brown & Velásquez, 2017}; ^{Buvinić et al., 2013}). Children from displaced households had not academically performed as well as other students (^{Akresh & De Walque, 2008}; ^{Gimenez & Barrado, 2020}; ^{Kadir et al., 2019}).

The impact of forced displacement is evidenced in the quality of life of household members, generating a negative context for their educational processes (^{Avendaño-Villa et al., 2015}; ^{Gimenez & Barrado, 2020}). There are factors associated with forced displacement, such as household poverty, which generate direct consequences on the academic performance of students (^{Bernardo & Baranovich, 2014}; ^{Navarrete et al., 2017}). For example, ^{Ordóñez (2009)} found that the nutritional deficiencies of IDP’s households affected the school development of children.

Also, the discrimination that boys and girls have been suffering in school for being a member of an internally displaced household can be so pervasive as to limit their academic attainment (^{Bertoni et al., 2019}; ^{Mooney & French, 2005}). Boys and girls from IDP households are repeatedly absent from school, more than other children (^{Ibáñez, 2006}; ^{Munevar et al., 2019}). They often miss school because their labor is needed at home or to generate income for households’ economic survival (^{Ibáñez & Querubín, 2004}).

Forced displacement is a traumatic event for household members. Many victims of forced displacement have suffered double or triple displacements, and it was more prevalent for IDPs belonging to ethnic groups. It could affect their physical and psychological health, which also could affect children’s academic achievement (^{Mancini, 2020}). The effects on the victims’ mental health of the armed conflict can cause disruptions in different dimensions of the household' members, such as children’s education development (^{Kecmanovic, 2012}; ^{Manrique-Palacio et al., 2018}; ^{Rodriguez et al., 2006}).

The intersectionality of IDPs’ disadvantages is a crucial category to understand educational inequalities. Depending on the context, the magnitude of the disruption frequently varies between boys and girls (^{Buvinić et al., 2013}; ^{Hattar-Pollara, 2019}; ^{Sutanuka & Prakarsh, 2016}). Studies that analyze the relation between children’s education out-comes and the presence of an IDP in a household are scarce. Still, the literature is more scarce if gender is also considered. Boys and girls run a different risk of suffering different consequences in their educational outcomes for being part of a family forced displaced by armed conflict (^{Contreras et al., 2012}).

Why might grade retention be different for girls and boys from IDP’s households? Some literature has evidenced that boys could achieve lower outcomes than girls on various non-cognitive dimensions. Boys could have attention and behavioral difficulties. Boys’ levels of inhibitory control and perceptual sensitivity are lower than girls’. Boys are less able to delay gratification, and more likely to be diagnosed with attention deficit and hyperactivity disorder (^{Autor et al., 2019}; ^{Bertrand & Pan, 2013}; ^{Brenøe & Lundberg, 2018}; ^{Kadir et al., 2019}).

Besides, the behavior of boys is more responsive than girls to other indicators of disadvantage, such as forced displacement households, poor-quality schools, and less-educated neighborhood (^{Bertrand & Pan, 2013}; ^{Brenøe & Lundberg, 2018}; ^{Lei & Lundberg, 2020}). Within the possible explanations for this gender gaps, one could be that family circumstances and parental time dedication could affect more boys’ academic outcomes than girls’, or because low-socioeconomic status households invest relatively more in girls than in boys (^{Autor et al., 2019}; Autor et al., 2016).

Also, a family disadvantage, such as being part of an IDP household, can strengthen the negative effect that neighborhoods and schools have on boys’ academic outcomes. Male children may be more exposed to violence in recipient communities and educational institutions. In some scenarios, authority figures could be more severe with boys than with girls; perhaps boys affront greater disciplinary and criminal sanctions from teachers and police (^{Autor et al., 2019}, 2016).

What we seek to understand is whether belonging to a household victim of the Colombian armed conflict impacts boys and girls equally or if it exacerbates gender educational disparities in a society affected by it. Among the possible measures of educational outcomes, we considered school retention to analyze the case of Colombian boys and girls from IDP’s households.

This educational outcome could be another source of discrimination. Evidence demonstrates that grade repetition is not adequate for most students; in some cases, it could worsen the student’s subsequent academic achievement instead of improving it (^{Lorence, 2014}). Besides, repeaters’ self-esteem could be negatively affected because of their age difference with their same-grade peers (^{Kamal & Bener, 2009}). Older students have a higher probability of report health problems (^{Needham et al., 2004}).

This research analyzed the link between grade repetition and being a boy or a girl in a household with any victim of armed conflict. The empirical approach used 2015 Demographic and Health Surveys data from Colombia, as the country with the most significant number of IDPs worldwide: 7.4 million, 15 % of the Colombian population, 18 % of worldwide IDPs in 2019 (^{IDMC, 2020}; ^{Unidad para la Atención y Reparación Integral a las Víctimas, 2019}).

Specifically, our research question is: Did forced displacement affect the repetition rate of Colombian girls and boys? With this purpose based on the 2015 DHS data and through logistic regressions, we estimated the net effect of forced displacement on the probability of repeating a grade for boys and girls.

We extended prior research in two aspects. First, we studied the relation between grade retention and one of the principal consequences of armed conflict, forced displacement. Second, we estimated the net effect of being a boy or a girl from an IDP household on the probability of failing a grade.

Methodology

Variables

Our universe of the study was a sample of Colombian girls and boys of school-age (6 to 15 years old). According to Colombian law, a child begins formal education at the age of five, and the limit for compulsory education is at the age of 15 years old. Under normal conditions, a boy or a girl who is 15 years old should have completed primary school and should have completed four years of high school.

We measure the impact of belonging to a household with at least one IDP member in the school repetition rate, as one measure of academic achievement and the adaptability of boys and girls belonging to households victims of armed conflict. Our explanatory variable of interest is a dummy variable that classifies a household as IDP if at least one member has migrated due to the Colombian armed conflict. We also had included as controls some available variables at the individual, household, and community level: age, characteristics of the household head (sex, age, and educational level), household wealth index, household geographical region and municipal index of IDPs.

We used a logistic regression model to estimate the partial effect of forced displacement on the probability of repeating a grade for boys and girls. The estimated model is as follows:

In our notation, Y*_i denotes the dependent variable of repeating a grade. The variable Female is a dummy variable that is one when the child is a girl and zero when the child is a boy. The variable IDP _hh is a dummy variable that identifies the forced displacement of any member of the household. We also include the characteristics of boys and girls at individual (X _ind ), household (X _hh ), and context (X _cont ) levels. Because we considered that the effect of all the explanatory variables differs across boys and girls, we specify model (1) so the intercept and all slopes can be different depending on the sex of the child (see δ coefficients).

Ethical Issues

Our primary source of information came from the Demographic and Health Survey (DHS) Project. This program collects information in over 90 countries. DHS has been one of the most recognized public sources of representative data on themes such as child health and nutrition, family planning, among others.

The DHS World Surveys maintain strict standards to anonymize respondents’ information. Also, each interview was conducted only with informed consent. On the other hand, the Institutional Review Board has reviewed the procedures and questionnaires of each country as well as legislation of the nation in which it is applied (^{Ministerio de Salud y Protección Social & Profamilia, 2015}).

Table 1  Descriptive Statistics by Sex  

Note . Data sources: Colombian DHS (2015), RUV, Dane. Standard Deviation in parenthesis.

Results

Based on the methodological strategy previously described, we focus on grade repetition as a measure of educational achievement. Table 2 presents the results of the estimating model (1).

Table 2  Logistic Regression Results  

Note . Data sources: Colombia DHS 2015, RUV, Dane. Robust standard errors in parenthesis. *p<0.1, **p<0.05, ***p<0.01

We estimate the effect of forced displacement and gender condition on the probability of repeating a school year (Table 3). First, our empirical strategy identified whether there was a difference in the likelihood of repeating a school grade of children belonging to IDP households, as compared to families that did not have. At the same time, we considered other covariates in the characteristics of a typical boy and girl in the sample. There was an increment of 6 % in the expected probability of repeating a grade if a Colombian boy was a member of an IDP household. This increase due to the forced displacement condition was statistically significant at 10 %. Therefore, we reject the null hypothesis that being part of a forcibly displaced household did not affect the probability of repeating a grade.

After controlling for all other independent variables, there is at most a 10 % probability that the effect of being forcibly displaced on the expected likelihood of repeating a grade was the result of chance. However, in the case of girls from an IDP household, this effect was not statistically significant. The probability of repeating a grade increased 0.5 percentage points, but it could be due to chance; there was no statistical evidence that the IDP condition affected girls’ school repetition.

Table 3 Selected Marginal Effects on Grade Repetition Probability 

Note . Data sources: Colombia DHS 2015, RUV, Dane. Robust standard errors in parenthesis. *p<0.1, **p<0.05, ***p<0.01

Second, the grade repetition probability difference between boys and girls students is negative in both scenarios of the presence and absence of IDPs in the household. The more considerable gap of gender is for non-IDPs; it is about 8.7 % less likely to repeat a grade for a girl than for a boy. That is a sizable as well as a significant difference. Among the households with victims of the armed conflict, girls had 3.2 percentage points less probability of failing a grade than boys in the same situation. Reviewing, although Colombian girls were less likely to repeat a grade, girls’ advantage was lower when they were from IDP households.

Discussion

Regarding the answer to our research question, we estimate that the probability of repeating a school grade of boys belonging to internally displaced households did increase by 6 %. In the case of girls belonging to internally displaced families, the probability of repeating a grade increased the retention rate by 0.5 %. However, there was no statistical evidence that this increase significantly affected IDP girls’ school repetition.

Colombia has the most significant number of boys and girls belonging to households with IDPs worldwide. Children from IDPs’ homes were 2 million until 2017 (Save the Children, 2018). The proportion of forced displaced youth who were still in primary school was almost double that of the non-displaced. The above indicates a delay in admission and more repetition among children in internally displaced households (^{Ferris & Winthrop, 2010}). Our findings are consistent with the literature, which reported that girls had better chances of completing school grades than boys in a family disadvantage circumstances as IDPs households (^{Autor et al., 2019}; Autor et al., 2016; ^{Bertrand & Pan, 2013}; ^{Brenøe & Lundberg, 2018}; ^{Lei & Lundberg, 2020}; ^{Schneider, 2020}).

However, forced displacement due to violence also generated a negative impact on the educational trajectory of Colombian girls, since the advantage of girls concerning boys decrease in 5.5 percentage points when we compare IDP households with non-IDP households. Colombian armed conflict has been reinforcing the inequalities that trap countries in cycles of violence.

Also, it is relevant to review the budget of the education sector. According to data of World Development Indicators, during 2018, the percentage of Colombian GDP dedicated to education was 4.1 %, while the proportion of military expenditure was 3.2 %. An indirect effect of Colombian armed conflict has been prioritizing, in some sense, the military instead of education spending. The government and donors should prioritize the development of an inclusive education system. It is relevant because education acts as a force for peace and gender equity (United Nations Educational Scientific and Cultural Organization [Unesco], 2011).

Some interventions could be useful in rebuilding education achievement for children from households with victims of armed conflict context. It is necessary to ensure that all critical national education system’s mechanisms consider policies specifically for boys and girls from IDPs’ households and recognize how gender affects their educational trajectory.

Mainly, the key is to implement a gender-responsive education plan for armed conflict victims. Public funds allocation could play a crucial part in narrowing learning gaps. Funding is relevant, among others, to implement a specific training and recruiting process for teachers in the municipalities with a higher prevalence of IDPs. It is necessary to attract qualified teachers and give them adequate support for promoting gender equity and inclusive education for boys and girls from households with victims of armed conflict. It is also relevant to avoid teacher and student absenteeism through early warnings to protect them in contexts with a high prevalence of armed conflict (^{Rodríguez & Sánchez, 2012}).

In these contexts, public education should offer additional support for children from households with victims of armed conflict. Among others, it is necessary to introduce mandatory gender training for teachers that have the most contact with boys and girls from IDPs’ households. Remedial education programs can also help boys and girls with learning difficulties related to disadvantage families, as the victims of armed conflict.

Also, it is required to introduce mandatory support of student affairs centers in the schools for students that provide advice, information, and support services for boys and girls with lower grades, preventing early grade retention. This support could be formal or informal in addressing specific student needs by household’s migration status and gender.

Formal programs could be interventions adhering to established school rules, looking to promote harmony and address students’ needs. The informal ways are other types of actions beyond the programmatic approaches established by the education system to calibrate the peculiarities of the students from IDPs households, such as the mediation of disputes among students because their family was armed conflict victim, and more frequent among boys.

The balancing of these efforts would encourage the resilience of these victims of war who generally are also involved in an economic vulnerability context. In effect, a proper follow-up could maintain students’ possibilities of benefit from the education system. In this sense, our results highlight the vital role of equity in education and gender in contexts with a high intensity of armed conflict such as Colombia.

This study contributes to the growing body of empirical research that analyzes the incidence of armed conflict on children’s academic achievement. This paper studied the effect on grade repetition of being a Colombian boy or girl of an IDP household. The purpose of this research refers to propose educational and governmental policy tools to prioritize funding, resources, and efforts to reduce these inequalities. This study is relevant, given the significant influence of armed conflicts on the civilian population, the importance of education for the well-being of boys and girls, and the possible poverty traps among the victims. Also, we analyze the Colombian case because it is the country with the most significant number of children IDPs in the world.

We found evidence that the children from households victims from Colombian armed conflict had more likelihood of grade repetition, especially boys. As shown in our results section, children from homes with IDP members increased the chance in grade retention of boys and girls, but the increase in probability was higher and significant for boys from IDP households. In consequence, the effect of being a child belonging to a family with any member victim from armed conflict on failure a grade becomes a new source of inequality of educational opportunities, especially for boys. When compared to children of the same age, an IDP has a lower accumulation of human capital (^{Ferris & Winthrop, 2010}).

We present some of the limitations of our study. First, the data is limited to DHS surveys that do not guarantee the representativeness of all forcibly displaced children in Colombia. Second, we define our academic achievement measure as school repetition, leaving aside other indicators of educational attainment such as attendance or results in standardized tests. Third, our sample only considers boys and girls who continued in the education system, leaving aside the children who, due to the armed conflict, could not continue studying.

Our findings support that the Colombian government must guarantee educational policies differentiated by the condition of forced displacement and by sex to ensure adequate school performance. In any case, the State must guarantee the reduction of the armed conflict, which has a significant impact on children’s academic performance, as well as compromising their future well-being. Governments bear the primary responsibility to guarantee a child’s right to education; it is necessary to respect and defend it even more in adverse situations.

In the Colombian case, the government has been unable to fulfill that social responsibility, so other actors, such as the international community, must be more involved. It is essential that Unesco and Save the Children monitor the effectiveness of policies and budgets aimed at reducing educational and gender gaps, especially in armed conflict contexts such as the Colombian. Equity education and equity gender in armed conflict contexts must be treated as a distinct and priority concern in all monitoring and reporting activities by UN agencies.

In doing so, researchers must consider analyzing the possible transmission channels to understand the link between gender and armed conflict on educational gaps. Besides, it is needed more research to locate ethnographic gender analyzes within the broader framework of debates shaping the academic understanding of education and gender gaps in diverse armed conflict contexts.

It is relevant to analyze if the educational system has been exacerbated these inequalities in boys and girls that have been exposed to Colombian armed conflict. Future researches could carry out case studies comparing the performance of boys and girls victims of the armed conflict in different types of educational institutions or between IDPs and international refugees. It is also essential to document the existence and effectiveness of educational policy initiatives that focus on the population victim of the armed conflict with a gender vision.