Using Longitudinal Social Network Analysis to Evaluate a Community-Wide Parenting Intervention

"Norms are shared attitudes and behaviors expressing beliefs about the appropriate conduct of individuals; they are strong guides for a range of social behaviors including parenting..."

Violence prevention is a public health priority in low- and middle-income countries (LMICs). Harnessing naturally occurring social influence processes may extend the reach of traditional parenting skills training programmes that are designed to prevent violence. This paper presents a social norms intervention implemented in a deprived community in South Africa to prevent harsh parenting attitudes and behaviours that negatively impact children's development. The researchers use social network analysis (SNA) to explore how the benefits from the intervention components diffuse through the community.

This study took place in Touwsranten, a rural, low-income community in the Western Cape Province of South Africa that is comprised of two language groups: a larger mixed-race Afrikaans-speaking community (n=2,218, 78.6%) and a smaller isiXhosa-speaking group (n=604, 21.4%). In spatial terms, the two groups live in distinct sections of the community and constitute two separate social networks.

The intervention consisted of two components aimed at shifting harsh parenting practices to more positive parenting practices across the entire community:

• A community development process that used social activation methods to mobilise the community around parenting - A community-based and community-developed "brand" for positive parenting was developed, alongside a number of child-oriented community activities to support related prosocial values. Many households displayed stickers with the logo on the doors of their homes and signed the "Saamstaan" (Standing together) manifesto describing values related to change in Touwsranten.
• Four age-specific Parenting for Lifelong Health programmes rolled out after the mobilisation was initiated - They consisted of: (i) Thula Sana, a home visiting programme starting during pregnancy to improve attachment between mother and child; (ii) a cognitive development book sharing programme for toddlers; (iii) the Sinovuyo Caring Families programme for children aged 2-9, which is designed to increase positive parenting and reduce harsh parenting and child behaviour problems; and (iv) the Sinovuyo Caring Families programme for parents and teens, which aims to reduce violent discipline and teen aggression.

The mechanism (theory) of change for the intervention was conceived in terms of diffusion through a social network: Caregivers who attended the parenting skills training programmes are expected to tell their friends and neighbours what they had learned, and so spread the positive parenting concepts. The community mobilisation process would facilitate this diffusion by helping non-attendees become more receptive to the new concepts and by reinforcing concepts once they had been acquired.

Afrikaans-speaking female caregivers (n=235; mean age 35.92 years) with children between 1.5 and 18 years old were included in the study; two waves of data were collected (January-April 2016 and June-October 2017). To detect the caregivers' social network, they were asked to nominate up to five female caregivers (with children under 18 years old) whom they "talk to about parenting in the community of Touwsranten". To account for the fact that parenting behaviour differs as a function of children's age, two age-specific questionnaires were administered at both waves. Both questionnaires assessed positive (e.g., warm, consistent) parenting skills.

Of the caregivers in the analytic sample, 98.7% (n=232) reported being aware of and receiving the community mobilisation component: 21.9% were attendees. The survey results indicated that around a fifth (25.1%; 42.3% of attendees; 21.4% of non-attendees) had participated in group activities and signed the manifesto.

The results indicated community-wide increases in positive parenting behaviour (involvement, supervision, consistent discipline, and reduced corporal punishment). There was, however, no evidence that attending a programme resulted in greater increases in positive parenting scores; nor was there evidence that attendees had systematically different parenting scores from non-attendees.

The researchers then looked to see how the network changed during the course of the intervention. There was an increase in network density from Wave 1 to Wave 2, and caregivers reported communicating about parenting with more network members. Attending at least one session of the parenting skills training programme (n=51; 21.7%) significantly predicted increases in network centrality. Attending sessions of the parenting skills training programme was associated with an increased likelihood of both forming new network ties or maintaining old ties and receiving new network ties, suggesting that attendees may have had more opportunities to influence other caregivers in the network. In that vein, caregivers appeared to use similar parenting behaviour to other caregivers they were connected to within the network, especially when those others attended a parenting skills training programme.

"Collectively, these results indicate that caregivers who attended parenting programs became more connected within the caregiver network than did non-attendees, and caregivers influenced the behavior of those to whom they were connected - particularly, if they attended a parenting skills training program. Therefore, due to the increased likelihood of attendees speaking about parenting, and the significant socialization effect, it appears that on average attendees may be influencing the behavior of other network members more than non-attendees do."

In conclusion, the results "suggest a means (community mobilization) and a process (norm change via social networks) for amplifying the effects of individually oriented parenting skills training programs, which may be more cost effective than simply delivering stand-alone programs." Per the researchers, the study also illustrates the value of SNA for ascertaining the processes by which the intervention achieved its impact.