Externalizing problems among Kosovar adolescents: pubertal correlates in girls and boys

ABSTRACT

Pubertal timing and pubertal developmental status or stage have been associated with the rates of externalizing problems among youth, particularly in vulnerable developmental contexts including cultures undergoing rapid transformations. The present study tested the extent to which pubertal timing and pubertal developmental status were associated with delinquent and aggressive behaviours, in girls and boys. It also tested whether sex moderated these relationships. Data were collected from N = 1,342 Kosovar adolescents (665 girls; M age = 13.26 years, SD = 1.27; 677 boys M age = 13.19 years, SD = 1.31). No associations were found for the relationship between pubertal timing and externalizing problems in boys and girls. The findings provided support for a positive association between pubertal developmental status/stage and both delinquent and aggressive behaviours, in both girls and boys. The study highlights that pubertal timing and pubertal status/stage have important yet independent effects on externalizing problems.

KEYWORDS:

• Aggression
• rule breaking
• Body mass index
• pubertal development
• sex

The Kosovar developmental context and puberty

The Kosovar Society has been going through substantial social and economic transitions over the past two decades, particularly since the last war in 1998/1999. Changes are evident in the family structure, economic development as well the cultural shifts from collectivism and individualism (Shahini et al., 2015), with patriarchal societal norms about gender expectations still present nowadays (Balkans Policy Research Group, 2019; Kelmendi & Hamby, 2022). The war left Kosovo in a devastated state with a floundering economy, missing job opportunities, and still remains one of the poorest countries in Europe (Mehmeti, 2013). A recent report from Kosovo agency of statistics (KAS) (2020) indicated that unemployment rates among population are 25.0%, particularly among the young people These higher rates of unemployment adversely impact mental health and well-being of youth. In the post war years (1999 and onwards), families continued migrating from rural areas to the big cities, for a better job opportunity; these changes are also included a shift from traditional extended families (parents, grandparents, aunties) to nuclear ones (parents and children; Shahini & Landsman, 2008). Changes in family composition also constitute an additional stressor for adolescents, thus increasing their vulnerability.

Studies in the Kosovar context have shown that externalizing problems are more prevalent among boys than girls (Kraja & Ahmeti, 2015; Shahini et al., 2015), however, little is known about Kosovar adolescent bodily development. The associations between pubertal development and externalizing problems can be better understood within the context that the observed behaviours occur, such social conditions, for instance. Low SES has been shown to be associated with increases in overweight categories, in comparison with adolescents coming from average or higher high SES backgrounds (Sherwood et al., 2009), earlier sexual development (Oelkers et al., 2021), as well earlier pubertal timing in both sexes (Sun et al., 2017). On the other hand, socio-demographic characteristics, such poverty in neighbourhoods or low SES, can be associated with aggressive behaviours among adolescents (Bozzini et al., 2021; Van Oort et al., 2011).

Puberty begins with the development of secondary sexual characteristics, which are unique to each sex. This also includes changes in height, skin, and growth of pubic hair. In girls, developmental changes also include breast development and the onset of menarche. In boys, the voice continues to deepen, and growth of facial hair begins (Petersen et al., 1988; Susman & Dorn, 2009). Although these changes are universal, the particular timing at which they occur varies across individuals; for some, these changes develop earlier or later in comparison to peers. Two key concepts of pubertal development have been found to be associated with psychosocial adjustment outcomes among adolescents, namely pubertal status or stage (namely maturational changes with no particular reference to age), and pubertal timing (maturational changes relative to same age peers; Ge et al., 2006). Both changes and advances in pubertal status/stage and in pubertal timing have been associated with externalizing problems in both boys and girls (Dimler & Natsuaki, 2021; Ge et al., 2006; Ullsperger, 2020).

Very limited research has been carried out on Kosovar youth in general; this also includes work on pubertal development and associated bodily changes as well as the relationship between pubertal development and measures of psychosocial adjustment. This is an important gap because of the unique Kosovar socio-cultural and economic developmental context which provides a unique opportunity to gain a better understanding of the extent to which idiosyncratic contextual factors might impact both pubertal development as well as associated measures of adjustment and developmental outcomes.

The present study sought to test the extent to which measures of pubertal development were associated with externalizing problems among Kosovar youth (both boys and girls) and the extent to which the links between pubertal development and externalizing problems were moderated by sex. In addition, the present study tested whether socioeconomic status was associated with pubertal development status/stage and externalizing problems as well as pubertal timing and externalizing problems.

Body mass index, pubertal development, and externalizing problems

Research provides evidence that BMI is positively associated with earlier onset and timing of sexual maturation among girls (Biro et al., 2018; Juul et al., 2006; Lee et al., 2007; Rosenfield et al., 2009) and boys (Marceau et al., 2011). Comparatively, lower levels of BMI have been found to be associated with less advanced sexual development and delayed pubertal timing in boys (Oehme et al., 2021). In addition, Oelkers et al. (2021) found that lower SES was associated with earlier thelarche (breast development) in overweight/obese girls, providing evidence of how low SES is associated with indices of pubertal development. On the other hand, higher BMI (being overweight) has been found to be associated with externalizing problems in both boys and girls (Drosopoulou et al., 2021; Ter Bogt et al., 2006), and with higher physical aggression in boys (Tso et al., 2018).

Pubertal development and externalizing problems links

Pubertal status/stage has been linked to rates of externalizing behaviours in both boys and girls (Ge et al., 2002, 2006; Najman et al., 2009). In particular, Ge et al. (2006) tested the association between pubertal stage, based on the PDS, and externalizing problems in a community sample of 867 youth, ages 10 to 12 years. They found a positive relation between pubertal stage and externalizing problems, in both boys and girls. A large cross-sectional study on Dutch youth also revealed that Tanner Stages were positively associated with externalizing problems in both girls and boys, such as rule-breaking behaviours (Oldehinkel et al., 2011). Conversely, Nottelmann et al. (1987) found that pubertal development stage was negatively associated with aggressive and cruel behaviours in girls. In boys, a combination of lower pubertal stage and greater age (late maturation) was associated with adjustment problems, including lower self-image, sadness and anxiety, poor social and family relationships, but not externalizing behaviours. The study also provided evidence that steroid levels were negatively associated with behaviour problems in boys, but not in girls. Thus, adolescents who experience pubertal development later tend to be more prone of adjustment problems, particularly boys. In contrast, other work has found no significant associations between pubertal stage/status and externalizing problems (e.g. Kanwar, 2020).

Similarly, early pubertal timing has been shown to have a consistent association with externalizing problems (Dimler & Natsuaki, 2015, 2021; Graber et al., 1997; Kanwar, 2020; Kowalski et al., 2021; Marceau et al., 2011; Ullsperger & Nikolas, 2017), including aggression (Lynne et al., 2007; Ullsperger, 2020) and rule-breaking behaviours (Susman et al., 2007; Ullsperger, 2020), with mixed evidence about sex differences (Cui et al., 2012; Ge et al., 2006). Susman et al. (2007) found that pubertal timing in boys was associated with rule breaking behaviours; for girls, it was only associated with relational aggression. However, other work has not provided support for these associations with externalizing problems in boys (e.g. White et al., 2013) or with aggression in girls (Graber et al., 2006). Previous research has also provided evidence that high-stress environments, such as neighbourhoods characterized by high rates of poverty, for instance, might exacerbate the link between early puberty and externalizing problems (Ge et al., 2002). Based on the somewhat mixed findings from previous scholarship, more research is needed to better understand the extent to which the unique Kosovar developmental context might impact the link between measures of puberty and externalizing problems. Thus, it was hypothesized that (1) early pubertal timing would be positively associated with externalizing problems, namely delinquent behaviours and aggressive behaviours, in comparison to on-time or late maturing youth; it was also expected that this link would be stronger in boys in comparison to girls (moderation effect), (2) advanced pubertal status/stage would be positively associated with externalizing problems, namely both delinquent and aggressive behaviours. It was also expected that the observed relationship would be stronger in boys in comparison to girls. It is important to understand whether externalizing problems are associated with pubertal status/stage, but not timing, whether they are associated with pubertal timing, but not pubertal status/stage. The former would indicate that advancing through puberty places youth at risk for developing externalizing problems, while the latter would indicate that youth reaching a more advanced pubertal stage at an earlier age might be at greater risk for developing externalizing problems.

Theoretical framework

The developmental readiness hypothesis holds that early maturational timing among adolescents can increase the risk for behavioural problems in adolescents, particularly among girls (Negriff & Susman, 2011; Petersen & Crockett, 1985). In addition, accelerated sexual development, namely pubertal development, strengthens the asynchrony between physical and cognitive development, as well the asynchrony between chronological age and the current status of physical development which in turn makes youth more vulnerable to engage in externalizing behaviours. The gap between the physical and cognitive development due to cognitive and physical disparities, early maturing youth have limited knowledge and support, on ways which facilitate navigating those earlier changes than their peer counterpart, which results in detrimental developmental outcomes. In addition, the puberty and externalizing problem links, might be exacerbated or diminished depending on the environmental conditions or circumstances which includes the family, the neighbourhood, as well as the larger cultural context; this is sometimes referred to as the contextual amplification hypothesis (see Ge & Natsuaki, 2009, for a comprehensive review).

Methods

Procedure

Data were collected from N = 1,342 early adolescents (665 girls; M_age = 13.26 years, SD = 1.27; 677 boys M_age = 13.19 years, SD = 1.31) across seven municipalities in Kosovo. The study was reviewed and approved by the Ministry of Higher Education; school principals from selected schools in the seven largest municipalities across the country (Pejë, Prizren, Gjakovë, Fushë-Kosovë, Gjilan, Malishevë, and Mitrovicë), reviewed the study and independently decided participation. At participating schools, parents were sent a consent document, informing them of the study and requesting their consent for their child to participate in the study. Consent forms were returned to the schools. Of approximately 2,000 consent documents, 1,478 were returned. The remainder were missing either due to parental refusal to participate in the study, or due to failure to return the consent forms to school. Thus, the total recruited sample included 1,478 early adolescents: due to missing data, the final study sample was reduced to N = 1,342 adolescents. Data were collected in school classrooms of each municipality, between November 2019 and March 2020. Participants also completed an assent document prior to participation. The survey consisted of an anonymous self-report paper and pencil instrument which was completed during school hours, approximately 40 minutes. The study followed all ethical guidelines of Helsinki declaration.

Measures

Demographics variables

Descriptive statistics (frequency and mean age) were used for reporting age, sex, and socioeconomic status (SES).

Participants were 11 years old (13.7%), 12 years 14.2%), 13 years (25.4%), 14 years (32.0%), and 15+ years old (14.8%). The sample included 6^th (13.7%), 7^th 14.2%), 8^th (25.4%), and 9^th grade students (46.7%) who were mostly of ethnic Albanian (95.6%) ethnicity, approximately equally divided by sex, namely 665 girls and 677 boys.

Ses

Parental employment was used to assess family socioeconomic status (SES), based both on mother’s and a father’s employment, ranging from 1 =” owner/professional official/high degree”, 2 =” small business owner/professional/IT/large farm owner/military officer’’, 3 =” semi-professional worker/skilled craftsman’’, 4 =”cleric staff sales representatives/artist/other military personnel’’, 5 =”machine operator/semi-skilled worker such cook, waiter or janitor’’, and 6 =” labourer or service worker” and 7 = ‘unemployed’. Responses were reversed coded and then averaged, so that higher scores indicated higher family SES.

Pubertal Developmental Scale (PDS)

Puberty was assessed by using Pubertal Developmental Scale (Petersen et al., 1988). The self-report scale consists of 5 items, which focus on the development of secondary sexual characteristics. The first three items ask about body growth in height, pubic hair and skin changes, which are for both sexes. Then, it is followed by the items for facial hair, deepening of the voice for boys only, and breast development and menarche for girls only. Except menarche question (a dichotomy), the five pubertal items use 4 rating points Likert scale starting from 1 there is no development, 2 developments have barely begun, 3 development was definitely underway, and 4 development was already completed. The girls were also asked to indicate whether they experienced menarche or not. Those who experienced menarche reported it in months and in years. The measure was internally consistent (girls α = 0.67; boys α = 0.74), indicating acceptable reliability. PDS scores were computed by summing across the five items to obtain a total score; the sum of the scores on the five indicators was divided by five in order to preserve the original (1–4) metric.

Body Mass Index (BMI)

The Body Mass Index was calculated using weight in kg and height in m reported by participants: BMI = weight (kg)/[height (m)]².

Youth Self Report (YSR)

The Youth Self Report (Achenbach & Rescorla, 2007) included 112 items. The present study focused on two subscales, namely aggressive behaviours (17 items, consisting in items such ‘I physically attack people’), and delinquent or rule-breaking behaviours (15 items, consisting in items such ‘I break rules at home, school or elsewhere’), both part of externalizing problems scales. The Albanian version was used with permission from official ASEBA’s package representatives for Kosovo (see Shahini et al., 2015, for its use in Kosovar youth). Participants rated the items on a 3-point Likert type scale, selecting from (0) not true at all (1), somewhat true, and very true (2). In the current study, items part of the two measures were averaged to compute the two scale scores, ranging from 0 to 3. Raw items mean scores have regularly been used for some time in work which focuses on behavioural constructs rather than diagnostic criteria or thresholds (e.g. Phares & Compas, 1990). In addition, whether using a sum or mean score will not affect associations among variables, whether based on correlations or regression analyses, as rank ordering is identical.

The two scales were internally consistent as was the larger externalizing broad-band scale, both for girls (aggressive behaviours α = .82; rule breaking or delinquent behaviours α = .69; and externalizing broadband scale α = .83;) as well as boys (aggressive behaviours α = .78; rule breaking or delinquent behaviours α = .69; and externalizing broadband scale α = .81;). The Youth Self Report has been established as a reliable Instrument for assessing externalizing problems among Kosovar youths (see Shahini et al., 2015). Adolescent self-reports and parent reports of externalizing problems are generally found to be modestly associated (Rescorla et al., 2013).

Plan of analysis

To determine pubertal timing, namely early, on time, and late categories, first, scores of the Pubertal Development Scale were standardized (changing into the distribution with a mean of 0, and standard deviation of 1) for each age: 11, 12, 13, 14, and 15 (few cases of 16 were also collapsed with the age 15 group). Then, early was defined to be having pubertal development scores greater than 1 (1); on time = scores between −1 and 1 (2); and late = scores below −1 (3) (see Flannery et al., 1993; Steinberg, 1987).

In a first step, descriptive statistics of the study variables were computed for boys and girls, followed by computation of bivariate Pearson’s correlations. Next, a series of Ordinary Least Squares (OLS) regressions were completed to test the study hypotheses. Due to high correlations between the PDS score and pubertal timing, regression would be carried out separately for each measure. Each regression included socioeconomic status as a control variable on an initial step, followed by BMI, and finally, the PDS score or pubertal timing was added. Analyses were conducted in SPSS 26.

Results

Descriptive statistics of the study variables are reported in Table 1. T-tests were completed to test for mean-level differences in measures of externalizing problems by sex; significant sex differences were found for delinquent (t = −7.77, p < .001), but not for aggressive behaviours (t = −1.53, p = .127). Table 2 presents Pearson’ s correlations of study variables which included both the PDS score and the pubertal timing. Although pubertal timing was not associated with either delinquent or aggressive behaviours in both boys and girls, the PDS score was positively associated with both measures for girls (r = .20 and .33, for delinquent and aggressive behaviours, respectively) as well as boys (r = .21 and .15, for delinquent and aggressive behaviours, respectively). As expected, the two measures of externalizing problems were highly correlated. In girls only, BMI was also positively associated with each externalizing problem (r = .08 and r = .11, for delinquent and aggressive behaviours, respectively).

Table 1. Means and standard deviations of study variables for girls and boys.

	Mean	SD	Min; Max
Girls (N = 665)
Age	13.26	1.27	11; 16
Family SES (parental education)	3.59	1.68	1; 7
BMI	19.76	3.52	8.15; 40.97
Pubertal indices:
1. Growth spurt	2.76	0.70	1; 4
2. Body hair (underarm and pubic hair)	2.62	0.88	1; 4
3. Breast development	2.58	0.70	1; 4
4. Skin change	2.10	0.85	1; 4
5. Menarche (%yes)	71%		0; 1
Pubertal Development Scale (PDS) (alpha = 0.67)	2.63	0.62	0.80; 4
Pubertal timing:
1. Early (%)	17.8
2.On time (%)	68.4
3. Late (%)	13.8
Externalizing symptoms:
1. Delinquent behaviours (alpha = 0.69, 15 items)	0.17	0.18	0; 1.27
2. Aggressive behaviours (alpha = 0.82, 17 items)	0.38	0.29	0; 1.65
Boys (N = 677)
Age	13.19	1.31	11; 16
Family SES (parental education)	3.59	1.71	1; 7
BMI	20.74	4.56	9.23; 51.88
Pubertal indices:
1. Growth spurt	2.57	0.81	1; 4
2. Body hair (underarm and pubic hair)	2.57	0.76	1; 4
3. Facial hair	1.76	0.72	1; 4
4. Skin change	2.03	0.88	1; 4
5. Deepening of the voice	2.39	0.87	1; 4
Pubertal Development Scale (PDS) (alpha = 0.74)	2.26	0.57	1; 3.80
Pubertal timing:
1. Early (%)	16.1
2. On time (%)	65.2
3. Late (%)	18.6
Externalizing symptoms:
1. Delinquent behaviours (alpha = 0.69, 15 items)	0.26	0.21	0; 1.33
2. Aggressive behaviours (alpha = 0.78, 17 items)	0.36	0.27	0; 1.47

Alpha refers to Cronbach’s alpha.

Table 2. Correlations of study variables.

Variables	1	2	3	4	5	6	7
A. Girls
1. Age
2. Family SES	.04
3. BMI	.23***	.04
4. PDS	.60***	.11*	.28**
5. Pubertal Timing: Early	−.04	.07	−.01	−.20***
6. Delinquent behaviour	.22***	.08	.08*	.20**	−.01
7. Aggressive behaviour	.32***	.08*	.11**	.33***	.08	.70***
B. Boys
1. Age
2. Family SES	.06
3. BMI	.08*	.08
4. PDS	.46***	.09*	.17***
5. Pubertal Timing: Early	.02	−.01	.02	−21***
7. Delinquent behaviour	.30***	.04	.05	.21***	.00
8. Aggressive behaviour	.23***	.08	.04	.15***	−.02	.67***

*p < .05, **p < .01, ***p < .001. Timing = pubertal timing. PDS = Pubertal Development Scale; 1 = early, 0 = on time/late.

Results from regression analyses are reported in Table 3. Because pubertal timing was unrelated to delinquent or aggressive behaviours, regression analyses were only carried out using the PDS score. The PDS score was positively associated with both delinquent (β = 0.16) as well as aggressive behaviours (β = .12). The model explained 3.0% in delinquent and 2.2% in aggressive behaviours. For boys, the PDS score was positively associated with both delinquent (β = 0.20) as well as aggressive behaviours (β = .14); the model explained 4.3% in delinquent and 2.7% in aggressive behaviours. Neither socioeconomic status nor BMI uniquely explained variance or was significantly associated with either measure of externalizing problems; this was found in the analyses focused on girls as well as boys.

Table 3. Regressions predicting externalizing problems.

	1. Delinquent behaviours			2. Aggressive behaviours
Predictors	β	SE	p	β	SE	p
A. Girls
Step1:
Family SES	0.04	0.01	.31	0.08	0.01	.054
Step 2:
BMI	0.05	0.00	.24	0.04	0.00	.36
Step 3: PDS	0.16	0.01	<.001	0.12	0.01	.004
R^2	0.030			0.022
B. Boys
Step1:
Family SES	0.04	0.01	.31	0.08	0.01	.054
Step 2:
BMI	0.05	0.01	.24	0.04	0.01	.36
Step 3:
PDS	0.20	0.00	<.001	0.14	0.00	<.001
R^2	0.043			0.027

PDS = Pubertal Development Scale.

Discussion

The present findings replicate evidence that advanced pubertal status measured by the PDS score was positively associated with delinquent and aggressive behaviours in both Kosovar boys and girls, whereas early pubertal timing was unrelated to either. Thus, this provided evidence that it is important to consider both pubertal status as well as pubertal timing separately when testing them as correlates of adolescent externalizing problems.

The evidence showed that boys scored significantly higher on delinquent behaviours in boys in comparison to girls, which is consistent with previous work (Bartels et al., 2011). In addition, Shahini et al. (2015) also found sex differences in rates of rule-breaking behaviours in Kosovar boys in comparison to girls, where boys reported higher rates as compared to girls (for similar evidence on alcohol or substance use, see Carkaxhiu et al., 2011; Tahiraj et al., 2016). Certain behaviours which involve risk taking during puberty, might provide an explanation for these findings. A greater tendency for alcohol consumption has been found in Kosovar girls and boys (Tahiraj et al., 2016), but girls smoke more cigarettes, while boys are more likely to use drugs (Carkaxhiu et al., 2011).

The first study hypothesis that early pubertal timing would be associated with elevated levels of externalizing problems of delinquent and aggressive behaviours in both boys and girls was not supported. This was consistent with previous findings by Susman et al. (2007), who found no support for a positive association between pubertal timing and aggressive behaviours in boys or a positive association between pubertal timing and delinquent behaviours in girls. Similarly, other work has also provided evidence that pubertal timing was unrelated to aggressive behaviours in girls (Graber et al., 2006), or with externalizing problems in boys (White et al., 2013). On the other hand, some other work did find evidence of a link between pubertal timing and externalizing problems (Ge et al., 2006; Kowalski et al., 2021). These discrepancies warrant some explanation. Pubertal events are inextricably linked with the contextual, social, and cultural environment. It is known that girls understand better and have clearer expectations about their bodily developmental changes, based also on communicating with their parents (Belgrave, 2009). Despite the fact that early onset of puberty might negatively impact perceptions of girls’ bodies, thus increasing psychological distress, particularly in societies where being thin is considered the ideal (Natsuaki et al., 2015), several external supports, including maternal support and communication, might promote better adjustment outcomes, including mental health (Branje et al., 2012). On the other hand, for boys, an adult-like physical appearance might even be an asset, increasing self-esteem, which might act as a protective factor against engaging in aggressive or delinquent behaviours. Thus, the developmental readiness hypothesis which asserts that early pubertal timing places youth at greater risk for adjustment problems, particularly early female adolescents, was not supported.

The second study hypothesis that advanced pubertal developmental status/stage measures by the PDS score would be associated with greater externalizing problems, namely delinquent and aggressive behaviours, was supported for both boys and girls, consistent with previous evidence (Ge et al., 2002, 2006; Najman et al., 2009; Oldehinkel et al., 2011). Najman’s et al. (2009) reports similar increases of aggression and delinquency in boys and girls as they progress through puberty. Pubertal status has also predicted substance use in boys in other work (Castellanos-Ryan et al., 2013) as well as alcohol initiation, both in boys and girls (May et al., 2021). The present study findings are different from some previous work which found no support for a relationship between pubertal status/stage and aggressive or delinquent behaviours (e.g. Finkelstein et al., 1994; Nottelmann et al., 1987). The PDS score was positively associated with externalizing problems in Kosovar adolescents, both in boys and girls. Advancing though puberty presents numerous challenges to adapt to bodily changes as well as the role expectations; these adaptation processes might contribute to a greater likelihood of engaging in aggressive or delinquent behaviours. In addition, the observed rates of externalizing problems among Kosovar youth might be related to low socioeconomic status as well as economic development (Mehmeti, 2013); in addition, they might also be related to the numerous stressors resulting from very large social transitions. The contextual amplification hypothesis proposed by Ge and Natsuaki (2009) speaks to this very issue, namely youth who experience pubertal transitions in a particularly disadvantaged developmental context are at substantially greater risk for poor adjustment. In addition, a lack of parental oversight and monitoring might further exacerbate the opportunities to do so (DeVore & Ginsburg, 2005). Boys and girls are more likely to engage in risk taking behaviours in the presence of peers, particularly when they are more advanced in pubertal development; thus, advanced pubertal development is also associated with a greater susceptibility to peer influences (Kretsch & Harden, 2014). Furthermore, during puberty, said it was a sensation seeking increases while impulse control decreases, thus greatly increasing the likelihood of engaging in norm violating and deviant behaviours (Vazsonyi & Ksinan, 2017; Warren & Brooks-Gunn, 1989). Not surprisingly, rates of aggressive and violent behaviours have been shown to be comparatively higher in pubertal stages III and IV (measured by Tanner stage criteria); and the odds of violent behaviours in mid- to late-puberty are three times higher than in the early puberty, independent of age (Hemphill et al., 2010).

No evidence was found that SES was associated with externalizing problems; this was consistent with previous work in the same developmental context, which found that SES was unrelated to rule-breaking behaviours (Tahiraj et al., 2016). However, some work has shown that lower SES is associated with earlier puberty in girls (Oelkers et al., 2021). Contextual challenges during pubertal transitions could play a role in puberty and externalizing problems interactions. Contextual amplification hypothesis asserts that transitioning through puberty at earlier age, in disadvantageous contexts increases the risk for youths’ adjustment behaviours (Ge & Natsuaki, 2009). The inherently challenging societal transitions in Kosovo (e.g. from collectivism to individualism; Shahini et al., 2015) and high rates of poverty (Mehmeti, 2013), might have played an indirect role in the present study findings. In addition, neither sex, nor pubertal status, or hormones are sufficient to be able to capture the complex influences on pubertal development that are known to have an impact in aggressive behaviours (Finkelstein et al., 1994).

It is important to highlight that scholarship on Kosovar adolescents to date has not focused on pubertal development. In addition, no national level data have been collected about pubertal development of adolescents, thus making it more challenging to contextualize the present study findings. This also makes it more challenging to draw comparisons with youth in different developmental contexts. The present investigation found the mean age of menarche to be 12.35 years which appears to be substantially lower than the one reported by Boshnjaku et al. (2016), namely 13.5 years. If this observed change is indicative of the age of menarche nationally, then this evidence might be consistent with the observed secular trends of declining age of menarche elsewhere. In addition, it might also be consistent with age of menarche observed elsewhere, as Biro et al. (2018) recently reported a median age of 12.25 years for girls in the United States; this would also mean that to some extent, age of menarche not context dependent.

It is also important to note that brain structure and function change during adolescence and puberty. The prefrontal cortex of the brain region is responsible for planning ahead, weighing risks versus rewards as well as controlling impulses (Steinberg, 2017). On the other hand, the limbic system is responsible for processing novel experiences, including social information and reward as well as punishment (Steinberg, 2017). Conversely, limbic and motor systems mature earlier in comparison to the prefrontal cortex largely responsible for impulse control (Susman & Dorn, 2009). It is this delayed development which places adolescence at particularly heightened risk for poor decision-making, impulsive behaviour, and thus also externalizing behaviours. With regard to structural brain development, adolescents experience considerable pruning, while the cortical white matter increases; pubertal development plays an important role in cortical reorganization and neuroanatomical changes (Juraska & Willing, 2017). The same findings related to loss of volume has also been made in rats (Juraska & Willing, 2017). In fact, Giedd et al. (1999) find that the volume of white matter increased linearly with age, less so in females in comparison to males, but a non-linear changes in the cortical grey matter.The changes in cortical grey matter were regionally specific, with developmental curves for the frontal and parietal lobe peaking at about age 12 and for the temporal lobe at about age 16, whereas cortical grey matter continued to increase in the occipital lobe through age 20. If the frontal and parietal grey matter peaks approximately one year earlier in girls, tied to pubertal development,, then this suggests potential influences by gonadal hormones (Giedd et al., 1999), which is likely also associated with observed behaviours and adjustment problems.

Lastly, puberty is a complex neuro-endocrine process, and a continuum of events, across different bodily systems influenced by hormone levels. Some behavioural changes might also be associated with changes in hormone levels at puberty. For instance Vermeersch et al. (2008) found a positive association between total and free Estradiol and both aggressive and non-aggressive risk taking behaviours in girls. The relationship between free and total Estradiol and aggressive risk-taking behaviours was stronger for the girls in the middle phase of the menstrual cycle, while non-aggressive risk-taking behaviours were unrelated to cycle effects. In addition, higher levels of androgen (testosterone and testosterone and dihydrotestosterone) were associated with externalizing behaviour in boys (Maras et al., 2003). Based on this evidence, hormonal effects cannot be ruled when trying to better understand how puberty is associated with problem behaviours and should therefore also be considered in future investigations.

Study Limitations

The cross-sectional study findings contribute to a better understanding of the associations between early sexual maturation and externalizing problems in Kosovar adolescents, a gap identified in Kosovar scholarship. The large sample and the developmental context are certainly study strengths; however, several limitations require some discussion. First, data were cross-sectional in nature and therefore no causality can be inferred. Related to this, because data were only assessed at one point of time, developmental processes including sequences of pubertal indices, tempo or Peak Height Velocity could not be investigated. Each of them is important and known to contribute independently to the associations between puberty and psychosocial adjustment in adolescents. Thus, additional longitudinal work is needed. Another limitation is related to measurement. It was not possible to assess ejaculation, or Oigarche among boys, a known proxy for pubertal development among boys much like menarche among girls; however, this challenge and limitation is a common problem in the literature. A further study limitation includes the sole reliance on self-reports which therefore introduces the potential for monomethod bias. In other words, the observed relationships might be inflated due to shared method variance. Future work needs to employ multiple informants, which might include physician reports or parent reports in order to address some of the main study constructs. Next, because the present investigation relied exclusively on adolescents attending school, a sample selection bias likely limits the generalizability of the present study findings. For instance, students not attending schools or not attending different types of schools, such apprenticeships or technical training, were not included in the present investigation. In this sense, the present results based on this sample provide a very modest first step on understanding puberty among Kosovar youth, by no means nationally representative of the Kosovar developmental context. Future work clearly needs to address this limitation by sampling youth across different educational tracks as well as accessing youth that might be not enrolled in structured school settings and simply working, for instance, as well as youth from across the country. Lastly, due to the fact that only specific adjustment correlates were considered and tested, a number of other known correlates were not, thus likely further limiting the implications of the present work.

Conclusions

The current investigation extends previous knowledge about adolescents from causal role by showing that accelerated pubertal development presents a risk for the development of externalizing problems. Specifically, advanced pubertal stage/status posits a higher risk for both sexes, and early pubertal timing posits no risk for either girls or the boys in the study. The following conclusions can be made: a) advanced pubertal development is an important factor in understanding variability in externalizing problems among Kosovar youth; b) early pubertal timing was unrelated to externalizing problems in the present sample, thus additional replication evidence is needed to contextualize study findings; 3) pubertal timing and pubertal development have important, yet independent effects on behavioural adjustment during pubertal transitions in Kosovar boys and girls; and finally, 4) although not directly tested in this study, a series of contextual stressors prior to puberty, due to growing up in a highly disadvantaged socio-economic context, characterized by constant transitions, might also have contributed to the present study findings. Future work will need to try to assess this question in greater detail by potentially carrying out cross-cultural comparative studies to better understand these potential effects.

The current work fills important gaps in the literature to better understand the extent to which pubertal development is associated with externalizing problems among Kosovar adolescents. Findings highlight the need for a closer look at other cofounding variables, including parental and peer influences, to more fully understand the associations between puberty and externalizing problems.

Ethical statement

The study followed all ethical guidelines of Helsinki declaration. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

Informed consent from parents and adolescents were obtained from all participants included in this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets generated during and/or analysed during the current study are available from the corresponding author, on reasonable request.

Additional information

Funding

The funding for the present publication was provided by the Czech Ministry of Education, Youth and Sports for specific research (IGA_FF_2021_008).

Notes on contributors

Elona Krasniqi

Elona Krasniqi is a PhD student at the Department of Psychology, at Palacky University Olomouc. Her dissertation work explores the correlates of pubertal development as well as the impact of puberty on adolescent emotional and behavioral adjustment. She also focuses on developmental changes in parent-adolescent relationship during puberty as well the way maternal parenting behaviours impact a daughter’s emotional adjustment is a PhD student at the Department of Psychology, at Palacky University Olomouc. Her dissertation work explores the correlates of pubertal development as well as the impact of puberty on adolescent emotional and behavioral adjustment. She also focuses on developmental changes in parent-adolescent relationship during puberty as well the way maternal parenting behaviours impact a daughter’s emotional adjustment.

Alexander T. Vazsonyi

Alexander T. Vazsonyi is Professor of Family Sciences, Professor of Psychology, and Professor of Sociology at the University of Kentucky. He conducts research in cross-cultural comparative developmental science and in criminology focused on adolescents and has published over 180 peer reviewed papers and chapters. His major research interests include self-control or self-regulation, adjustment and well-being (problem behaviors, deviance, violence, and health compromising behaviors), with a focus on development in context (family, school, neighborhood, and culture).

Panajotis Cakirpaloglu

Panajotis Cakirpaloglu is a Professor of Social and Personal Psychology, Psychological Axiology, Developmental Neuropsychology at the Department of Psychology at Palacky University Olomouc. His major research focuses on personality psychology, psychology of values, as well adolescent personality. He has been publishing peer-review papers, book chapters, as well delivered international guest lectures including United States, Macedonia, Portugal, and so fourth.