Blog
Why developing more measures of social and gender norms really matters for gender equality
03 Mar 2021

A young girl in Jordan.
Supriya Sthapit recently completed an internship at UNICEF Office of Research – Innocenti. In this blog, Supriya reflects on her work reviewing existing gender equality measures on social and gender norms, and how they can be used to strengthen research and evaluation studies, including on gender equality and social protection.
In August 2020, I joined the Gender-Responsive and Age-Sensitive Social Protection (GRASSP) Team at UNICEF Office of Research – Innocenti as a research intern. Working with my colleagues Elena Camilletti and Zahrah Nesbitt-Ahmed, we mapped and reviewed available measures of gender equality outcomes, including ones on social and gender norms across the life course used or tested in research and evaluations in low- and middle-income countries.Why measure social and gender norms?
In recent years, there has been growing interest in social and gender norms (Box 1 provides definitions), in part due to their role in shaping gender equality outcomes. Depending on how restrictive or progressive these norms are, they can either foster positive social transformation or hinder progress.
Methodology
Using existing repositories of measures from EMERGE and Population Council (Gender and Power Metrics), as well as additional surveys and evaluations[1], the measures were mapped and assessed to identify whether they:- align with the underlying theories on empowerment and norms
- cover a range of thematic areas in the GRASSP conceptual framework, such as health
- cover different life course stages from childhood to old age, and
- cover different countries and regions.
while journal articles and research reports claim to measure norms, they in fact measure attitudes
Some Preliminary Findings
Most of the existing measures of gender equality are on attitudes - only few focus on social and gender norms. So far, 419 measures of gender equality have been reviewed. Of these, only 31 were found to be on social and gender norms. This is significantly lower than 226 measures found on attitudes and perceptions. Upon further review, while journal articles and research reports claim to measure norms, they in fact measure attitudes. This is in line with findings from ALIGN on gender norms, and suggests the need for a more careful conceptualization and operationalization of the concepts of norms and attitudes.Only a few of the thirty-one measures of social and gender norms are reported to be reliable. Of the twenty-three measures on norms that could be tested for reliability, nine of these report at least one measure of reliability, such as internal consistency. Of those, only seven had Cronbach alphas of more than 0.80. For the remaining 14 measures, journal articles reviewed did not report information around their reliability. This calls for greater attention to validity and reliability, and rigor, by researchers when developing or employing measures in research and evaluations.Most of the thirty-one measures on social and gender norms focus on empirical expectations. Twenty-one of these measures on norms focus on empirical expectations (see Graph 1 below for the breakdown on the number of questionnaire items). For instance, a 2019 paper drawing on findings from GAGE Bangladesh and Ethiopia baseline surveys includes a measure on “Community-Level Restrictive Gender Norms”. This assesses empirical expectations related to gender and education by asking respondents to agree or disagree with statements such as "Adolescent girls in my community are more likely to be out of school than adolescent boys".A further 13 measures are on normative expectations. For example, a South Africa survey contains a measure on "Norms about Partner Violence Scale", asking respondents to agree or disagree with the statement "Most people in community think a boy can assault a girl".Finally, two measures were found to assess what actions are taken if individuals transgress prevailing norms in their communities (sanctions). Additionally, one measure was found on the nature and influence of the reference group (who and how individuals who matter to a person influence his/her decisions).
