Leigh Stefani
Theresa Hwang
Publication Date


"CARE has found that there are persistent gender inequitable behaviors that are not changing despite working on attitudes and providing information, and where good policies are in place. CARE sought out to test whether a deeper understanding of social norms could shed new light on what is holding certain behaviors in place, and lead to more effective strategies for transforming gender norms and behaviors that seem stuck."

CARE has been working to develop a deeper understanding of social norms and how they can shed new light on what is holding certain behaviours in place. Beginning in 2014, a small team across CARE came together to develop and pilot new measures for social norms through an iterative learning process across 3 pilot sites in Sri Lanka and Ethiopia. This process included an initial training on social norms theory and measurement from researchers from the University of Pennsylvania Social Norms Group (UPenn SoNG). This report shares experience and learning on translating social norms theory into practical measurement tools for development programming, featuring: the use of CARE's Social Norms Analysis Plot (SNAP), a framework developed to measure if and how norms are changing; qualitative vignettes; and survey questions. By adapting social norms theory into practical design and measurement tools, they hope to better equip practitioners working to address social norms within international development programmes.

As outlined here, social norms are

  • behavioural rules constructed and shared by a group, and are different from individually held beliefs or attitudes;
  • made up by one's beliefs about what others do ("empirical expectations", or EE) and by one's beliefs about what others think one should do ("normative expectations", or NE);
  • maintained by social influence - that is, by the anticipation of social approval or disapproval for one's actions, also called positive or negative social sanctions, respectively; and
  • involve our "reference group" or "reference network", who are the relevant other people who matter to us.

The application of social norms theory proposes a divergence from traditional behaviour change interventions, which seek to change individuals' personal attitudes and knowledge and consequently their behaviours, usually through awareness raising and information sharing. Social norms, however, as the rules of behaviour within a group, may be blocking this pathway of change. Individuals may personally disagree with a social norm, but act in accordance with it out of a desire for social belonging and to avoid social backlash. Thus, CARE stresses that attitudes and knowledge are still important factors to address for behaviour change, but addressing social norms as well may be a key piece in transforming some of the more intractable behaviours - especially those that are kept in place by gender norms.

Noting a dearth of existing work on measuring and monitoring shifts in social norms in a way that more closely follows social norms theory, CARE developed and tested out new measures for social norms based specifically on Bicchieri's synthesis. This included starting from Bicchieri's assumption that social norms are held in place by both empirical and normative expectations (EE and NE), and that the presence of both these expectations together indicate the presence of a social norm. CARE developed and piloted a combination of quantitative and qualitative tools and processes to explore ways to answer these questions within a standard implementation process. The team paid special attention to ease of use of tools for resource-constraint settings, so chose to adapt already well-known and used methodologies (e.g., surveys, focus group discussions, or FGDs). Table 1 on page 5 outlines the purpose, methods, and learning aims for each stage of measurement during implementation.

The case studies that form the backdrop of this report has been an iterative piloting and learning process, as follows:

  1. ReNEW (Redefining Norms to Empower Women), focused on engaging men and boys to reduce intimate partner violence (IPV) on tea plantations in Sri Lanka, funded by Johnson & Johnson (J&J) (2014-2016);
  2. TESFA (Towards Improved Economic and Sexual Reproductive Health Outcomes for Adolescent Girls), focused on the needs of ever-married adolescent girls in the Amhara region of Ethiopia, also funded by J&J (2015-2017); and
  3. Abdiboru (Improving Adolescent Reproductive Health and Nutrition through Structural Solutions), an operations research intervention focused on reducing early marriage and improving health and nutrition outcomes for young adolescent girls in the Oromia region of Ethiopia, funded by the Bill and Melinda Gates Foundation (BMGF) (2015-2020).

During the formative phase, the following key pieces of information were identified: What, if any, social norms are at play for a specific behaviour in question? Who are the most influential reference groups for the specific norm? What social sanctions are anticipated for deviating from the norm? For example, in the Abdiboru project, primary data was collected over 4 days using a brief set of semi-structured FDGs and key informant interviews (KIIs) with a cross-section of community members. A sample question asked: What would others say about parents whose daughters are not married by [insert ideal age as identified in prior questioning]? How does this affect the parents? Whose opinion would matter most? The additional data in the formative stage that emerged proved a useful and worthwhile step in Abdiboru, as it uncovered some important surprises and nuances that impacted the design of the baseline tools. For instance, FGDs revealed that girls themselves could also be a driving force in the decision to marry early, and were influenced to do so by their peers. The implementation strategy may have otherwise missed this important reference group because traditional thinking was that adults were the main driving force, not girls themselves.

Tools at baseline and endline involved quantitative surveys, which were used to measure personal attitudes, knowledge, behaviour, and social expectations, using response scales to try to better track incremental changes over time. Social norms questions were added to existing knowledge, attitudes, and practices (KAP) surveys by adding sections with prompts that asked about others' behaviours and attitudes (EE and NE). Normative expectations were measured in surveys by either asking respondents what behaviour they think others would approve of, or whether respondents expect any negative social sanctions to take place in response to behaviour that deviates from the norm under investigation. Answers to both types of questions reveal whether there is a normative expectation to behave in a certain way. Data were also analysed to compare EE to actual behaviour, and NE to personal attitudes to reveal whether people hold incorrect assumptions about what others do and think ("pluralistic ignorance").

In addition, CARE used qualitative methods. Vignettes tell short stories about imaginary characters in specific contexts, with guiding questions that invite people to respond to the story in a structured way. This helps to diagnose whether a norm exists. CARE used vignettes at project baseline to understand how norms were influencing behaviour and to identify weaknesses in the norms that could be addressed by programme activities. At endline, CARE used vignettes to identify and track signs of norm change. For TESFA, CARE developed an analysis framework for vignette data (the SNAP - see tables 4 and 5 on pages 12 and 15), and in the span of one week CARE staff designed vignettes, trained researchers, and piloted the vignettes. The aim of the approach was to be light and quick, and to develop and test the use of vignettes for the first time. While this data were being analysed, CARE built on the promising early results of this experience and developed vignettes as part of the endline evaluation for the ReNEW project, using the SNAP again as a guide. The SNAP framework has also been used to guide analysis of potential signs that norms might be weakening or shifting, or if not, what factors in particular seem to be holding norms in place. CARE pilot efforts to date have begun experimenting with practical ways to integrate the SNAP framework into monitoring efforts, in a way that does not overburden staff. For example, for public activities that facilitate critical discussion on harmful norms, observation prompts of group dynamics and reactions can be added to monitoring forms, such as: If participants' voice resistance or support to ideas that challenge the norm, how do (most) other participants react?

Prior to the report's annexes, CARE reflects on the key insights and practical challenges that emerged during the process of adapting and testing this approach across the 3 pilot settings, as well as some remaining questions:

  • Some insights: More iteration and learning is needed to understand when it is necessary to collect primary data in the formative stage to identify social norms; so far, this measurement approach has come with high demands on staff capacity and time; staff reflection on their own beliefs and values is important to be able to identify and problematise gendered norms that are acting as barriers to health and development outcomes; and a focus on social norms seems particularly appropriate in cases where programmes seek to change intractable behaviours held in place by gender norms and shielded from scrutiny by taboo.
  • Some challenges: Some of the terminology in Bicchieri's theory was hard to translate, and could be alienating to practitioners; there are some concerns about the hypothetical nature of survey questions that ask about what others "generally" do or theoretically would do under certain circumstances; and questions about EE and NE increased the length of already-long surveys, which can be especially challenging for multi-sectoral projects.
  • Some remaining questions: Is measuring EE and NE worth the extra investment to understand influences on behaviour? Are other qualitative methods, such as Most Significant Change, direct observation, or individual interviews, more appropriate to detect the emergence of entirely new norms? If working to replace a norm about marrying early with delaying marriage, could we go further to create an environment where it is also acceptable for individuals to choose not to marry at all? A diversity of norms (barring those supporting harmful behaviour) may be a more sustainable and just path forward for social change.

"Applying Theory to Practice: CARE's Journey Piloting Social Norms Measures for Gender Programming. Copyright 2017 Cooperative for Assistance and Relief Everywhere, Inc. (CARE). Used by Permission." Source: Girls Not Brides website and CARE website, both accessed on August 15 2017. Image credit: © 2016 Josh Estey/CARE