Intergenerational (IG) continuity of parenting practices delves into sharing parenting behaviors, values, and techniques across generations. Researchers started intergenerational parenting studies based on assumptions that parenting knowledge, practices, and styles are passed down to the following generations, and the following generations mimic parenting behaviors of their parents knowingly or unknowingly. IG parenting studies were started with retrospective data based on these appealing intuitions. However, because of the limitations of recall biases and people’s tendency to validate their parenting behavior, researchers initiated prospective IG studies in the 1990s [1]. Terms such as intergenerational transmission, intergenerational continuity-discontinuity, and intergenerational stability-instability have been used in IG studies that communicate similar meanings with slight differences. Van Ijzendoorn [2, pp 76-77] defined intergenerational transmission of parenting as “the process through which purposively or unintendedly an earlier generation psychologically influences parenting attitudes and behavior of the next generation.” However, it extends beyond individual families, as societal, cultural, and environmental factors play vital roles in this process [3,4]. Intergenerational stability denotes similarities in rank order, whereas intergenerational continuities denote similarities in mean levels of behaviors across generations [1]. However, the term intergenerational continuity-discontinuity appeals to us in this fundamental study in the Bangladeshi context because we believe that the exploration of similarities and differences across generations should come first before stability rank order. We consider that intergenerational continuity is an active process where, consciously or unconsciously, both generations play an active role in the process.

Child developmentalists have been intrigued by the question, “Why do parents parent in the way they do?” [5,6]. Researchers determined to answer this question focusing on IG transmission/ continuity of parenting practices. Numerous studies investigating IG parenting have provided supporting evidence using multigenerational samples [2,7–10].

A literature review by Kerr and Capaldi [1] reported the paucity of IG parenting studies, and that the available studies mostly focused on problem outcomes such as psychopathology and substance abuse other than parenting itself. There are countless studies on the intergenerational transmission of corporal punishment and abusive and harsh parenting [5,11–17], while other parenting aspects received less attention. A couple of studies are found in the area of intergenerational transmission of attachment [18,19], parental monitoring [11], parental rejection [20], emotional warmth and acceptance [21], constructive parenting [22], and initial reaction [23]. Yet, exploration of intergenerational continuity of specific parenting dimensions that comprise a set of parenting practices is limited. As such, research on the intergenerational continuity of protective parenting dimension is limited except for the practice of parental monitoring, which is one of the practices of protective parenting dimension. While researchers in Western countries have examined intergenerational parenting transmission; parenting in general, protective parenting practices in particular, remains understudied in Bangladesh [24]. This study addresses research gaps by exploring the intergenerational continuity of protective parenting practices among urban-dwelling middle-class Bangladeshi mothers.

Modernization effects, economic and historical changes in society, culture, and the external world affected the context, attitude toward certain parenting behaviors, and outcomes [1]. After approaching millennium two, the number of women joining the workforce increased remarkably, which empowered the current generation of mothers in many ways, more importantly, financially and in decision-making. However, it also caused a mother’s absence in childcare for a long period of a day. This indicated a possibility of intergenerational change in parenting practices and sharing parenting responsibilities with the father and/ or other family members. Increased substance usage in society, availability of contraceptives to decrease teenage pregnancy and sexually transmitted diseases, availability of cell phones, and parents’ and children’s exposure to the internet and social media, have a bearing on real and perceived threats for children [1]. This appeals to different approaches to parenting while fundamental parenting concepts and practices could remain similar.

The impact of modernization might cause the dynamics of intergenerational continuity of parenting practice to be more complex than ever before. Traditional parenting knowledge and cultural practices, that conserve distinct customs and contribute to building resilience [25], might be lost if not passed over through media, mainstream education, and documentation [26]. Besides, intergenerational continuity of parenting practices in a shared communal culture like in Bangladesh might not be the same as in individualistic Western societies. In-depth exploration would inform the planning of parenting and child development programs for Bangladeshi communities to promote positive outcomes while reducing the negatives. In-depth exploration would also suggest effective modes of sharing and preserving parenting knowledge and would enrich Indigenous and global parenting literature.

Psychologists and anthropologists regard social learning theory as a universal learning mechanism for cultural knowledge and practices [27]. Therefore, social learning theory is commonly discussed as a key mechanism of intergenerational continuity of parenting behaviors [9,17,28]. However, the social cognitive framework [29], a later version of social learning theory, was most suitable for this study. In addition to incorporating all the learning mechanisms, such as observation, direct instruction, and modeling, the social cognitive theory emphasizes a person’s cognitive and decision-making aspects in learning and practices.

Although we started this study with a social cognitive framework, it was found inadequate to explain the data that emerged in the context of urbanization and modernization. Therefore, Wirth’s [30] theory of urbanization, and Inkeles and Smith’s [31] theory of modern man were added to the theoretical framework. The functionalist perspective of the theory of urbanization was used, which suggests that urbanization brings both positive and negative outcomes, and the residents respond to the effects with greater tolerance and creativity. The participants’ creativity in responding to the urbanization-induced insecurity challenges in Dhaka city was addressed with this theory of urbanization.

The lens of the theory of the modern man [31] was used during data analysis, which states that a modern person- 1) keeps themselves updated about major events and news, 2) believes that their actions might affect his life course, and that his active efforts can improve his as well as his family conditions. 3). is independent of the traditional authority sources, and 4) is open to novel experiences. The theory also suggests that a modern person uses their choice and decisions and does not ‘soak up’ everything modernization brings [p 82]. This perspective aligns well with the cognitive aspect of social cognitive theory [29]. The emphasis on a person’s free choice and decision-making skills of modernization theory also aligns well with the theory of urbanization [30], which emphasizes a person’s creativity in responding to the effect of urbanization.

This study focused on the parenting practices of the middle-class population of Dhaka, the capital of Bangladesh. The following matters were considered for choosing the middle class as the study population. The lives of the middle class generally center around their family; preparing their children for a bright future is their top priority. History suggests that the middle-class population played irrefutable roles in socio-cultural activities, intellectualism, political change, and leadership [32]. Besides, the middle class constitutes a large portion (25%) of the entire population in Bangladesh [32].

The study field, Dhaka, is densely populated, accommodating around 20 million people [33]. As the center of education, administration, and financial opportunities [34], this cosmopolitan city attracts around 30,000- 40,000 inland migrants every year [35]. It hosts people from various subcultures of Bengali culture. Consequently, we considered Dhaka an opportune setting to obtain enriched data on protective parenting.

In Dhaka city, huge-scale urbanization started in the 1990s when a large influx of inner migration and housing for the middle-class population started there. A second wave of urbanization took place in the 2000s with remarkable development in residential and commercial aspects. The urbanization process configured a new cultural identity for the migrants and tarnished the inherent nature of the community [34]. Sultana [36] reported that children in Dhaka city grew up in isolation inside homes with no interaction and not knowing the closest neighbors. This impacted their social bonding, social values, and responsibilities. Gradually, Dhaka has become a city of insecurity and crime. A rise in drug use, kidnapping, and violence against women was reported in 2006 [37], which is also valid to date [38]. The study also reported increasing loss of trust and loss of mutual understanding among the residents.

In the last two decades, the Shahbagh, Uttara, Mirpur, and Mohammadpur areas of Dhaka city have been primarily occupied by middle-class people. Among these areas, Mirpur was selected as the study location based on one primary condition: greater access to the study site. The first author grew up in Mirpur and was an insider by residence and lifestyle. This helped to achieve trust and access in the community for data collection. Besides, around 47 percent of the population in Mirpur belongs to the middle class [34]. Around 45% of the total income of Mirpur residents comes from fixed jobs [34]. This helps them plan for the future of their children and families.

A threshold of USD 2-3 per capita per day income is considered to be included in the middle class [39]. A survey on the emerging middle-income class in Bangladesh revealed that the monthly family expenditure of 85% of middle-class households was between BDT 30K to 99K, 15% had BDT 100K to 250K, and a negligible number of households had BDT 250K [40]. The family income stated in Table 1 indicates that G2 participants belonged to the middle class. The family income during children’s school-going stages of G1 participants was not collected because that would not explain their former socioeconomic status now. However, none of the participants reported or indicated poverty, financial hardship, or financial abundance while sharing their parenting experiences.

Ethics approval was obtained from the Human Research Ethics Committee (non-medical) of the University of the Witwatersrand, South Africa (protocol number: H16/06/38), and the Department of Clinical Psychology Ethics Committee, University of Dhaka, Bangladesh (project number: PH160603). For better communication with participants, the Bengali version of the participant information sheet and informed consent form were used. Besides, the first author, who was also the interviewer, explained the participant information sheet and informed consent form in Bengali for the participants’ better understanding. Written informed consent was collected from all but one G1 participant (Khadija), who gave verbal consent because of having no trust in written consent.

The ethics application stated that the interviews would be conducted in private. Yet, the plan changed following the pilot study when the researchers discovered a shared sense of confidentiality in the population as part of a communal culture. Additionally, conducting interviews closed door affected rapport and brought family members’ disappointments towards both the participants and the researcher. Participants also did not agree that closed-door privacy was necessary. Therefore, the following interviews were conducted keeping the door semi-closed. A semi-closed door communicated a sense of privacy and allowed interested family members to enter the room when deemed necessary.

Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the Supporting Information (S1 Checklist).

The parenting practices of two generations of Bangladeshi mothers were explored using semi-structured in-depth interviews of 22 purposively selected participants of 11 mothers (second generation; also referred to as G2) and 11 maternal grandmothers (first generation or G1). The participants were selected based on three criteria: a) the mothers (G2) with at least one grade 1 to grade 12 school-going child, b) the maternal grandmother (G1) be accessible and fit for an in-depth interview, and c) the participants with not developmentally or mentally challenged children. The grandmothers were the biological mothers or primary caregivers of the mothers. Participants were assigned pseudonyms, and the recorded interviews were translated, transliterated [41], and transcribed [42] into English. We emphasized cultural equivalence [43] in the process. NVivo 10 software was used for coding and data management.

Along with data collection, the first author occasionally wrote field notes and memos, which helped with data analysis and interpretations. The field notes were about their experience, the environment of the interview sessions, and relevant and noticeable information. The memo was about their biases, hunches, interpretations, hypotheses, and connections among the codes, ideas, and discussions. Table 1 depicts the participants’ socio-demographic background.

Inductive data analysis was conducted using a hermeneutic circle. UHJ repeatedly read each transcript and the field notes to understand the stories’ parts and wholes and generated initial codes. UHJ was attentive to the participants’ altered perceptions and interpretations while reading the texts. They applied axial coding to categorize the initial codes based on functional connectedness and perspectives. UHJ read and re-read the code categories, texts, research objectives, and research questions to generate broader themes and sub-themes and to interpret participants’ lived experiences. Initial themes were generated. Then, the initial themes were refined based on internal homogeneity and external heterogeneity [44]. The protective parenting theme was divided into four sub-themes: practices, beliefs, learning, and influencing elements. The data on each sub-theme were compared and contrasted between G1 and G2. This procedure was carried out multiple times till the manuscript was finalized.

MN verified the coding, themes, and analysis. UHJ’s knowledge and experience as an insider, as well as the literature on culture and traditions, helped contextualize and interpret the data. MN and BND checked for personal biases throughout. Accordingly, we were persistently aware of researcher biases and sensitive to the text’s ‘alterity’ [45, pp 268-273].

Data triangulation was achieved by collecting data from participant dyads of two consecutive generations. Their inputs supplemented and validated each other’s data without sharing information. The credibility of this study was achieved in three phases. During interviews, open-ended questions were used to avoid imposition of the researcher’s meaning on data. After translation and transcription, a language editor matched the Bengali interview audio with the English transcript. Finally, during the data analysis phase, the findings and interpretations were repeatedly checked against the raw data to maintain ‘referential adequacy’ [46, p 301]. Besides, descriptive notes were taken during the research process, member checking was conducted, and peer debriefing was also maintained from the formation of the questionnaire up to the end of writing this report. Confirmability of this study was achieved by an audit trail and inquiry audit, which were conducted by the second and third researchers, respectively.

The participating mothers (G2) acknowledged learning about chaperoning and saving money for their children’s future through observations of their mothers’ (G1) practices. Chaperonage was present in both generations’ parenting practices, but G2 mothers exhibited changes and variations shaped by evolving safety concerns in their social context.

Four mothers (G2) affirmed that they witnessed their mothers (G1) chaperone their (G2) younger siblings and identified the strength of the practice in revealing their younger siblings’ secrets and deceptions. Thus, chaperonage appeared to be a powerful tool for mothers in protecting children from derailment. For example, Piu (28-year-old mother, G2) described how chaperonage helped in stopping her sister’s deceptive behavior with their parents,

Thus, Piu vicariously learned about the usefulness of chaperonage and practiced it to prevent her children from being deceitful and to monitor their friend circles. Dali (50-year-old mother, G2) also reported a similar learning experience.

Parents’ responsibility to ensure their children’s financial security was also learned from observing their mothers’ practices of saving money for their children’s future. Jhumjhum (34-year-old mother, G2) shared that her parents (G1) spent money on her and her siblings’ education and weddings, and she was also saving for her only son’s (G3) education and wedding program, following her parents’ example. Jhumjhum said,

Charu (a 48-year-old mother, G2) also planned to finance her daughters’ weddings, as her mother Khadija (a 66-year-old grandmother, G1) did for her. Vicarious learning and observational learning played a critical role in adopting mothers’ practices of chaperonage and ensuring financial security for children, thus facilitating their continuity intergenerationally.

By comparing the reports of the G1 and G2 participants, practices such as training children to share their daily whereabouts, monitoring their social circle, and imposing gender boundaries were found common in both generations and often with slight modifications in the mother generation (G2). This indicates the carryover of certain practices from one’s own childhood experiences.

Dali’s (50-year-old mother, G2) mother, Hena (G1), trained her to report her whereabouts to parents. After becoming a mother, she trained her children to do the same and continued the practice until they were adults.

Developing a child’s practice of updating parents about their whereabouts appeared to be a form of monitoring children from a distance, and an attempt to control children’s movement. This practice was also an effort to generate self-control within their children by keeping them mindful of their accountability and of their parents’ constant monitoring.

All the participants from both generations valued gender boundaries unexceptionally for their capacity to protect girls from sexual abuse and harassment by limiting girls’ movement outside the home and by reducing girls’ exposure to men. Daughters were prohibited from talking to male relatives in private, visiting neighbors alone, or sleeping over without the mother. Fatima (60-year-old grandmother, G1) validated that restrictions to girls’ free movement were motivated by Islamic values. Fatima’s parents were strict religious practitioners and did not let her spend time with her male cousins alone. Although she was not a strict religious practitioner, she followed the same practice and did not allow her daughter Shirin (49-year-old mother, G2) to talk much with male cousins. Fatima said,

Schooling in co-education, Shirin’s (G2) daughter (G3) had more access to male classmates. Yet, Shirin maintained similar restrictions, did not allow their own daughter to interact with men or male cousins, and always interviewed if she ever talked to them. Shirin said,

Cell phones and internet access have become easy in the modern world, and that appears to have created anxiety among Bangladeshi parents about losing a grip over children and about children exceeding their boundaries. Therefore, the G2 participants had to extend monitoring of their child’s interaction with their opposite-gender relatives or friends. Although in some cultures, first cousins are considered close as siblings, and marriage is prohibited among them [47,48], this is not the case in Bangladeshi culture. This might have prompted the participants to protect their daughters from male cousins in the same way they would from any non-relative male. Each mother undertook different strategies to monitor their children’s friendships. Khadija (66-year-old grandmother, G1) and Jobaida (80-year-old grandmother, G1) used to host their children’s friends and establish connections with their parents. Khadija’s daughter Charu (48-year-old mother, G2) and Jobaida’s daughter Farida (46-year-old mother, G2) similarly hosted and quizzed their adolescent and young adult children’s friends about their family backgrounds, to assess their suitability as friends. Farida extended the monitoring strategy by befriending her children’s friends to control their activities indirectly because her children were not very obedient, and she wanted to safeguard them from negative influences. Farida said,

Both the G1 and G2 mothers’ concerns and efforts regarding children’s friendships reflect their conviction in a Bengali proverb that says, ‘A bad association can cause an iron rod to float,’ which means that a good friend positively influences lives, and a bad friend influences negatively. Childhood experience armed the participants with some protective parenting practices, and the G2 participants adapted or extended those practices based on their assessment of the current social safety context and their children’s characteristics to avert potential harm to their children.

Maternal grandmothers (G1) were found to have a considerable influence on teaching parenting practices. Mothers (G2) specified maternal (G1) guidance and direct instructions as a source of learning about protective parenting practices such as gender boundaries, restricting children’s movements, and traditional rituals.

Four mothers reported being explicitly instructed by their mothers (G1) to accompany their daughters when they visit friends. Grandmothers (G1) expanded their role by guiding mothers (G2) on safety measures for their grandchildren (G3). The safety measures included- not leaving daughters at home alone, restricting their children’s movements, and ensuring that children did not stay outdoors in the evening. Khadija (66-year-old grandmother, G1) advised her daughter Charu (48-year-old mother, G2),

The participating mothers (G2) mentioned that their mothers (G1) also educated them about the traditional beliefs, customs, and child protective rituals against evil spirits. The reported rituals were restrictions in attending to groceries soon after bringing them home, not bringing fish or meat home after evening, and cleansing fried and sweet food by whirling on the fire. Thus, the grandmothers’ generation (G1) played the role of responsible guardians. Kahi (57-year-old grandmother, G1) reported learning pregnancy-related customs and rituals from her mother and elders and instructed her daughter Jhumjhum (34-year-old mother, G2) to follow those during pregnancy,

The G2 mother, Jhumjhum, was highly educated, with a Master’s degree. Yet, she strongly believed in traditions and chose not to give up tradition- and culture-supported practices in child protection. Both generations of participants expressed honor and trust in their mothers and identified them as important agents in their process of learning protective parenting. Sometimes, the mothers (G1) modeled those practices, and sometimes, they instructed their daughters (G2) before and after childbirth.

Participants mentioned other learning opportunities besides learning from their mothers. G2 participants sometimes received advice and directions from fathers and husbands about escorting children to school. Primarily driven by safety concerns, family members supported and encouraged protective parenting practices. Elders in the community, local proverbs, and urban legends reinforced practices such as not praising children, selecting friends from good families, and not giving money to young children.

Besides the direct instructions from the grandmothers (G1), the significance of chaperonage children was emphasized by the mothers’ peer groups, media reports, and other family members. Sometimes, children’s behavior and actions were reported to influence protective parenting. Shoshi (47-year-old mother, G2) and Farida (46-year-old mother, G2) sometimes found their sons lying and engaging in culturally disapproved behavior, and they had to be creative to extend parental protection. Shoshi explained, “I called to check whether he (son) went to the coaching center truly… [I become sure of it] by listening to the teacher’s voice and the children’s voice. Boys are not trustworthy, (.) understand?”

Children’s ‘inappropriate’ conduct thus prompted the adoption of extended monitoring methods, including frequent phone calls, befriending children and their friends to control children’s whereabouts, and having access to daily updates. Extended monitoring strategies have been reported to bring some positive changes in children’s conduct. However, the extent of mutual influence cannot be reported here as this study did not explore children’s accounts.

Access to global media exposed children to socially unacceptable content, resulting in the development of new protective parenting strategies, which were absent and irrelevant in G1 participants’ childrearing time. Tonu (37-year-old mother, G2) explained how technology has grounded extended monitoring strategy,

Participants had heightened concern about children’s media access and that called for their creativity to put a balance between access to modern technology and preserving local traditions and cultural norms. Mothers (G2) also engaged in open conversations with children and were vigilant of warnings of substance abuse, while the mothers (G2) reported an increase in the availability and misuse of substances in recent years.

Thus, the intergenerational continuity of protective parenting practices occurred through an interplay of direct instruction, childhood experience, observations of mothers’ practices, and other learning opportunities in the context of urbanization and modernization in a traditional communal society. Based on these findings, we propose a process model of intergenerational continuity of protective parenting practices in Fig 1.

According to this intergenerational continuity process model, learning of protective parenting practices starts at home within a complex interaction with social endeavors. A person carries their childhood experience of being parented. Besides, while growing up, they observed their parents’ protective practices and parents carrying protective values. After becoming a parent, the person often receives various parenting advice and instructions, mostly from their mother, and also from a partner, parents-in-law, and extended family members. As a member of society, a parent interacts with cultural and traditional knowledge, practices, and beliefs, urban legends, peer groups, seniors, community members, emerging technological and global changes, and various aspects of the community, including security. A person continuously internalizes knowledge and practices from all these resources. When they raise their children, they project their knowledge onto their individual situations. Using their rational thinking, openness, knowledge, and traditional, emotional, and moral values, a parent then chooses which parenting practices to implement with their children and which ones they would modify or alter according to current needs and situational demands. Through this process, a parent develops and determines the parenting practices they would implement in raising their children. Children’s behavior and characteristics remain as live modifiers in the process, which explains why a parent’s parenting practices might differ for two children in the same household.

The inclusion of maternal grandmother and mother dyads enabled triangulation of data. However, the inclusion of maternal grandmothers in the sample might have created a bias in the findings, as maternal grandmothers were mostly mentioned. The G2 mothers were cognizant that their mothers were being interviewed, and this may have primed their responses to be about interactions with the maternal figure.

The findings presented here originate from a study centered on general parenting practices of Bangladeshi mothers, and protective parenting aspects did not receive a focused exploration. Besides, geographical constraints might have a culturally confined view of aspects of protective parenting. A dedicated study conducted in different geographical regions in Bangladesh has the potential to enrich and broaden the applicability of the findings.