Mayra Buvinic (United Nations Foundation) served as moderator for this session. She explained that the concept of women’s empowerment gained traction during the early 2000s with the work of Amartya Sen and then Naila Kabeer. According to Buvinic, theoretical work focused on concepts related to empowerment dominated the field for more than a decade due to the intrinsic difficulty of measuring empowerment. Buvinic said that serious efforts to empirically measure women’s empowerment—particularly economic empowerment—have emerged over the past 5 to 6 years.
THE LITERATURE ON THE RELATIONSHIP BETWEEN WOMEN’S EMPOWERMENT AND FERTILITY
In 2014, Ushma Upadhyay (University of California, San Francisco) published a review of the literature on the relationship between women’s empowerment and fertility.1 Among the 60 articles she reviewed, women’s empowerment had been measured by 120 different proxy conceptualizations. Upadhyay acknowledged that there are different ways to operationalize the concept of empowerment for research purposes. Researchers must first determine the level of empowerment that they wish to measure: individual, interpersonal, or societal (local/community, regional, or national). On the individual level, she said, measurements usually take the form of indices, such as gender attitudes, that relate to the individual. Interpersonal measurements can quantify a woman’s power in relation to others—such as her partner or her health care provider—at a very nuanced level; interpersonal measurements can also characterize family support structures and interspousal differences. At the societal level, empowerment measures can capture phenomena such as rates of voting or child marriage, and in multilevel analyses they can do so alongside individual or interpersonal measures. Upadhyay pointed to a need to perform more varieties of measurements in multiple contexts to interrogate how contextual changes influence individual power.
Empowerment measures must capture a complex process, Upadhyay noted, one that includes bargaining, negotiating, and making choices in the face of resistance. While most studies utilize single-timepoint indicators of empowerment, longitudinal studies encounter the logistical challenge of collecting these indicators over time.
Any selected measure of empowerment also must capture actual choices. Upadhyay emphasized that many people internalize social norms in ways that constrain their choices or face constraints based on available options. For example, evoking the case of activist Malala Yousafzai, Upadhyay argued
___________________
1 U.D. Upadhyay, J.D. Gipson, M. Withers, et al., Women’s empowerment and fertility: A review of the literature, Social Science & Medicine 115, 111–120.
that measuring empowerment through education may not be appropriate in cases where education entails a threat of extreme violence that renders it an extremely difficult choice to make. Empowerment is therefore most clearly demonstrated by a relative exercise of power within a gender-based system of expectations, and can be measured by relative levels of adherence to gender norms. Measures must also reflect the appropriate domain of empowerment for a given study, such as the legal, economic, or reproductive domain. Upadhyay explained that combining domains of empowerment into one measure reduces the predictive validity of that measure and can mask relative empowerment in one domain compared with the others.
In addition to these methodological challenges, Upadhyay identified current gaps in measures and research on empowerment, including a need for more research in low- and middle-income countries. Currently lacking are measures of women in nonmarital and nonheteronormative relationships, unpartnered women, and adolescents and young adults. In addition, more research is needed concerning the impact of declines in fertility and increases in contraceptive use, as well as evaluations of how gender-transformative approaches affect women’s empowerment. In Upadhyay’s view, the goals of reducing fertility or increasing family planning should not drive motivation to increase women’s empowerment; instead, increasing empowerment as a goal in itself as well as promoting reproductive autonomy—with access to safe and legal abortion as a central feature of that autonomy—should serve as primary motivators for research and efforts related to women’s empowerment.
Upadhyay introduced the Sexual & Reproductive Empowerment Scale for Adolescents & Young Adults (SRE Scale for AYA), a collaborative effort to consider empowerment as a predictor of desired contraceptive use among adolescents and young adults (ages 15–24). In quotient form, this scale considers individuals who use a contraceptive method that aligns with their preferences in the numerator and individuals who want to use a contraceptive method other than that which they currently use in the denominator; if no contraceptive method is desired, the individual is sorted based upon his or her ability to act accordingly. The 23-item scale was developed between 2015 and 2019 and reflects seven domains of sexual and reproductive empowerment: sexual safety; self-love; sense of future; sexual pleasure; comfort talking with partner; choice of partners, marriage, and children; and parental support. Although the scale was developed in the United States, Upadhyay hopes to see the scale adapted and tested internationally. Finally, Upadhyay introduced a new resource for women’s empowerment measures, the University of California Global Health Institute (UCGHI) Database of Gender and Women’s Empowerment Measures.2
___________________
During the discussion that followed the presentation, Alex Ezeh asked why outcomes related to women’s fertility and family planning were used to validate the SRE Scale for AYA despite Upadhyay’s position that these should not be the ultimate outcomes in measurement of women’s empowerment. Upadhyay noted that the validation measure for the SRE Scale for AYA was not contraceptive use but desired contraceptive use, which she believes reflects the ultimate goal of reproductive autonomy. The links between women’s empowerment, contraceptive use, and fertility reduction are well established in the literature, but if these are the primary motivations for work on women’s empowerment, the goal of reproductive autonomy may be overlooked.
Aletheia Donald (World Bank) asked about the degree of overlap between the UCGHI Database of Gender and Women’s Empowerment Measures and the EMERGE database.3 Upadhyay explained that the former focuses on empowerment measures that have been used to assess health outcomes.
Sarah Baird (George Washington University) requested advice on what actions researchers could take when social norms pose challenges to asking a particular survey question. Upadhyay advised that surveys avoid language that marginalizes some respondents. Surveys may also combine encompassing and clarifying language so that questions are understood by the majority. For example, the primary language of a question may use the term “partner” and follow it with a clarifier for those who may not understand this language to refer to their “husband” or “wife.”
MEASURING WOMEN’S EMPOWERMENT IN THE CONTEXT OF FAMILY PLANNING
Sonalde Desai (University of Maryland) presented three domains of measurement that describe women’s empowerment in family planning and other outcomes. The first domain is institutional, in which measurements describe the structure that governs women’s agency and interactions. Here Desai drew a distinction between formal institutions (e.g., laws, labor markets, education systems, health services, transportation) and informal institutions (e.g., social norms, stratification systems). The second domain of measurement involves agency itself, which can be classified as individual agency (e.g., psychological, social, material, cognitive) or collective agency (e.g., organization, representation, social support, knowledge transfer). The third domain measures interactions, such as negotiations with partners and
___________________
3 EMERGE, which stands for Evidence-based Measures of Empowerment for Research on Gender Equality, is a project of the Center on Gender Equity and Health (GEH) at the University of California, San Diego.
family as well as health care providers and bureaucrats. Desai described these domains as micro- and macro-level “building blocks” to consider when measuring empowerment.
Desai said that most literature to date focuses on individual-level agency or intra-family negotiations, partly due to the simple fact that relevant data exist, for example from demographic and health surveys. More recent efforts to expand these measurements attempt to evaluate more dimensions of agency, such as self-efficacy and gender attitudes, and the linkages between social institutions and individual agency, such as through access to paid work, land ownership, and financial inclusion. Desai pointed out that several of the “building blocks” to measure empowerment remain missing, including those that focus on informal institutions, collective agency, interactions with service providers, as well as the influence of changes and interactions over the life course that ultimately lead to increases in agency; these gaps were the focus of Desai’s presentation.
A key challenge in measuring informal institutions (such as norms) is the question of whether empowerment varies more among individuals or communities. Desai discussed a study of five Asian countries that demonstrated far greater variation on almost all measures of empowerment among communities than among individuals within a community: 40 to 80 percent of intercommunity variation in empowerment measures (e.g., decision-making ability, freedom of movement, domestic violence) could be explained by aggregated community responses without including any individual traits (e.g., individual educational attainment).4 She cited the study author’s conclusion that the results were “consistent with the assumption that empowerment is a group-based [rather than individual] process.”
However, Desai noted that aggregation of data is not especially satisfactory because it may identify spurious relationships. For example, she said that one could posit an effect in which the marital status of her neighbors in graduate school had influenced her own marital status at the time, whereas the fact that she lived in married student housing actually explained that similarity. Thus, measurement of informal institutions must consider whether an attribute describes an individual or a community. For illustrative purposes, Desai provided examples of how surveys might word questions about social norms to capture this distinction between individual gender attitudes and community-based gender norms:
- Measures individual attitudes: “In your opinion, is a husband justified in hitting or beating his wife in the following situations.…”
___________________
4 K.O. Mason and H.L. Smith, Husbands’ versus wives’ fertility goals and use of contraception: the influence of gender context in five Asian countries, Demography 37(3), 299–311 (PMID: 10953805).
- Measures community norms: “How many people in your community would think it normal for a man to beat his wife in the following situations.…”
- Measures community norms: “If a man failed to punish his wife in the following situations, he would lose respect in your community.…”
When measuring agency, it is critical to assess whose agency is being measured: the individual’s or the collective’s. Desai pointed to a body of work that measures individual agency (e.g., self-efficacy, intrahousehold decision-making power, physical mobility), but noted that some situations may call for measures of collective agency. Participation in self-help groups, community involvement in hiring health care workers, informal social networks for discussion of childbearing and contraception, and community-focused sex-education programs are all relevant measures of collective agency for family planning research. Desai pointed out that collective agency is an especially suitable arena for policy intervention.
Desai explained that measurements of social interactions must consider the locus of power in the interaction, for example whether power is exercised vis-à-vis family members or institutions. Measures of empowerment tend to focus on intrahousehold negotiations, yet these interactions may not play a large role in outcome measures. Women and men largely agree on fertility preferences, for example, rendering intrahousehold negotiations less consequential than interactions with health systems.5 Desai noted that many women in India opt for sterilization over intrauterine devices (IUDs) or birth control pills because health providers emphasize the former practice over the latter, and thus family planning research on this issue should focus on the power imbalance between women and these institutions, not on that between women and their male partners. Data on resistance to or general understanding of overtly or subtly coercive family planning policies can also shed light on women’s empowerment, because women with more agency may be better equipped to navigate these systems. Examples of such resistance include subversion of the one-child norm in China and negotiation of provider resistance to offering abortion.
To measure the influence of interactions on women’s agency over time, data on stability versus change are needed. Desai explained that empowerment and family planning are dynamic processes and that excessive reliance on cross-sectional data creates a false sense of empowerment as a stable attribute of individuals. Although work exists on LFP and fertility over time, longitudinal data on other aspects of gender empowerment will enable
___________________
5 K.O. Mason and H.L. Smith, Husbands’ versus wives’ fertility goals and use of contraception: the influence of gender context in five Asian countries, Demography 37(3), 299–311 (PMID: 10953805).
family planning research in more areas, including growth in intrahousehold power following the birth of a son in patriarchal societies, the ability to continue paid work with longer birth spacing, childcare constraints on community participation, and the identification of key ages or life stages where family planning interventions matter most. Desai emphasized that the field may be able to understand only some of these dynamic relationships through the collection and analysis of longitudinal data. Finally, she concluded by acknowledging that variation among countries and cultures, as well as intra-country variation (e.g., intersectionality by race, gender, class), were not addressed in this presentation despite their importance.
Following Desai’s presentation, Mayra Buvinic called upon funding agencies to prioritize longitudinal studies of women’s empowerment and family planning. Sajeda Amin (Population Council) asked whether complications with data aggregation can be addressed through strategic study design and the use of the level of aggregation as a reference group. Desai responded by highlighting two potential issues specifically related to aggregating data from the reference group of neighbors. First, spurious relationships in the data may be generated, which Desai acknowledged can be rectified through strategic research design. Second, social norms may not be reflected in the behavior of neighbors; for example, even if norms dictate that women working outside the home is a sign of low status, no one in a poor neighborhood could live up to that norm even though they subscribed to it.
Jocelyn Finlay inquired whether the research community can press the Demographic and Health Surveys (DHS) Program to embrace longitudinal studies. Desai advocated for more longitudinal research in general but also explained that it may not be easily achieved in large, cross-national data sets. Sunita Kishor added that DHS cannot tackle every methodological problem and that a push for longitudinal DHS studies is aspirational. Kishor also said that longitudinal data will likely be more useful when focused on a specific set of key indicators, and noted that donors and survey funders are interested in innovative survey designs that might allow for a subset of longitudinal measures.
Nancy Birdsall inquired whether anyone on the panel had leveraged natural experiments to measure women’s empowerment. Desai shared that three natural experiments occurred between two survey waves in India: first, the implementation of a job guarantee program that had a positive impact on women’s LFP; second, the introduction of cash transfers for hospital deliveries, which also had a positive impact on women’s LFP; and third, the introduction of transportation networks within a village, which led to women’s participation in nonagricultural work. Birdsall added that another critical opportunity for a natural experiment would be to follow a policy change specifically related to contraception access, family planning, or general reproductive health.
MEASURING WOMEN’S AGENCY IN FAMILY PLANNING RESEARCH: A CALL TO TRANSFORM PARADIGMS
Kathryn M. Yount (Emory University) described a process to conceptualize and measure agency, with emphasis on the fact that although agency may be an abstract concept, it is nonetheless important to define it as a validated multidimensional construct that can be applied consistently across contexts.
Yount presented findings from a literature review that evaluated the relationship between measures of women’s agency and contraceptive use in low- and middle-income countries.6 This review spanned multiple decades and languages, and focused on quantitative analyses of population-based surveys that narrowly defined agency as freedom of movement and decision-making. The review was restricted to 12 studies that met the inclusion criteria; these studies were conducted between 1996 and 2014, primarily in sub-Saharan African and South Asian low- and middle-income countries, and generally relied on DHS data. Yount described the quality of the included studies as either low or moderate, and highlighted the various ways that researchers referred to empowerment and contraceptive use as well as the ways they inconsistently applied or vaguely defined measures. Given this considerable variability in clearly defined, operationalized constructs in the literature, it was unsurprising to Yount and colleagues that the relationships identified among measures of decision-making, freedom of movement, and contraceptive use were mixed. One study, however, was conducted in a sub-Saharan low- or middle-income country and relied on more items of measurement, two latent constructs, and specific measures of contraceptive use; this study did find a positive relationship between both aspects of agency and contraceptive use.7
Yount then presented a sampling of measures of women’s sexual and reproductive health (SRH) agency. Rather than offering a comprehensive measure of women’s overall SRH agency, these measures tend to focus on specific concepts, such as sexual activity, HIV risk, prevention of pregnancy using contraception, adolescent SRH, and agency in pregnancy. They vary substantially in dimensionality, item number, level of validation, and method of psychometric assessment. In addition, few studies have evaluated these measures with regard to test-retest reliability or replicated
___________________
6 L. James-Hawkins, C. Peters, K. VanderEnde, L. Bardin, K.M. Yount, Women’s agency and its relationship to current contraceptive use in lower- and middle-income countries: a systematic review of the literature, Global Public Health 13(7), 843–858 (doi: 10.1080/17441692.2016.1239270; Epub 2016 Oct 1. PMID: 27690750).
7 M. Corroon, I.S. Speizer, J.C. Fotso, et al. The role of gender empowerment on reproductive health outcomes in urban Nigeria, Maternal and Child Health Journal 18(1), 307–315 (doi:10.1007/s10995-013-1266-1).
cross-cultural validation, which is a concern for longitudinal research that tracks agency over time. Yount asserted that the field currently lacks a comprehensive measure of women’s SRH agency that is consistently defined, operationalized, and validated across groups, context, and time.
Yount offered a multidimensional construct of agency defined as a process of conscientization whereby women—individually and collectively—become aware of and act to overcome their disempowerment. The three dimensions of this construct are intrinsic agency, instrumental agency, and collective agency. Intrinsic agency includes state of mind, awareness of rights and aspirations, confidence in capabilities, and internal motivation to act. Instrumental agency includes acting strategically, exercising voice, influencing decisions, and moving freely in historically restricted spaces. Collective agency includes connection, joint efficacy, engaging in or leading networks with shared goals, and joint action in pursuit of shared goals to effect change.
In addition, Yount and colleagues have designed a sequential, mixed-methods process to develop scales of women’s agency that are both general and specific to SRH. Steps in this process include qualitative research; questionnaire review; item pool development; expert review to assess the item pool; cognitive testing with target populations and critical stakeholders; form pretesting; pilot survey; psychometric validation; and replication across groups, contexts, and time. This process can identify comparable measures of women’s agency and validated contextual items, although Yount emphasized that skipping any of these steps could compromise the resulting scale.
Given the variation in methods used in the field, Yount reviewed the specific steps taken to perform the psychometric validation phase of this process. These steps include univariate and bivariate analysis of included items; exploratory factor analysis in a random-split sample to assess dimensionality; confirmatory factor analysis in a second random-split sample; and measurement invariance analysis across groups, across contexts, and over time. After generating a valid, long-form measure of women’s agency, short-form subsets of measures may be identified that are adequate when less time is available to survey.
Yount provided two examples of measures recently validated by the proposed construct definition and development process: the Women’s Agency Scale 61 (WAS-61),8 a general agency measure composed of 61 validated items tightly linked to dimensions of intrinsic, instrumental, and collective agency; and the Reproductive Agency Scale 17 (RAS-179), a
___________________
8 K.M. Yount, Z. Khan, S. Miedema, Y. F. Cheong, and R. T. Naved, The Women’s Agency Scale 61 (Was-61): a Comprehensive Measure of Women’s Intrinsic, Instrumental, and Collective Agency (available at SSRN).
9 K.M. Yount, L. James-Hawkins, and H.F.A. Rahim, The Reproductive Agency Scale (RAS-17): development and validation in a cross-sectional study of pregnant Qatari and non-Qatari Arab Women, BMC Pregnancy and Childbirth 20(1), 1–15.
more concise measure of reproductive health agency developed in Qatar that consists of 17 items that are also linked to the aforementioned three key dimensions of agency. Yount concluded that if the field can agree on a set of concepts that focus on intrinsic, instrumental, and collective agency, then the development of consistent and widely applicable general and SRH agency measures is achievable.
During the discussion, Yount added that in conducting a survey experiment in Bangladesh, her team has found that responses are sensitive to changes in question wording and order and that it is important to bring rigor to the measurement of agency in this regard. Amy Tsui asked whether any panelist had experience measuring change in empowerment at the community level in a randomized experiment. Yount responded that Emory University is collaborating with CARE on a large community-based randomized trial that will assess a social-norm-change initiative and a movement-building initiative among adolescent girls. The trial will have the capacity to measure intergenerational and community-level changes in intrinsic, instrumental, and collective agency.
MEASURING WOMEN’S AGENCY
Aletheia Donald (World Bank) described agency as a central component that binds resources to outcomes, such as child nutritional status and contraceptive use, and emphasized the importance of measuring empowerment consistently in order to definitively characterize this relationship. Citing Naila Kabeer’s definition of agency, Donald explained that agency is the “ability to define one’s goals and act on them” and further outlined three required dimensions of this ability: goal-setting, perceived control and ability (“sense of agency”), and acting on goals.10
In order to understand an individual’s ability to set goals (the first dimension of agency), one must assess the individual’s capacity to define goals as well as to align those goals with the individual’s own values and preferences. Goal-setting is often measured by standardized questionnaires and scales in industrial and organizational psychology research, and in the context of entrepreneurial interventions in Africa it is often measured by post-coding respondents’ goals based on specificity and detail.11 Donald noted that questions about goals and preferences have more recently been used to complement questions about decision-making roles in recognition
___________________
10 N. Kabeer, Resources, agency, and achievements: reflections on the measurement of women’s empowerment, Development and Change 30(3), 435–464.
11 M. Frese, S. Krauss, N. Keith, S. Escher, et al., Business owners’ action planning and its relationship to business success in three African countries, The Journal of Applied Psychology 92, 1481–1498 (doi: 10.1037/0021-9010.92.6.1481).
of the fact that a woman’s desire to make a decision is a critical factor in assessing whether her involvement in that decision truly reflects agency. The most common measure of whether an individual’s goals align with her (or his) values is the Relative Autonomy Index.12 This index measures the extent to which motivation is external (e.g., a product of coercion or avoiding repercussions), introjected (i.e., to adhere to norms), or identified (i.e., an autonomous preference). Work in this area typically utilizes hypothetical vignettes to ground respondents’ answers.
Individuals’ “sense of agency” (the second dimension of agency) has been studied in many disciplines. Measurement of this sense of control in cognitive science has employed brain imaging studies and sensory tests that capture different perceptions of self- and externally generated stimuli. The social sciences have historically conceptualized sense of agency according to a framework developed by Ellen Skinner in the 1990s, which characterizes the relationship between resources, or means, and goal-related outcomes, or ends. Donald outlined two prominent concepts related to this framework: locus of control and self-efficacy. Locus of control is the degree to which one believes an event is caused by one’s own actions, while self-efficacy is the belief in one’s capability to bring about an action. Locus of control is traditionally measured using the Rotter I-E Scale, which distinguishes between external and internal loci of control, or the Multidimensional IPC (Internality/Powerful Others/Chance) Scale, which further parses external locus of control according to an individual’s belief that an outcome is a function of chance or the will of a powerful “other.” Both scales have been validated for well-educated professional sub-populations, and Donald noted their increasing use among low-literacy populations. Self-efficacy has been measured as both generalized and domain-specific self-efficacy; Donald pointed to the Condom Use Self-Efficacy Scale as a validated example, which has been used in Ghana to evaluate one’s efficacy to use condoms or negotiate the use of condoms with partners.
Donald said that because scales that measure sense of agency can be quite long, some measurements have been developed that attempt to describe the “means-ends” relationship directly. The most popular is the Control Scale, in which respondents grade their perceived sense of control over the direction and outcome of their lives on a scale of 1 to 10; internationally, women score significantly lower than men on the Control Scale. The Fate Scale, which is more closely related to locus of control, was introduced more recently and asks respondents to grade their belief that they control their destiny on a scale of 1 to 10. The World Value Survey
___________________
12 R.M. Ryan and E.L. Deci, Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being, American Psychologist 55(1), 68–78 (doi: 10.1037/0003-066X.55.1.68).
collected nationally representative data from 90 countries for both the Control and Fate Scales and demonstrated that response discrepancies vary systematically across income levels; the responses to both scales overlap nearly perfectly in high-income countries, and the discrepancy between responses becomes progressively larger with lower income levels. This relationship between income level and response discrepancy also differs by gender. Indeed, Donald pointed out that the only region where the response discrepancy is small and similar across genders is North America. Donald warned that researchers must take care when conducting cross-country analyses to consider the systematic influence of poverty and gender on outcome measurements.
The capacity to act on goals (the third dimension of agency) has been measured since the 1960s through surveys of intrahousehold decision-making roles.13 Donald highlighted two main drawbacks to the most commonly-used decision-making questions: first, the questions fail to capture negotiations or bargaining processes as they relate to decision-making in a household, and second, ambiguous phrasing can lead to varied interpretation of questions by respondents in ways that are systematic and exceed random measurement error. Some measures, including the Women’s Empowerment in Agriculture Index, aim to address the weaknesses in these surveys by asking questions specifically related to decision-making processes, such as “How much input do you have in making decisions about a certain topic?” or “To what extent do you feel you can make your own decisions if you want to?”
About 11 years ago, the DHS started to separately ask husbands and wives questions about household decision-making roles. Donald presented results from a study that analyzed these data from more than 71,000 couples across 23 sub-Saharan African countries, focusing on questions about decision-making roles for major household purchases.14 Consistent with previous public health research, the study showed that well-being outcomes, such as those concerning women’s general and reproductive health, are often highest when spouses agree on the woman’s power, yet nearly half of couples disagreed about roles and decision-making power.
Compared to cases of uncontested power or of husbands “giving” power, cases of women “taking” power were more positively correlated with key economic empowerment proxies (e.g., higher earnings than their husbands, land ownership). Women “taking” power was more strongly
___________________
13 R.O. Blood, Jr., and D.M. Wolfe, Husbands and Wives: The Dynamics of Family Living (Free Press Glencoe, 1960).
14 J. Annan, A. Donald, M. Goldstein, P. Gonzalez Martinez, and G. Koolwal, Taking Power: Women’s Empowerment and Household Well-Being in Sub-Saharan Africa, Policy Research Working Paper No. 9034 (Washington, DC: World Bank, 2019 (https://openknowledge.worldbank.org/handle/10986/32494).
related to beneficial household welfare outcomes (e.g., children’s health, women’s reproductive health) than husbands “giving” power. In some cases, women “taking” power was even associated with similar benefits as them having uncontested power, such as for girls’ stunting and wasting, or an increased prevalence of outcomes such as women terminating a pregnancy. However, Donald also noted a stark relationship between women “taking” power and higher prevalence of emotional violence. Donald said that analysis of response discrepancy is an active field, and that what may at first seem like unreliable self-report data can actually provide insight into different aspects of women’s agency.
Donald called for more systematic analytical work to test alternative forms of measurement and the appropriateness of measurements in the context of given policy circumstances and research questions. Some of this research should evaluate all three dimensions of agency together across contexts in order to consider women’s preferences alongside self-perceptions and to provide a more comprehensive understanding of agency. She also recommended that measurements of agency be expanded to include other age groups and designed to systematically adapt to local contexts. A new collaboration called Measures for Advancing Gender Equality (MAGNET) aims to fill some of the existing gaps. Donald noted that advances in agency measurement will help sharpen theoretical insights on how households actually work, and lead to the design of better programs and policies that can more effectively target particular cognitive or behavioral pathways in enhancing women’s empowerment.
In the discussion, Anita Raj asked for examples of advancements in the measurement of self-efficacy and decision making. Donald replied that scales for measuring self-efficacy have been introduced in contexts where they were not previously used, such as in entrepreneurship impact evaluations. For measures of decision-making, questions have increasingly been designed to unpack the process that generated the final decision, such as a woman’s sense of input and her decisional outcomes in the event of conflict. They have also moved toward analyzing dimensions of agency, such as by assessing decision making alongside measures of motivational autonomy or locus of control.
Kathryn Yount added that labor force panel surveys in Egypt and Jordan contain longitudinal data on decision-making, attitudes about violence, and freedom of movement that show a subset of items that are time-invariant over a 12-year period, indicating a capacity to develop indices that can track change over time. Yount also commented that there is a tendency to include general self-efficacy measures when assessing women’s agency and advocated for making those general measures more specific (e.g., confidence in self-expression, confidence in influencing decisions).
This page intentionally left blank.
