Background

Ample evidence has demonstrated that biology, social power and social experience combine within a gender system to create health-related inequities. This gender system is held together by gender norms, the implicit rules that govern the attributes and behaviours that are valued and considered acceptable for men, women and gender minorities in various social settings and at various stages of their lives.1 Though they are contextual and amenable to change, gender norms are embedded in our institutions, laws and policies. They interact with other axes of advantage and oppression to reproduce and reinforce hierarchies of power and privilege, which determine an individual’s social position and, thus, their health throughout life.

Recognition of the strong connections between gender inequality and health has driven global commitments, including the 2030 Agenda for Sustainable Development, to increase attention towards ‘gender-transformative programs’.2 These programmes seek to critically examine and transform gender dynamics and norms as a means to improve health and achieve gender equality.3 Despite their intentions, however, these programmes are often not designed or evaluated along the dimensions that are fundamental to achieving and measuring sustainable gender norm change and gender equality. Rather than focusing on systems and social structures that apportion power and privilege, they often define and deconstruct gender norms so that they can be addressed and measured at the individual level.4–6 In our view, these efforts come up short in reaching their optimal health and development impact and in achieving and measuring long-term gender equality outcomes.

A focus on the individual level reflects a more atomised and psychological approach to understanding gender norm change, diverting attention and responsibility from the underlying social structures and systems that contribute to gender inequality. A failure to focus on these root causes also results in a failure to address the discriminatory policies, laws and institutional practices that lead to other intersecting forms of oppression.7 As one article argues, an overfocus on norms, as they are currently conceptualised, has prevented the field from addressing a host of other ‘institutional, material, individual and social factors’ that keep inequalities in place.8

Further evidence of this overemphasis on the individual lies in the fact that programme success is centred not on gender equality, but rather on the development outcome of interest. There is little incentive, therefore, to go beyond the individual and attempt to measure broader norm change. Even if that incentive were there, aid programming does not allocate sufficient time or funding to track the impact on norms. The process of identifying and addressing gender-related barriers is often a ‘must-do’ in the short term, but attaining and proving gender-transformative outcomes is only seen as a ‘nice-to-have’ in the longer term.

Our effort to address these issues has led us to develop an analytical framework for assessing programmes that attempt to achieve gender equality and transform restrictive gender norms considering their stated commitment to social change. This framework derives from our earlier systematic review, which identified rigorously evaluated ‘gender-transformative’ health programmes published between January 2000 and November 2017 that measured potential changes in health-related outcomes, at least, and both health and gender-related outcomes, at best (87 evaluations of 85 programmes).5

The framework offered here contributes to the field by expanding our understanding of how programmes engage pervasive systems of gender inequality, and not just the manifestation of gender inequality in the lives of individual programme beneficiaries. It also highlights the need for greater ambition in how programmes are designed, implemented and evaluated. We believe that this framework will help structure gender-transformative programmes so that they are more effective at attaining gender equality impact and will support more ambitious evaluations for capturing norm change.

Theory: the link between gender norms and health

Although gender norms are socially constructed and can shift over time and space, heteronormative cultures tend to present male masculinity and female femininity as fixed, timeless, natural and inevitable. They deeply disfavour individuals who fail to conform to culturally sanctioned gender categories, often rendering alternatives to these rules as impossible or aberrant, a contravention of ‘Nature’s way’. Essentially, gender norms serve as the mechanism through which gender-biased ideologies, relationships and social structures persist and are reinforced.6

Research tells us that combined with other hierarchies like race and class, gender norms are a part of an elaborate system of rewards, punishments, expectations and constraints that deeply affect the health of individuals and entire communities.1 9 This system contributes to determining health outcomes independent of the expertise of health practitioners or the availability of healthcare itself, resulting in health disparities between groups of different genders, and exacerbating power differentials and injustice. Norms are sufficiently powerful influences on individual thought and behaviour that they can drive individuals to act in ways that conflict with their personal preferences and result in harmful outcomes.8

Analysing gender norms helps us understand why people engage in unhealthy behaviours, including different forms of violence, risk taking and avoidance of healthcare, even when they know they are endangering their own (or others’) health. For instance, young, gay men may engage in risky sex to reaffirm their masculinity, even though they know the dangers inherent in such behaviour.10 11 Similarly, women engage in potentially harmful body-altering procedures as a consequence of cultural messages about the inadequacy of their bodies and the need to improve themselves to be attractive to men; these messages have created a massive market for these procedures in diverse cultural settings that nonetheless share a patriarchal gender hierarchy.1

Are gender-transformative programmes really changing gender norms?

Given the evidence linking gender norms and health, many donor and civil society organisations have committed to implementing ‘gender-transformative’ programmes.12 These programmes understand health as a dynamic social phenomenon, influenced by a broader system of feedback related to social and structural determinants of health.5 They attempt to shift the unequal power relations and harmful systems that uphold them by focusing on two kinds of gender norms that contribute to this system: those that are restrictive, as well as those that are ‘opportunistic’ and provide space for positive influence.8

According to the Interagency Gender Working Group of the United States Agency for International Development (USAID), ‘Gender-transformative approaches encourage critical awareness among men and women of gender roles and norms; promote the position of women; challenge the distribution of resources and allocation of duties between men and women; and/or address the power relationships between women and others in the community. (p8)’13 As others have elaborated, these program approaches ‘aim to go beyond individual self-improvement among women toward transforming power dynamics and structures that act to reinforce gendered inequalities.(p10)’14

The rationale for gender-transformative programming is well established in theory, and growing data show that these programmes do, in fact, lead to improved gender-related and health-related outcomes. That said, the evidence of their impact on sustained, longer term norm change and gender equality is lacking. Operationalisation and measurement of the pathways to change are complex and dynamic and are influenced by multiple factors. As one study has stated, ‘To make sense of these complexities, organizations often operate around theories of change, which are sets of hypotheses (best guesses) on how change happens’ (Hillenbrand et al14 p 12). How do we know, then, whether and how our programmes are transforming gender norms and improving health?

Regarding ‘whether’ and to what extent gender norms are actually transformed, it is difficult to say. According to the 87 evaluations of 85 programmes included in our systematic review,5 programmes were not always evaluated in ways that could prove that norm change had occurred. First, we found that donor constraints and time frames often precluded assessment over a period long enough to measure social change; only five (6%) examined impact 3 years or more after programme completion. Second, despite implementing activities that were intentionally designed to address restrictive gender norms, approximately 40% of the evaluations included in our analysis either did not measure gender-related outcomes at all or did so using qualitative methods alone. Evaluations that did assess gender-related outcomes (n=49, 56%) most frequently focused on activities that were implemented to improve internal locus of control, measuring self-reported outcomes such as self-confidence, ability to negotiate for oneself, freedom of movement and age at marriage. Changes in gender-related attitudes—which are often conflated with norm change8—were measured in roughly half of the programmes; 11 (13%) programmes measured changes in knowledge (eg, related to laws and policies, rights and entitlements, access to services, etc); and five measured educational attainment, a strong predictor of gender equality as a means of improving health.5

As for the ‘how?’, the findings of our earlier analysis5 indicate that most gender-transformative programmes are not designed along the dimensions that would be fundamental to achieving lasting gender norm change and gender equality. Among the 85 programmes, nearly all implemented activities focused on the individual and/or interpersonal levels, rather than the system as a whole. For example, most of the programmes (n=83, 98%) implemented education or awareness-building activities among individual beneficiaries to foster critical awareness of existing norms and inequalities and to discuss how one’s life might be advantaged and/or limited by them. Generally, these programmes also implemented activities to foster equitable interpersonal relationships and social integration through diverse activities including sports, life skills training and mentorship. 74 programmes (87%) focused on community-level activities such as local health fairs, community drama presentations and the like. Though community-level activities are an important approach to norm change, when done alone they do not ensure the transformation of structures and policies that enable long-term and large-scale changes in power relations.

Framework

Given the barriers to designing, implementing and measuring gender-transformative programming, we offer a practical and simple framework that builds on norm change theory and is substantiated by an analysis of the literature included in our prior systematic literature review.5 We propose that programmes are more likely to transform gender norms and increase gender equality if they demonstrate a multiplicative effect, sustainability, spread and scalability.

These concepts arose from an understanding of the glaring gap our analyses had pointed out. As we have highlighted here, the inadequacy of norm change evaluations logically calls attention to the need to assess impact beyond specific beneficiaries, include broader time frames and measure change at scale. To start addressing these shortcomings we drew first from the literature on systems thinking, and its associated concepts of complex, dynamic and interconnected social connections that foster sustainable outcomes.15 Definitions of gender-transformative programming also provided inspiration in their calls for programmes to challenge social hierarchy; and again, logically, if gender-transformative programming reflects these elements, then they need to affect the power dynamics that shape people’s lives and not just their individual experience. Finally, if gender-transformative programming occurs ‘upstream’ from specific outcomes, as others have described it,16 it should have an impact on multiple outcomes. We built on this earlier research and the understanding that gender norms reflect and exert influence on a complex, dynamic system of hierarchy and inequality to construct our framework of four essential dimensions of norm change: multiplicative effect, sustainability, spread and scalability.

The framework is presented in table 1 with definitions for the four dimensions and the criteria for each. To achieve ‘measured high-quality’ status (ie, programmes that are highly likely to transform—and/or have transformed—norms and improve gender equality), a programme must explicitly measure at least one criterion of each dimension in its evaluation. (See box 1 for a case study describing the measured high-quality programme, SASA!.) Programmes that are designed along the four dimensions of our framework yet fail to measure all four of the dimensions explicitly, would be described as having ‘inferred high quality’ (ie, we can only infer that the programme is gender transformative). To achieve an inferred high-quality rating, a programme must at least reference each of the four dimensions in its design or description of programme implementation, even if the measurement of these dimensions is not reported in its evaluation. Online supplemental table A lists the programmes from our systematic review that met measured high-quality status (n=16, 19%) and inferred high-quality status (n=9, 11%), arrayed according to the dimensions of the framework.

The first dimension of our framework captures the multiplicative effect of gender-transformative programmes. The interconnected nature of a system means that as one element in the system shifts another will also be affected, and so on. Changing those elements changes the whole system. Gender-transformative programming works to change systems of norms that are not related solely to specific health outcomes of interest but reflect a broader set of values and cultural beliefs that influence any number of outcomes. Therefore, we posit that if gender-transformative programmes are working on the system, then there should be downstream ripple effects beyond, and in conjunction with, the specific health-related outcomes of interest.

The ’Strengthening Household Ability to Respond to Development Opportunities’ (SHOUHARDO) project in Bangladesh, for example, demonstrated improvements across a wide range of outcomes related to health, well-being and agency.17 In its design and evaluation, the programme recognised the interdependent relationship between the disenfranchisement of women and those in poverty and numerous outcomes, including the nutritional status of children, stunting, food access, quality of household diet, sanitation, empowerment of women and the poor and the threat of natural disaster. Programme activities included women’s participation in Empowerment, Knowledge and Transformative Action, village development committees and other SHOUHARDO groups, which increased women’s contributions to household decisions, freedom of movement, cash income and reported changes in patriarchal attitudes. SHOUHARDO project evaluators noted that the complementarity and diversity of interventions led to a significant improvement in a broad range of health and development outcomes.18

The second dimension of our framework is sustainability. Within a system, reinforcing and balancing feedback loops perpetuate the system. As has been documented, feedback loops are powerful in part because their effects accumulate as movement across the system occurs.15 Hence, once the structure is changed, then the elements within that structure are less likely to revert to the way they were before.

Not many programmes are achieving this structural change, but there are a few. SASA! is a remarkable programme that worked first in Uganda to reduce gender-based violence and violence against children through a broad, community-based intervention.19 SASA! acknowledges that intimate partner violence and HIV result from a complex interaction of factors at all levels of society. To accumulate and reinforce social capital and influence for long-term change, the intervention engages a variety of stakeholders including the public, community activists, local community leaders and professionals. Furthermore, SASA! implements activities that reach each level of influence: individual, interpersonal, community and society. Five years after SASA! began, a cluster randomised controlled trial demonstrated positive sustained effects. Because of its success, SASA! has since been adapted for implementation in over 30 countries by more than 75 organisations.20

SASA! is sustained by its community mobilisation approach, which recognises that even if an individual makes or tries to make a change, it is very difficult for that individual to maintain that change unless supported by the people and environment that surround her/him. This brings us to the third dimension of our framework, spread. Theorists suggest that a person’s preferences and behaviours are influenced by their ‘reference group’, which is loosely defined as the people whose opinions matter most to them.8 Therefore, to change a gender norm, it is critical to engage the entire network of people who share the norm in question.21 22 As with the ripple effect described above, it makes sense then that a gender-transformative programme will demonstrate evidence that gender-related outcomes have directly or indirectly passed to individuals outside the intervention. Over time, this ‘spread’ will gain momentum and diffuse through the population and/or social system.23 Somos Diferentes, Somos Iguales, a multifaceted intervention implemented across much of Nicaragua, serves as a good illustration of this spread effect.24 This communication for social change strategy was implemented over a 3-year period and included a peer-to-peer youth leadership training, among other important activities. By making use of social networks, Somos Diferentes, Somos Iguales influenced their intended target population’s broader reference groups and web of contacts.

Finally, by definition, a social change agenda must occur at scale,25 so gender-transformative programmes should have been or can be taken to scale, the fourth dimension of our framework. While spread is the actual transference of programme influence from beneficiaries to their networks, scale focuses on increasing the number of beneficiaries by increasing the scope of the programme. Ultimately, scale contributes to spread.

Scalability often requires thinking from the beginning about longer term programme expansion. One programme that exemplifies built-in scalability is The Fourth R: Skills for Youth Relationships, a school curriculum developed in Ontario, Canada.26 The researchers argue that beyond its impactful implementation, The Fourth R has the potential for widespread dissemination. Since the intervention’s 21-lesson curriculum meets Ontario’s mandated educational requirements, it was possible to integrate it into schools’ existing health and physical education classes. The curriculum included all that was needed to implement the intervention: detailed lesson plans, rubrics, videos, handouts and role-play exercises. Apart from a 6-hour training workshop on the curriculum, no additional time outside of school hours was required of either teachers or students. It can easily be taken up in other schools or across the province.