Attachment disorganisation and poor maternal discipline in early childhood: independent contributions to symptoms of conduct problems

ABSTRACT

Conduct problems (CP) are common behaviour difficulties in young children. Poor parenting and caregiver-child attachment relationships are important risk factors of CP, but more research is required to understand their concurrent contribution to CP in the early childhood years. We examined the association of maternal-reported conduct problems (CP) (Strengths and Difficulties Questionnaire) with maternal-reported harsh and inconsistent discipline (Preschool-Alabama Parenting Questionnaire) and children’s representations of attachment (Manchester Attachment Story Task) in a sample of children between 4 and 6 years old (n = 66). Correlation analysis showed that CP were associated with attachment insecurity, high levels of attachment disorganisation, and maternal harsh and inconsistent discipline. Regression analysis showed that disorganisation and inconsistent discipline explained unique variance in CP ratings even after controlling for the effects of confounding variables. The findings suggest that attachment disorganisation and aspects of poor maternal discipline in early childhood reflect parallel processes with unique contributions to CP.

KEYWORDS:

• Conduct problems
• attachment representations
• discipline
• parenting
• early childhood

Conduct problems (CP) such as oppositionality, defiance, and aggression are amongst the most commonly reported types of behaviour difficulties in early childhood (Vasileva et al. 2021). Persistent symptoms of early childhood CP are associated with a higher risk of long-term psychosocial disadvantage including poor mental health and substance abuse (Bevilacqua et al. 2018), and poor academic achievement, welfare dependence, and offending (Fergusson, John Horwood, and Ridder 2005). In addition to costs for the individual, early childhood CP were related to high societal costs, including higher levels of public service use across the criminal justice, health care, and social welfare systems, and extra educational provision (Rivenbark et al. 2018). The impact of early childhood CP on the individual and the society highlights the importance of investigating the factors that contribute to their development.

Poor parental discipline is one of the stronger predictors of symptoms of CP. Explanatory models of antisocial behaviour rooted in social learning theories (Dishion and Patterson 2006) and empirical studies have shown that the frequent use of harsh discipline is often associated with behaviour difficulties (Pinquart 2017). A common form of harsh discipline is physical and verbal punishment (Gershoff and Grogan-Kaylor 2016; Wang and Kenny 2014). Physical punishment refers to non-abusive use of physical force to change behaviours such as spanking (Gershoff 2002; Paolucci and Violato 2004). The application of physical punishment during the preschool years is found to be associated with high levels of behaviour problems, such as aggression and CP, that persist or increase across development (Flouri and Midouhas 2017; Gershoff and Grogan-Kaylor 2016; Olson, Choe, and Sameroff 2017). Harsh verbal discipline refers to the use of psychological force to cause emotional discomfort, such as shouting, and has been linked to symptoms of CP (Wang and Kenny 2014). Another form of poor parenting is inconsistent discipline, which refers to the inconsistent application of non-harsh clearly articulated rules of appropriate behaviour and of consequences to misbehaviour (Grusec and Davidov 2010). Examples of inconsistent parenting include repeating instructions over and over or the intention to apply consequences without following through. Parents of children with symptoms of CP have been reported to find it difficult to follow-through with their commands in discipline situations (Cheung et al. 2018; Gardner 1989; Stanger et al. 2004; Stormshak et al. 2000).

In addition to parenting, the child’s early relationships with the primary caregiver are associated with positive developmental outcomes. According to attachment theory, a secure parent–child attachment relationship fosters an appropriate context for child socialisation and socioemotional development (Kochanska and Kim 2012). Because children with secure attachment relationships manage emotions and interpersonal relationships successfully, they are less likely to develop behaviour problems (Boldt, Kochanska, and Jonas 2017; Bureau et al. 2017; Fearon et al. 2010; Scott et al. 2011; Theule et al. 2016). The disorganised types of insecure attachment and externalising behaviour, including symptoms of CP, are consistently linked in the literature. An earlier meta-analysis found an association between infant disorganisation and externalising behaviour in early childhood (Van Ijzendoorn, Schuengel, and Bakermans-Kranenburg 1999). A following meta-analysis showed that disorganisation measured by observational measures in children 12 years of age or younger was more strongly associated with externalising behaviour (i.e., aggression, oppositional problems, conduct problems or hostility) than other types of insecurity independent of child clinical status, socioeconomic background, or type of assessment (Fearon et al. 2010).

Concurrent examination of the role of poor parenting and attachment relationships in the development of CP can help disentangle further different explanatory models of behaviour difficulties and increase specificity in targets for intervention. An important question is whether maternal poor discipline and mother–child attachment relationships are parallel processes with unique contributions to symptoms of CP or whether they influence one another. Extant research shows that attachment security and organisation in the infant years provide a socialisation context, which can influence the impact of power assertive parenting on early childhood externalising difficulties (Kochanska and Kim 2012; Cyr et al. 2014; Keller, Spieker, and Gilchrist 2005; Granqvist et al. 2017). The concurrent examination of attachment and caregiver discipline and their implication for children’s emotional and behavioural outcomes beyond infancy is limited and less understood (Koehn and Kerns 2018). Nonetheless, the effects of parenting behaviour are better appreciated when examined in the context of the specific developmental period wherein they are more conspicuous (Rothenberg et al. 2020). As children grow older, parents use more discipline (Lawrence et al. 2019) and late early childhood is a developmental period where the use of harsh discipline has been found to peak (Berthelon et al. 2020). This is because often parents in their efforts to control the child’s growing autonomy use assertive forms of discipline such as physical and verbal punishment. Additionally, the developmental changes that take place in the preschool years bring about changes in the expression of attachment disorganisation, which are reflected in children’s controlling and punitive modes of interaction (Keller, Spieker, and Gilchrist 2005). The increased use of assertive discipline by parents and the child’s capacity to negotiate and to express in a more complex manner may have implications for parent–child interaction that are specific to this developmental age. Therefore, the consequences of insecure attachment and poor parenting may become more prominent later in the preschool years (Fearon et al. 2010; Keller, Spieker, and Gilchrist 2005), which makes them a critical period for the examination of their impact on child behaviour.

Existing studies that could shed light on the association of CP with child-mother attachment relationships and maternal discipline in the preschool years support an independent effect of harsh parenting on children’s CP. However, they did not examine attachment disorganisation. For instance, a recent study of a large US-based cohort sample of low-income urban families found that maternal spanking at the age of three predicted externalising difficulties at the age of five independently of attachment security assessed by mother-report and after controlling for the effects of maternal age, race, education, depression, parenting stress, household income, marital status, and child gender and temperament (Ward et al. 2020). A study of a Belgium-based sample of clinic-referred children found that high maternal reported coerciveness, including physical punishment and attachment insecurity in children’s representations in early childhood, predicted independently increasing externalising behaviour over time (3–14 years old) (Roskam 2018). In contrast, studies using attachment-based measures, such as parental warmth and sensitivity, have demonstrated both independent (Lee, Altschul, and Gershoff 2013; Altschul, Lee, and Gershoff 2016; Xing and Wang 2017) and moderated effects (Germán et al. 2013) of physical discipline on externalising behaviour. The mixed findings, the great variety across studies in design and conceptualisation of maternal discipline, and the lack of measures of disorganised attachment make it difficult to come to firm conclusions about the role of maternal harsh discipline and attachment security and disorganisation in CP in the late early childhood years.

Another important limitation in this line of research is the restricted number of studies on the application of non-harsh discipline and parent–child attachment relationship (Koehn and Kerns 2018). Consistency in the application of discipline may provide children with a sense of predictability about parent–child interactions, which in turn may help to promote a secure parent–child attachment relationship (Wolchik et al. 2000). On the contrary, inconsistency may give mixed signals as to how the child should behave in certain situations. Difficulty in interpreting parental cues can lead to more instances of inappropriate behaviour and coercive transactions with the caregiver, which can reduce the chances of using the parent as a safe base and heaven (Wolchik et al. 2000; Kobak, Zajac, and Madsen 2016; Lyons-Ruth and Jacobvitz 2016). The investigation of the impact of early childhood attachment relationship and parenting on CP would benefit by introducing a wider phenotype of parental discipline.

Measurement of the attachment relationship is another important issue in this line of research. A lot of the research on the association between parenting and attachment takes place in the infant years (Koehn and Kerns 2018). Therefore, our knowledge in this area is significantly shaped by findings from studies that use the classic separation-reunion assessment method of caregiver–child attachment. However, the use of the classic method becomes challenging as children grow older due to developmental changes in the attachment behaviour (Allen et al. 2018). Equally, the use of self-report measures in later early childhood or early primary years is challenging because of the child’s limited cognitive and verbal skills to handle their demands (Jewell et al. 2019). Therefore, measures based on symbolic representation of attachment are common with preschool or early primary school age children (Kerns and Brumariu 2016). However, research using this method in the context of parenting is scarce (Allen et al. 2018).

While broadband dimensions of externalising difficulties reflect the great heterogeneity found in child psychopathology, narrowband dimensions help us to identify the outcomes associated with specific behaviours and understand intervention specificity (Achenbach et al. 2016; Kulawiak et al. 2020). A lot of the research on parenting and attachment is focused on externalising behaviours; however, CP symptoms are highly comorbid with attention deficits and hyperactivity (Biederman, Newcorn, and Sprich 1991; Harvey, Breaux, and Lugo-Candelas 2016). Some studies have identified a clear link between attachment security and inattention and hyperactivity (Storebø, Rasmussen, and Simonsen 2016) but others show that disorganised attachment is more likely to be associated with symptoms of CP (Forslund, Peltola, and Brocki 2020). Therefore, the study of the contribution of harsh and inconsistent parenting and insecure attachment on CP should account for associated symptoms of inattention and hyperactivity.

In summary, there is limited research on the concurrent impact of early childhood attachment relationship and parenting on children’s CP. Additionally, more research is required to understand the association using representational measures of attachment and accounting for the effects of associated inattention/hyperactivity. Therefore, the aim of the study was to examine whether the mother-child attachment relationship impacts on child behaviour independently of maternal discipline or whether maternal discipline can influence the child’s behaviour irrespective of the quality of their attachment relationship.

Methods

Participants and procedures

Seventy-two (N = 72) biological Greek mothers of children between 3 and 6 years old and free from developmental delays according to school records were recruited by the second author through informed consent. Recruitment took place in a nursery and preschool centre in a provincial town. We used the data of mothers who returned a completed questionnaire (n = 66; M_age = 37.6 years, SD = 4.81; range = 27.05–51.33 years) and whose children (n = 66; M_age = 5.03 years, SD = .63; range = 4.01–6.58 years; 48% girls) had a completed attachment assessment. Most participants were married (97%) working mothers (67%) and half of them (54%) had completed tertiary education (university/higher technical institution). National data on the Greek female population showed that in 2019 51.3% of females between 20 and 64 years old were in employment (Statistical Office of the European Communities 2021) and 36.6% of females between 25 and 64 years old had tertiary education (Statistical Office of the European Communities 2020). Therefore, the mothers in our sample were more likely to be in employment and have a higher education degree than the mothers in the general population. A summary of participant characteristics is presented in Table 1. Ethical approval was obtained by the Ethics Research Committee of the University of Roehampton. To thank the mothers for their participation, the first author delivered a 2-hour session on child development and ideas on parenting.

Table 1. Mother (n = 66) and child (n = 66) characteristics

	f	%
^aTertiary education	36	54
^bIn employment	44	67
^cMarried	64	98
Girls	32	48
Age	M	SD
Mother	37.6	4.81
Child	5.03	.63

^aTertiary (university/higher technical institution and/or postgraduate studies vs non-tertiary education (up to high school at 17 years and/or further education college)); ^bin employment vs non in employment; ^cmarried vs divorced/separated.

Measures

Attachment representations in children’s story stem narratives

The Manchester Child Attachment Story Task (Goldwyn et al. 2000; Green et al. 2000) is a common story stem narrative assessment method of attachment that triggers storytelling and symbolic play to access internal representations of caregiver-child attachment in children between 4 and 8 years old. Children use dolls to enact four-story stems (nightmare, hurt knee, feeling ill, lost in store) started by the administrator, and they are asked standardised follow-up questions to gain a fuller picture of the child and caregiver thoughts and feelings. The task was administered and video-recorded in a room in the centres’ premises by the first author, who is trained by the MCAST developers and experienced administrator and coder. A task manual is used to rate each story on a 9-point scale for a) task engagement and increase in arousal; b) child attachment-related behaviour (e.g., proximity seeking); c) caregiver behaviour (e.g., parental responsiveness/sensitivity); d) effectiveness of assuagement; e) narrative coherence; and f) attachment disorganisation based on the frequency of bizarre and disoriented phenomena. A 3-point scale is used to assess mentalising of the self and mother. The ratings are used to classify each story to the four categories of attachment: Secure (B), Insecure-Avoidant (A), Insecure-Ambivalent (C), and Disorganised (D). The child’s overall attachment style is based on the attachment classification of each story. To examine the level of disorganisation, we used the child’s total mean score (range: 1 − 9) of the frequency of disoriented and disorganised phenomena across the four stories. The first author coded all the videos from children whose mothers consented to participate (n = 72) and a second independent coder who is a clinical psychologist and is trained to gold-standard by the developers of the MCAST and blind to participant demographics and ratings coded 20 videos. Interrater agreement for the 2-way and 4-way classification was κ = .90 and .77, respectively. The intraclass correlation coefficient for the total disorganisation score was r = .86, p = .00.

Harsh and inconsistent discipline

We used the total score of the Inconsistent Parenting (seven items, e.g., You threaten to punish your child and then you do not actually punish him/her; range: 0–35) and Punitive Parenting (five items, e.g., You yell or scream at your child when he/she has done something wrong; range: 0–25) scale of the preschool (3–5 years) Alabama Parenting Questionnaire (APQ-PR; Clerkin et al. 2007). This is an adapted version of the widely used original 42-item APQ (Essau, Sasagawa, and Frick 2006; Shelton, Frick, and Wootton 1996) that lists 32 items rated on a 5-point scale (1 = never – 5 = always). The APQ-PR (Clerkin et al. 2007) has good internal reliability (Inconsistent Parenting α = .74; Punitive Parenting α = .63) and correlates with child psychopathology (Cova et al. 2017; Loginova and Slobodskaya 2017). A Greek version of the original 42-item APQ was obtained from the Greek team of the FemNAT-CD study (FemNat-CD 2016) to create the APQ-PR and included the following adaptations: the item about hitting with an object was modified by eliminating the examples (e.g., belt and switch) to avoid disquieting the parents.

Child behaviour

Mothers completed the Greek version of the Strengths and Difficulties Questionnaire (SDQ) for children between 4 and 17 years old (Goodman 2001, 1997) available at: https://www.sdqinfo.org/py/sdqinfo/b0.py. We used the total mean score of the 5-item CP and hyperactivity subscales rated on a 3-point scale (0 = not true, 1 = somewhat true, 2 = certainly true; range: 0–10). The SDQ is a screening questionnaire of behavioural and emotional difficulties that is used internationally (Stone et al. 2010). The internal structure and reliability of the Greek version of the parent-reported SDQ subscales are satisfactory (α range = .71–.81) (Bibou-Nakou et al. 2019).

Results

Preliminary analysis

Half (50%) of the sample had a secure attachment. Of the children classified insecure, nearly one in five was classified as avoidant (n = 13; 19%) followed by disorganised (n = 11; 17%) and ambivalent (n = 9; 14%). Table 2 presents the correlations between the study variables. Higher ratings of CP symptoms were associated significantly with being a boy (r = −.29, p < .05) and symptoms of hyperactivity (r = .36, p < .01). Additionally, there was a tendency for mothers with tertiary education to report lower rating of CP symptoms (r = −.22, p = .09). Insecure attachment classification was significantly associated with being a boy (r = −.30, p < .05), and higher ratings of CP (r = .32, p < .05) and hyperactivity (r = .34, p < .01). Higher disorganisation ratings were associated significantly with being a boy (r = −.25, p < . 05), lone motherhood (r = −.24, p < .05), and high ratings of CP (r = .34, p < .01). More frequent use of inconsistent and harsh discipline was associated significantly with not having tertiary education (r = −.30, p < .05; r = −.42, p < .00), and with higher ratings of CP (r = .51, p < .00; r = .31, p < .05) and hyperactivity (r = .34, p < .01; r = 31, p = < .05). Additionally, there was a tendency for higher inconsistent discipline to be correlated with lone motherhood (r = −.22, p = .07).

Table 2. Correlations among study variables

n = 66	1	2	3	4	5	6	7	8	9	10	11	12	M (SD)
1. ^aAttachment security	-
2. Disorganisation	.59**	-											3.36 (1.93)
3. ^bChild gender	−.30*	−.25*	-
4. Child age	−.10	−.11	−.17	-
5. Maternal age	.10	.08	−.07	.31*	-
6. ^cMaternal employment	−.06	.00	.02	−.05	−.17	-
7. ^dMaternal education	.00	−.05	.03	−.01	.15	−.06	-
8. ^eMarital status	.00	.24*	.01	.08	−.18	.06	−.02	-
9. CP	.31*	.34**	−.31*	.11	−.04	−.07	−.22	.01	-				1.89 (1.54)
10. HYP	.34**	.07	−.19	−.16	.00	−.07	−.13	−.23	.37**	-			2.09 (2.24)
11. Harsh discipline	.08	.13	−.12	.05	−.01	−.04	−.42**	.04	.31*	.30*	-		7.61 (1.96)
12. Inconsistent discipline	.18	.02	−.17	.06	−.08	−.18	−.30*	−.22	.50**	.34**	.37**	-	16.19 (4.16)

^aSecure = 0.

^bBoy = 0.

^cIn employment = 1.

^dUp to high-school and further education = 0/Tertiary (university/technical) = 1.

^eMarried = 0/Divorced/separated = 1.

* p ≤= .05; **p ≤= .01.

Hierarchical linear regression analysis

We performed two 4-step hierarchical regression analyses to examine the independent effects of attachment (insecurity/disorganisation) and maternal discipline (harsh/inconsistent discipline) variables and the effect of their combination on the variance of maternal ratings of CP. To examine the combined effects we created four interaction terms (Table 3) after mean-centring the continuous variables (Tabachnick and Fidell 2007). In the first step, we entered the study variables that were significantly associated with CP ratings to control for their effects (Tabachnick and Fidell 2007): child gender and hyperactivity (Table 2). Additionally, we included maternal education as there was a tendency for a medium strength association with CP. In the second step, we entered the maternal discipline variables; in the third step, the attachment variables; and in the fourth step, the interaction variables. Based on the recommendations for checking multicollinearity (Tabachnick and Fidell 2007) inspection of the Collinearity Diagnostics did not reveal any multicollinearity (conditioning index range across both regressions: 1.00–5.61).

Table 3. Summary of hierarchical regression analysis for variables predicting maternal reports of conduct problems (n = 66)

	Step 1		Step 2		Step 3		Step 4
	B (SE)	b	B(SE)	b	B(SE)	b	B(SE)	b
Gender	−.73 (.34)	−.25*	−.60 (.32)	−.20	−.50 (.33)	−.17	−.43 (.33)	−.14
Hyperactivity	.20 (.08)	.30**	.12 (.08)	.18	.09 (.08)	.14	.11 (.08)	.16
Maternal education	−.51 (.34)	−.17	−.14 (.35)	−.05	−.16 (.35)	−.06	−.26 (.35)	−.09
Harsh discipline (HD)			.06 (.09)	.08	.06 (.09)	.08	.15 (.13)	.19
Inconsistent discipline (ID)			.13 (.04)	.37**	.13 (.04)	.36**	.07 (.06)	.20
Attachment security (AS)					.41 (.34)	.14	.40 (.34)	.14
HD x AS							−.21 (.18)	−.22
ID x AS							.13 (.08)	.26
R^2	.19*		.29**		.30**		.31**
Gender					−.40 (.31)	−.13	−.39 (.32)	−.13
Hyperactivity					.12 (.07)	.18	.12 (.07)	.17
Maternal education					−.13 (.33)	−.05	−.13 (.34)	−.05
Harsh discipline (HD)					.03 (.09)	.04	.03 (.09)	.04
Inconsistent discipline (ID)					.14 (.04)	.39**	.14 (.04)	.38**
Disorganisation (D)					.22 (.08)	.28**	.22 (.08)	.29**
HD x D							.02 (.05)	.04
ID x D							−.01 (.02)	−.03
R^2					.36**		.34

Because the values in Step 1 and 2 of the second regression are the same with the values in Step 1 and 2 of the first regression they are not presented.

* p ≤ .05; **p ≤ .01.

The first regression (Table 3) showed that child gender and hyperactivity explained a significant amount of the variance (19%) in children’s CP ratings (Step 1). After the effects of gender, hyperactivity, and maternal education were held constant inconsistent discipline explained a significant proportion of the variance (29%) in children’s CP ratings (Step 2). Attachment security status did not explain a significant proportion of the variance in CP ratings (30%) while controlling for the effects of child and mother characteristics and maternal discipline (Step 3). However, inconsistent discipline was significantly related with CP even after considering the effect of attachment security. The contribution of the interaction effects to the variance of CP rating was negligible and not statically significant (Step 4).

The second regression (Table 3) equally showed that child gender and hyperactivity explained more than one-fourth of the variance (29%) in children’s CP scores (Step 1). Additionally, inconsistent discipline predicted change in CP ratings significantly across all steps. In step 3 higher rates of disorganisation made a significant and independent contribution to CP ratings and improved the prediction of change in CP from 29% (Step 2) to 36%. The interaction effects did not make a significant contribution to the variation of CP (Step 4).

Discussion

We examined the independent association of maternal harsh and inconsistent discipline and of children’s representations of attachment insecurity and disorganisation and their interaction with symptoms of CP in a community sample of children between 4 and 6 years old. The findings indicated that disorganisation in the attachment relationship and inconsistent parenting were significantly associated with CP even after controlling for the effects of child gender and hyperactivity, maternal education, and the presence of one another. These findings align with earlier research, which shows that CP are associated with inconsistent discipline (Raudino et al. 2012; Stormshak et al. 2000) and disorganisation (Fearon et al. 2010; Groh et al. 2012; Van Ijzendoorn, Schuengel, and Bakermans-Kranenburg 1999). Collectively, they suggest that lack of consistency in the application of discipline by mothers can impact child behaviour regardless of the quality of the child-mother attachment relationship. While harsh discipline was significantly associated with CP, the regression showed that harsh discipline did not predict children’s CP ratings once the effects of maternal education were held constant. The findings suggest that the relationship between harsh discipline and CP problems in our sample was moderated by maternal education. The literature shows that the number of years parents spent in education is strongly associated with how often the apply coercive discipline (Berthelon et al. 2020). More educated caregivers are more likely to apply effective discipline because they may have more access to knowledge around appropriate caregiving and resources to facilitate it (Roubinov and Boyce 2017). Therefore, a plausible explanation of the association between harsh discipline and CP is caregiver lack of awareness and knowledge around appropriate child-rearing practices. Attachment security did not make an independent contribution to CP ratings. The findings could be explained by earlier research using the MCAST, which has shown that the association of externalising difficulties with disorganisation ratings is more consistent than with insecurity classification (Futh et al. 2008; Barone and Lionetti 2012; Green, Stanley, and Peters 2007).

The interaction between the quality of the attachment relationship and aspects of maternal discipline did not contribute significantly to children’s symptoms of CP. The impact of the combination of coercion and attachment insecurity and disorganisation on a child's social and emotional symptoms can vary depending on the child’s family ecology (Kochanska and Kim 2012). In line with the literature on cumulative developmental risk, a higher number of family stressors is associated with more child vulnerability (Bernier and Meins 2008). In non-risk samples, because the chances of family stressors and child vulnerability are low, a more unusual and extreme parenting typically associated with disorganisation would be required to increase the chances of poor child outcomes (Bernier and Meins 2008). The adversity of the upbringing environment in our community sample was probably not significant enough to interfere with the child’s representations of attachment. It is plausible that in our sample use of poor discipline was implicated by effects such as caregiver education and caregiver awareness and knowledge around appropriate child-rearing practices than factors pertaining to pathological parenting.

Implications for theory, practice and research

The finding for an independent contribution of maternal discipline to children’s CP compliments findings by previous studies which showed that aspects of maternal discipline in the preschool years predicted externalising difficulties regardless of preschool mother-child attachment relationship in low income and clinical samples (Ward et al. 2020; Roskam 2018). Together, these lines of research suggest that maternal discipline and attachment relationship in early childhood may reflect parallel processes with deleterious effects on child behaviour across samples of varying levels of socioeconomic disadvantage. Moreover, they favour the argument that the attachment relationship in early childhood may not act as protective factor of poor maternal discipline (Ward et al. 2020). Future studies should consider whether this finding can be replicated with multiple measures of attachment and informants and in the middle childhood years. Additionally, future research could consider the role of the cultural context in the association between discipline, attachment relationships and CP in early childhood. The magnitude of the negative effects of corporal punishment varies depending on cultural perceptions about its normativeness (Altschul, Lee, and Gershoff 2016). The research on the upbringing values and practices of Greek parents is scarce but the application of strict parental control found by a few earlier studies (Olivari et al. 2015) may suggest that the Greek society is tolerant to the use of power assertive discipline. The sociocultural fabric of the Greek society combined with lack of access to knowledge about appropriate parenting may render some mothers in the community more likely to use harsh discipline. These well-intended, but otherwise ineffective discipline methods may influence the child’s behaviour without necessarily harming the mother-child relationship. In contrast, when the use of physical discipline is not acceptable, harsh discipline may interfere with attachment. An important future direction of this line of research is the examination of how the intersection between Greek parents’ upbringing values and sociodemographic characteristics can shape child social and emotional development. Finally, future examinations should adopt a developmental approach. Pre-schoolers may perceive the use harsh parenting less negatively than children in other developmental groups (Grusec et al. 2017). Research shows that parents engage in negative practices more as children grow older but for a time-limited period until children develop the skills to manage their behaviour as they move from the infant to the toddler and early childhood years (Berthelon et al. 2020; Lawrence et al. 2019). It is plausible that frequent use of non-abusive harsh discipline does not impact on the attachment relationship dramatically because it is considered normative by pre-schoolers especially in upbringing contexts where it is endorsed by dominant cultural values about appropriate discipline. On the contrary, the frequent use of physical force in the infant years might reflect atypical parenting and therefore have a more prominent impact on parent-child relationships.

The finding of a significant association between inconsistent discipline and CP regardless of maternal education suggests that the association should be examined more closely to identify why parents engage in this type of discipline. Advancing our understanding of specific parenting practices could inform the research on how individual components of parenting interventions (e.g., discrete disciplinary techniques taught to parents) are likely to influence intervention efficacy (Leijten et al. 2015). The link between harsh punishment and maternal education suggests that some mothers in the community are at a greater risk of using poor discipline and may require more organised support in the form of widely accessible parenting programmes or health visiting programmes. This is in alignment with research that supports the differentiation of treatment for disruptive child behaviour based on the needs of the target families (Leijten et al. 2019).

Previous research has underscored the role of neurological vulnerability (Bernier and Meins 2008; Spangler et al. 2009), comorbidity with ADHD, and emotional regulation in disorganised behaviour (Forslund, Peltola, and Brocki 2020). Our findings were free from the effects of neurological impairments according to school records and hyperactivity, which may have interfered with children’s performance on the story stem narrative task. Future research should consider the role of emotional regulation.

Limitations

The findings should be understood in the context of the study’s limitations. Most children had mothers with tertiary education and came from families based in a rural area; therefore, the findings may not be generalised to samples that do not have the same characteristics. However, they have implications for similar contexts, including rural communities in Greece. We can be confident in the associations detected, but future research should replicate the association with a larger sample to reduce the likelihood for Type II error. We used the APQ, a well-established measure of parent-reported use of disciplinary practices. Future replication and extension of the present research could consider a multi-method assessment approach and include independent observer ratings of parenting behaviour. Additionally, a single-method approach may limit the chances of sampling the construct of attachment successfully (Kerns and Brumariu 2016). Therefore, the use of multiple assessment methods (e.g., representational and reunion) may allow for a more thorough assessment of attachment. Another limitation is that the main coder of the MCAST was not blind, but a high level of convergence with the reliability coder who was blind was achieved.

Disclosure statement

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

Additional information

Notes on contributors

Angeliki Kallitsoglou

Angeliki Kallitsoglou is a senior lecturer in child psychology and special education and a chartered psychologist. She studies children’s social and emotional outcomes in the context of families and schools.

Vasiliki Repana

Vasiliki Repana is an early childhood care professional with expertise in early years settings.