Investigating the Conceptual Plurality of Empowerment through Community Concept Drawing: Case Studies from Senegal, Kenya, and Nepal

: Women’s empowerment is a driving concept in gender and development scholarship. This scholarship often engages quantitative indices of evaluation that are unable to account for culturally speciﬁc meaning and nuance that shape local understandings of empowerment. Recent efforts within the ﬁeld of international development are attempting to create methodological mechanisms for capturing this nuance. This study employs one such method, Community Concept Drawing (CCD), in rural villages within Kenya, Senegal, and Nepal. Findings indicate signiﬁcant differences between the ﬁeld sites in the local conceptualization of empowerment. Cross-examination of site-speciﬁc data yields an understanding of how cultural norms and values shape local perceptions of empowerment in ways that are critical for research that engages gendered understandings. Furthermore, such analysis is critical to a more accurate understanding of the locally speciﬁc context of gender inequity.


Introduction
In the early 2000s, international development institutions identified women's empowerment as a critical pathway for building sustainable livelihoods [1] by rectifying gender inequities in the productive and reproductive work carried out by men and women in households and communities. These development efforts have long been shaped by neoliberal principles that prioritize economic indicators of empowerment [2,3]. Previous research has shown that development programs rooted in the ideologies of industrialized nations often fail to engender sustainable and equitable empowerment in other contexts at the local level. According to Porter, this is problematic because "empowered communities are the primary site for transformative agency" [4] (p. 11).
Critique of neoliberal approaches in empowerment-oriented interventions is not new, and there is growing consensus that many pathways to empowerment exist [5] as do diverse visions of the concept itself [4]. These pathways and concepts are shaped by culture, norms, and structures that are specific to local context. Qualitative approaches have proven useful for bringing alternative visions of empowerment to the fore by complicating research findings that investigate the concept through proxy measures including income, productivity, time, and decision-making. Methodologies that facilitate iterative and reflective processes among researchers, practitioners, and communities can foster a deeper understanding of how empowerment operates in diverse settings. This may be critical to designing gender transformative interventions.
Community Concept Drawing (CCD) is a participatory visual method designed to facilitate interactive research between stakeholders [6]. In this article, we present the ways in which CCD illuminated diverse understandings of power within and across multiple contexts, and why this is important in moving the study of empowerment forward within rural development efforts. We argue that CCD provides information about empowerment that is useful for establishing stronger gender transformative approaches to development initiatives by including the voices of both women and men in discussions about empowerment and reorienting the conceptualization to the local level, instead of deriving concepts of empowerment from institutional and academic reservoirs.
We organize this article as follows. First, we review key lines of thought on how empowerment is conceptualized and measured within development scholarship. Next, we present methods and results from research to understand the conceptions of empowerment in Senegal, Kenya, and Nepal. These data were collected between 2014 and 2018 to understand the role of empowerment in improving food security of rural, smallholder farming households as part of the Consultative Group for International Agricultural Research (CGIAR) program on Climate Change and Food Security (CCAFS; Kenya and Senegal) and two USAID funded projects, Feed the Future Innovation Lab for Livestock Systems (LSIL) and Integrating Gender and Nutrition within Agricultural Extension Services (INGENEAS). Finally, based on these data, we draw our main argument: understandings of empowerment are diverse both within and between local contexts and tools that uncover these pluralities are essential in development. In this last section, we argue that development organizations interested in fostering empowerment in communities should begin their efforts by examining how global challenges are felt locally.

Background: Rethinking Empowerment
Scholars, practitioners, and community stakeholders call into question the theoretical underpinnings of the term "empowerment" and the extent to which concepts and measures associated with it reflect the contextual nuances that give "empowerment" its meaning [2]. One of the main points of contention within the empowerment debate is that development institutions often frame empowerment in ways that do not reflect local understandings and exclude how empowerment is experienced diversely by both men and women. However, these features contribute conceptual complexity to empowerment. Empowerment may hold variable meanings across contexts and among individuals according to social position [7,8]. This approach has been driven by neoliberal ambitions of development institutions, which have historically defined empowerment according to economic frameworks for development [2,9]. Scholars in this line of thought emphasize the potential for material-based indicators of empowerment, such as assets and wealth accumulation, to support a more equitable balance of power at the household and community level [9,10]. However, these models tend to measure women's status against the productive activities and assets traditionally held by men while failing to address distinct needs facing men and women [4]. Still, the notion of "empowerment as economic empowerment" is not in and of itself problematic. Rather, it is problematic that "empowerment as economic empowerment" has become dominant within development discourse and practice to the detriment of alternate ways of viewing and experiencing power, and that conversations around empowerment are often reduced to economic indicators [11,12]. Furthermore, components of empowerment that are process oriented, relational, and dynamic are often neglected or misunderstood when applying solely quantitative measures [13].
In the hope of bringing research "to scale", international development practitioners utilize quantitative tools to draw comparative assessments across multiple sites and contexts. One tool that has brought excitement to the fields of gender and international development is the Women's Empowerment in Agriculture Index (WEAI). The WEAI identifies five domains of empowerment resources, income, production, leadership, and time in which gendered inequalities in decision-making may exist at the household level [14,15]. The WEAI has generated renewed interest and experimentation in the measurement of empowerment. Dialogues across the field of gender and development have emerged, and new indices (e.g., ProWEAI, WELI, WENI) and modules (e.g., nutrition, livestock, etc.) have been developed to contribute to and improve upon the WEAI [16][17][18].
While the WEAI and other metrics offer new insights into the ways in which empowerment exists in development contexts, they fall short for many of the same reasons other quantitative tools have: they are not designed nor equipped to measure the relational and dynamic nature of empowerment. According to critical scholars, quantitative measures of empowerment tend to emphasize visible forms of agency-such as income and decision making [16]. In doing so, they risk obscuring the conceptual depth of empowerment. Meanwhile, others press that much of the transformative potential of empowerment is already lost because it has become disconnected from local contexts [2]. The complexities inherent in the concept have been reduced to basic indicators, which neglect its intrinsic inter-relational, social, and dynamic qualities. Still others fear empowerment has become a hollow concept, devoid of meaning amid development's attempt to make generalizations about gender inequity [2].
Recently, international development has sought to understand gender as a transformative process [13]. Gender transformative approaches (GTA) aim to move beyond indicators of equality between men and women and instead address the norms embedded within society and reflected in the attitudes and behaviors of its members, thereby transforming gender inequalities. However, to do this effectively, it is critical that researchers and practitioners employ research tools that allow them to understand the dynamic mechanisms at play in the communities with whom they are working. This must begin with the recognition that empowerment is not experienced uniformly, even within one specific setting; rather, it is intersectional by nature. Norms guided by culture, religion, and colonial experience can shape what factors are valued as empowering in local spaces. Quantitative tools that measure indicators of empowerment do not account for the diverse experiences of empowerment among women and fail to reflect intersectionalities, which may constrain empowerment along age, class, caste, or ethnic lines, both within and among communities [8].
There is much to be gained by investigating empowerment and its relationship with food security and nutrition, as well as other development objectives, but how it is studied must be rethought. Consensus across the feminist development literature suggests that there is not a "one size fits all" approach to empowerment, but rather these conceptualizations of empowerment can differ depending on the structural and normative dynamics shaping relationships in a given home or community and may change over time. Importantly, studying empowerment using methods or indicators developed largely by western researchers and practitioners risks reifying western concepts of empowerment, rather than studying the phenomena on local terms. Research methods should allow for multiple conceptual foundations of empowerment to be explored within a given context. In what follows we present the results of CCD, a research method that engages both men and women to understand multiple interpretations of empowerment within a given cultural context. We present findings from our engagement with CCD in three countries, where we sought to understand empowerment to better inform interventions for food security. In our conclusion, we argue that CCD provides us with the type of information critical to the development of gender transformative approaches for pursuing food security and other development goals.

Materials and Methods
Our team facilitated CCD, a participatory research activity, to explore the cultural nuances of empowerment among select sites. Here we provide an overview of the method and discuss case studies from Senegal, Kenya, and Nepal. As we examine each case, we present the diverse meanings of empowerment articulated by participants at each site.
CCD is a participatory visual method (PVM) that invites interlocutors into the process of imagining and defining a given concept [6]; in this case, the study participants defined the term "empowerment" as it exists and is experienced within their community. (A full description of the methodology, including the facilitator question guide are presented in a forthcoming article, see McOmber et al., 2022.) CCD is administered in a group discussion format with a trained facilitator. The discussion unfolds through four distinct phases-brainstorming, elaboration, reflection, and transformation. In brief, the facilitator asks participants to describe and engage in discussion about the concept of empowermentwhether they have heard it used in a particular context, whether there are groups that are included in the use of this concept, and those who are excluded. Once a general discussion of empowerment is concluded, the participants are then asked to draw those components of empowerment that resonate with the experiences of people in their community. The debates, discussions, narratives, and illustrations that emerge from this process allow both the researcher and participants to disentangle the concept of empowerment from the broader web of social complexity into which it is tightly woven. Each step is described in more detail below.
Prior to engaging the community members in brainstorming, careful attention must be given to translation of the term and the implications of that translation. If the goal is to make comparisons, it is important to translate empowerment into terms that are culturally relevant, and when possible, comparable in meaning across contexts. For example, in case studies carried out in Senegal, Kenya, and Nepal, the local translation of empowerment derived from the word for "to be able", thus that translation was used in in Wolof (in Malem Hodar and Kaffrine), Swahili and Kikamba (in Makueni County), in Swahili, Kalenjin and Luo (in Kisumu County), and Nepali and Newari (in Kavrepalanchok). The Kenyan group discussions were run primarily in Swahili but switched to local languages throughout the discussion, particularly when participants felt more comfortable speaking in their first language or if there was a need for further clarity. In Wolof the expression menal sa bopp, or "to be able to do for oneself" was used. In Swahili the term kuwezesha was use and in Kikamba the term "kutheka" was used in Makueni County. The term in Luo is mijingo and in Kalenjin the term kekimit were used in the Kisumu groups. All of these terms mean 'one who is able'..) The participants are then asked whether they can identify any types of people who are empowered in the community. Next, the facilitator asks the participants whether they are familiar with the term "disempowerment". Again, the facilitator inquires whether the participants can identify any types of people in the community who are disempowered.
Placing a blank poster on a wall or table, the facilitator invites the participants to elaborate on these initial ideas of empowerment by drawing an ideal typical empowered man in the community. Probing questions may include what the man's family structure looks like, what his home looks like, and/or what his roles and responsibilities look like in the community. When all ideas about the most empowered man have been drawn, the facilitator asks the participants to draw what the least empowered man looks like. Again, the same type of probing questions can be applied. Finally, the facilitator asks whether there are any middle categories between the "most empowered man" and the "least empowered man". If there are, the facilitator repeats the exercise for the middle-empowered categories, recognizing there may be one or more intermediate levels of empowerment. Later in the exercise, these steps are repeated to examine women's empowerment. The next step is reflection. During this step, participants are asked to identify the most important indicators for determining men's empowerment. The facilitator asks the participants: of all the items drawn for the most empowered man, which are the three most important in determining his empowerment? A facilitated discussion generates a consensus list. The last step guides participants through a discussion about transformation from one "level" of empowerment to the next. The facilitator asks the participants: what would be necessary to move the "least empowered man" to the intermediate category? Once this is complete, the researcher asks: what would be necessary to move the "intermediate empowered man" to the highest category? This can help the researcher to understand the structural barriers that are preventing empowerment in the community. If more than one barrier is listed, the community is asked to rank those that are most significant to hindering upward mobility in empowerment. Once this activity has been completed for men's empowerment, the process is repeated in its entirety to conceptualize women's empowerment.
As the CCD method is facilitated through a similar format as a focus group discussion, it is subject to many of the same methodological limitations [6,19]. The data from each group discussion are collected as an aggregate, a culmination of the group's deliberations. During a discussion, many viewpoints-including contradictory views-may emerge. It is this community dialogue and deliberation over a concept that produces the critical information necessary to deconstruct concepts and the processes and mechanisms which drive it. From a methodological standpoint, this messiness is what we are hoping to capture; we are not necessarily interested in consensus around any one idea or perspective. Instead, untangling multiple experiences of empowerment from the intersectional lives of the participants is a critical part of this process. To avoid the dominance of a singular voice or "group think," facilitation of discussions require skill to ensure all participants are heard and able to contribute their thoughts. Keeping focus groups small in size can help to encourage participation; facilitators can ensure each participant has a chance to draw and contribute to the community's discussion of empowerment and that their perspectives are noted in accordance or contradiction with the other participants. Additionally, because this is a qualitative ethnographic methodology, the explanatory power of our data is not intended to be generalizable. Indeed, we are seeking to understand the multiple pathways to empowerment in specific contexts. Our expectation is that, while there may be some similarities across cases, experiences and configurations of empowerment will likely be different.

Approach to Analyzing CCD Data
The CCD method produces two forms of data: visual and narrative. These data should be analyzed both independently and in dialogue with each other. While the oral narrative collected through the process of the group discussion is critical to understanding the story/stories of empowerment in a particular context, the illustrations can provide symbolic imagery that is not always vocalized within the narrative. Although CCD data can be analyzed in several ways [6], for this study we used thematic analysis to qualitatively analyze the group discussion transcripts. We first analyzed the transcripts independently, and then analyzed both the drawings and the transcripts alongside each other. While analyzing the transcripts, we identified common themes that emerged from the group discussions and organized them into relevant subcategories. The three main themes identified in our analysis, include (1) neoliberal conceptualizations of empowerment (2) cultural conceptualizations of empowerment, and (3) transitional processes of empowerment. Within each of these categories, discussion of economic, social, and reproductive components of empowerment were prominent. Our results and consequent discussion in this article are framed around these three themes.
We then returned to the community illustrations and analyzed each part of the drawing as it was being described in the transcripts. This simultaneous analysis helped to provide context for the drawings produced. Through their illustrations, participants utilize symbolic and realist techniques to represent their cultural context. For example, participants may use color, size, or the positioning of drawings to symbolize power relationships. They may also draw out a specific number of children to represent the perceived household size of the least/middle/most empowered individuals. While in the field, facilitators used probing questions wherever necessary to investigate the meanings behind participants' artistic decisions. Probing questions such as, "I noticed you use red when drawing the more empowered woman but a different color for the least empowered. Can you explain the reasons you have used different colors here?" can provide critical context behind the normative visual expressions of power which may be assumed within the group but may not be immediately apparent to the investigator.

Field Sites
CCD is designed to capture plural constructions of concepts as they exist and are understood across and within a given community. While there have been efforts to standardize indicators of empowerment through multiple indices (as discussed above), our intention is the opposite. Through this research, we understand that empowerment is experienced in unique configurations across space and time. We have chosen three diverse case studies that are culturally distinct, acknowledging that these cultural differences are critical to understanding the diverse experiences of empowerment both between and within cases. Culture can shape the ordering of power and its configuration at the state, community, and household levels. We anticipate that, from these diverse case studies, we will be able to also capture and understand diverse experiences of empowerment. Our intention is to highlight context-specific experiences of empowerment within our field sites to better understand how empowerment occurs and is felt within distinct spaces, and to see whether comparable lessons emerge.
For this study we conducted CCD with small groups of community members, organized homogenously. As we were interested in understanding power as it pertains to gender, we organized small groups along socially salient cleavages. In communities where there were strong ethnic distinctions-such as in Kisumu County-group discussions were organized by gender and ethnicity. In sites where there were not strong ethnic cleavages apparent, we organized groups based on gender and age (one group consisting of people between the ages of 18 and 50, and the other 50 years and older). These different groups gave insight into how power was distributed within and among groups, and the fluidity of that power across circumstances. Site and participant details are presented in Table 1 below.

Senegal
Both research sites in Senegal were located in the Kaffrine Region, which is situated in the center of the country, in the heart of Senegal's peanut basin. Peanuts are a major cash crop, driven by both domestic consumption and significant export. The semi-arid environment experiences an annual rainy season between the months of July and September. The Kaffrine region experiences high interannual variability of rainfall [20] and climate change has the potential to significantly impact the region [21,22]. The end of the Saloum River flows into the southern portion of the region. The villages chosen for data collection in 2014 were in two départments, Kaffrine and Malem Hodar, both within the Kaffrine Region. The major ethnic group in the region is Wolof, with minorities of Sérère and seasonal populations of Fulani herders. Wolof and Sérère dominate the agricultural production of peanuts and millet. Many Fulani herders, who graze livestock north into the arid Ferlo valley during the rainy season, return to spend the dry season in the Kaffrine Region, where pasture and water are more readily available livestock.
The Wolof population of within the region is overwhelmingly Muslim, in keeping with the overall population of Senegal. Most rural Wolof families in the Kaffrine Region practice polygamy, with men acquiring up to four wives over the course of adulthood. Many Wolof, both male and female, attend Koranic school, daara, starting at an early age to learn to recite the Koran and read Arabic. Although most of the population does not retain much Arabic literacy, attending Koranic school is seen as an important part in the proper development of a child's moral and social character [23]. A former francophone colony, the Koranic school exists alongside the Senegalese public school system, which is a direct derivative of the French school system. Many children do attend French school; however, few children continue past the first years of primary school. The poor infrastructure of the school system, a lack of resources and adequate staffing, and the obligation of children to contribute to household and agricultural labor contribute to this low retention rate.
NGO presence in Kaffrine Region is high; however, the degree of involvement with NGOs varies significantly from village to village. The majority of NGO engagement in Kaffrine Region during the period of data collection focused on agricultural interventions and climate change adaptation. Many villages in the Department of Kaffrine host Peace Corps Volunteers. Past and ongoing Climate Change and Food Security (CCAFS) involvement in Kaffrine Region has focused on providing access to climate information services for agricultural households. One of the research sites for this study, in the Malem Hodar Department, has been studied extensively in this context, while the other, in the Kaffrine Department, has not.

Kenya
The research in Kenya took place in two counties, Kisumu and Makueni, in May of 2014. Unlike the Senegal sites, where we gathered data from individual villages located a short distance from one another, the Kenyan village sites were in different regions of the country, with Kisumu county in the western region and Makueni county in the east. In Kisumu county, we conducted work in two villages consisting of different ethnic groups, Luo and Kalenjin. In both villages, Christianity is the predominant religion practiced. The Luo village is located in a drier region of the county and suffers greater seasonal rainfall variability. The rivers in this region often run dry, requiring farmers to walk long distances for water. There is one main planting season and many farmers have reported poor crop yields in recent years. While the Kalenjin and Luo villages border each other, the former tends to experience more consistent rainfall. Water sources are also more accessible in the Kalenjin village, which has two cropping seasons. There is also a difference in land ownership between the two village communities, where many Kalenjin households own between 5-10 acres and most Luo households own between 1-2 acres. Households in the Kalenjin village also have more livestock than the Luo and are more dependent on their livestock for income than crop production.
In Makueni County, we conducted CCD in two villages located in the Kathonzweni ward. This area is ethnically Kamba, and religiously Christian. While there are womenheaded households throughout the region, our fieldwork explored villages where dualheaded households were more common. The region is known for its semi-arid environment that has experienced increasingly dry conditions in recent years. Farmers in our study discussed more recent challenges of delayed onset of rains and an overall reduction in rainfall. Extension support through the Ministry of Agriculture and other NGOs have assisted in building climate resilient agriculture practices throughout the region.
There is a significant NGO presence focused on climate change and public health in both Kenyan field sites, but this presence is particularly apparent in Kisumu County. In this county, elevated levels of HIV/AIDS brought a wave of development intervention in the early 2000s, and the result has been an actively engaged civil society. Studies conducted by CCAFS found that there were 20 active community-based organizations in the vicinity of the two participating villages, with as many as 600 households participating as members. Membership of these CBO groups are highly gendered, with women making up over 70 percent of the membership [24]. This appears to reflect the demographics of the region, which has a high rate of female-headed households. In contrast, institutional capacity and overall presence of NGOs and CBOs appeared to be less organized and active in Makueni County at the time of our study; however, the Ministry of Agriculture and some development institutions (e.g., CGIAR, CCAFS) are actively engaged with communities to build climate resilience within agricultural livelihoods.

Nepal
Research for the Nepal case study was carried out during two field visits in the summer and fall of 2017. Study sites for this work included three communities in the Kavrepalanchok (Kavre) district of the Bagmati Pradesh province. These sites are located in a highly heterogeneous socio-cultural context where various castes and ethnic backgrounds were represented. Study participants spoke a variety of languages including Nepali, Tamang, Newari, among others. However, CCDs were conducted in Nepali because of its accessibility across diverse groups. Two Nepali Master's students in International Cooperation and Development Program (MICD) (both native speakers of Nepali, one Newari heritage speaker) and three local key informants from various development institutions helped to navigate these linguistic complexities during CCD sessions and identified approximate linguistic equivalents for empowerment. The terminology used to discuss empowerment varied slightly across field sites. In one village, participants often used the word saks . ama (capable) to speak about empowerment. In the other two villages, the English word for empowerment was used alongside saks . ama, particularly in discussions with women who were comfortable using both terms. These CCD sessions were transcribed word-for-word to draw out distinctions in word choice across each community.
With roughly 65% of Kavre's population dependent on agriculture as their primary livelihood activity [25], smallholder farmers and agricultural wage laborers have been severely impacted by climate-related pressures, natural disasters, declining soil fertility [25]. Amid the unreliability of agricultural work, many men have migrated out of Kavre in search of employment in domestic and foreign cities [26,27]. Over the past 30 years, remittances from migrating family members have come to characterize rural economies in Kavre and other regions of Nepal [28]. In the wake of male outmigration, women have entered new labor responsibilities including land preparation, cultivation of high-value crops, and price negotiation [29]. Previous studies in this region have drawn connections between the "feminization" of agriculture and evolving gender relations [30]. The findings from our case study suggest that the concept of empowerment itself may also evolve in response to emerging labor dynamics.
In the discussion below, we bring findings from Nepal into dialogue with those of Kenya and Senegal to examine how some interpretations of empowerment are distinct, shared, and conditional in different geographic contexts.

Results: Drawing Empowerment: Navigating Local Conceptualizations through Illustrated Storytelling
The illustrations of empowerment-and their accompanying narratives-revealed important factors in local constructions of the concept. Distinctions between age and sex, as well as between geographic locations, were important to communities' interpretations of empowerment. In this section, we break down the conceptualization of empowerment by looking at some of the important ways each of the group participants illustrated their interpretations of the concept.

Exploring Neoliberal Conceptualizations of Empowerment
The most prominent theme among all the research sites was the role of financial security in enabling empowerment. Communities most often talked about economic indicators of empowerment in terms of assets, focusing on material items such as houses and modes of transportation. Disparities between different levels of empowerment were often marked by the presence or comparable size and quantities of items; those in wealthier households tended to be represented in the most empowered category, while those in the least empowered categories tended to have smaller, poorer quality, and fewer financial assets. In Malem Hodar the most empowered were described not only as having a "nice" home, or family compound, but one that was constructed with good quality building materials and consisted of many dwellings. In the case of Makueni county, differentiation included the quality of sanitary facilities, such as latrines, within the homestead. In the Luo group discussion, there was special attention to the assets contained within the home, such as furniture. Nepali participants connected economic empowerment to the ownership of livestock animals. Buffalo were indicative of the highest degree of empowerment and pigs, the lowest. It is important to note that these livestock-related economic indicators overlap significantly with caste hierarchies. In all group discussions conducted across Senegal, Kenya, and Nepal, empowered people were illustrated owning cars. The least empowered individuals, on the other hand, were often described as relying on livestock for transportation. In Kenya, the intermediately empowered person was often drawn using a bicycle for transportation. Similar patterns were identified in Nepal, where intermediate empowerment was associated with the use of public transportation while the least empowered person typically traveled on foot.
Participants in all groups used their illustrations to demonstrate how physical appearance can reflect a person's wealth and was considered to be an indicator of economic empowerment within their respective communities. The empowered person was described as well-dressed in all four sites; the least empowered men and women in all four cases wore poor clothing. In Nepal, one group of participants expressed that an empowered man would be able to "bring tasty foods for their children" after returning home from his office in the city. When asked to describe these foods, participants listed nutrient-poor, albeit luxury, food items, including crisps and candies. In each of the seven field sites for this study, participants used clothing and bodily physique as a way of representing one's economic empowerment; often these depictions were reflective of cultural norms and expectations about health and well-being and its perceived connection with one's overall empowerment. In both Senegal sites, the participants focused on body size and overall shape and figure. For example, empowered women were drawn with an hour-glass shaped body and empowered men were drawn with stocky and strong physique. In contrast, disempowered men and women were illustrated as extremely thin. Participants in these Senegal sites explained that the empowered had access to more than enough food and were therefore well fed and healthy. They also noted that those who were empowered often did not have to engage in heavy physical labor. The disempowered, however, often did not have access to enough food or adequate nutrition and relied on the food produced from their own subsistence farming. Furthermore, the less empowered were unable to hire help and, as a result, put long and hard hours into working on their farms. The health and livelihoods of the empowered and disempowered were placed in stark contrast to each other through these drawings.
Other indicators of empowerment included occupations and levels of education. Agricultural livelihoods seemed to be a standard for middle empowerment. Those who could engage in additional livelihood activities (e.g., owning a business, holding civil service positions, or teaching) were recognized as more empowered. Those who could not manage their own farms, but instead worked as hired laborers on others' farms or herders of others' livestock, were considered less empowered. In Nepal, participants identified important connections between education and other indicators of empowerment already discussed in this section: economic status and diet. For example, one woman stated, "Even if we tried to gain education but we are not good economically even to buy paper and pen then we can't attend school. We can't even have enough food to eat, then we can't do any work without food. Without eating food, we can't concentrate on education." In Senegal, the type of schooling (French or Koranic) was not as important as whether one completed all levels of schooling.
In Kenya, one's level of empowerment was also closely associated with access to education-with those who are most empowered attending and attaining degrees from university, the intermediately empowered attending and often completing secondary level education, and the least empowered rarely attending school at all. There was no discussion of differentiation between men and women's opportunities to attend school during these discussions. Additionally, of particular interest in these discussions was the ability to send children to boarding school, public, or private schools; more pointedly, communities noted the ways in which economic resources can enable different access to different schools, and how this can ultimately both reflect and reinforce levels of empowerment.
Still, while economic indicators are relevant to the story of empowerment in this study, the directionality and complexity of the relationship between empowerment and economic development require attention and nuance. One woman in Makueni County explained, "you first have to be empowered in order to be wealthy". From her own explanation, it is one's empowerment through social capital-through social trust, respect, and reciprocitythat provides opportunities for their economic advancements in the community. In other ways, economic empowerment is reinforced through social relationships and the normatively perceived personal character that is associated with wealth. This is very apparent in the ways that the least empowered men and women are described using derogatory language; their circumstances are often attributed to their "laziness". Women participants in the Luo CCD described the disempowered man as dishonest and known to take advantage of others. A Luo woman participating in the discussion remarked upon how the village members "hate and do not trust him". In Nepal, empowerment was linked to a person's dependability and the registration of their trustworthiness by the community at large. This kind of reliability, according to men in one community, can be demonstrated by a person's ability to complete promised tasks, help others, and "do good deeds" for those less fortunate. In Makueni County, the participants explained how the disempowered man and woman was often isolated from the broader community, and that the shame associated with their situation often led to depression and substance abuse. Behaviors perceived as deviant seemed to exacerbate low economic status and other indicators associated with disempowerment. Substance abuse was discussed openly and critically, particularly in conversations about disempowered men in the CCD sessions in Makueni and Kisumu counties, and in all three Nepali sessions This topic was not raised in the Senegalese CCD sessions, an Islamic society where drugs and alcohol consumption are forbidden.

Understanding How Culture Frames Conceptual Knowledge about Empowerment
While economic components of empowerment were prominently featured in similar ways across the seven case studies, other mechanisms enabling empowerment were contextually specific and often reflected cultural norms. One area in which this was particularly evident was through discussions of family size. While participants in all the case studies understood children to be a potential gateway to improving empowerment within the household, the number of children required for improved empowerment varied across contexts. Households with fewer children, with longer spacing between births, generally associated with higher levels of empowerment within development literature [31][32][33]. This is considered an important indicator for women's empowerment in particular, as it is understood to be reflective of her ability to control her own reproductive health. While these indicators may reflect women's decision-making power and access to family planning resources, it does not reflect women's preferences. Instead, this literature assumes that, given the appropriate resources, women will choose to have smaller families. This assumption does not always hold true, an assertion made clear through CCD in these communities.
In Kenya, the association between birthrate and empowerment was in agreement with the development literature on empowerment; both field sites expressed the importance of family planning in attaining empowerment. It was established through CCD discussions and drawings that the optimal number of children for empowered men and women was between 2-4 children. Participants explained that with fewer children, parents would be able to provide for each of the children's needs equally, including high quality education at a boarding school. Those who were least empowered were illustrated as having many children (between six and twelve within a household). The participants explained that, with so many children the parents would not be able to provide critical needs for the children including food, clothing, education, and healthcare. One man from Makueni County explained, "he [the disempowered man] does not have the potential to educate his children". In Kisumu County, the Luo women explained how children in disempowered households were often naked because their mother does not have the resources to provide clothing for her many children.
Our research in Senegal revealed a gendered distinction important to understanding of the relationship between fertility and empowerment. Our participants explained that, while the most empowered man should have between 10 and 15 children, this is because he should also have many wives (some drawings indicated that the most empowered man had between three and five wives). Participants then explained that, for women, having many children may be indicative of disempowerment. Women with many young children, for example, were more occupied with reproductive activities within the household and have less time for income-generating activities and food production. Women who have children in close succession, a situation referred to as neff, were considered disempowered because neff may be the result of a woman's inability to negotiate family planning with her husband. Participants also recognized that children can be important contributors to household agricultural labor and, therefore, may also contribute to household empowerment. However, the social and economic tradeoffs that come with bearing and raising children may lead to a comparatively more precarious situation for women as compared to men.
In Nepal, families and communities face emerging challenges due to the out-migration of men. Local demographic changes stemming from the absence of men have contributed to shifting gendered labor dynamics, wherein women may take on a greater share of agricultural labor. To mitigate the pressures of increased labor, some communities discussed practicing a collective labor exchange arrangement called parma. For example, one woman explained that "nowadays, we don't go for [paid] labor work. We simply come to pay parma". In this scenario, parma is "paid" when reciprocated. Men also recognized parma as central to maintaining agricultural production in their absence. However, men from a different community in Nepal claimed that it can only take place where the community members have a "sound relationship". Participants further explained that one can become empowered through the process of paying and receiving parma. This kind of solidarity work can strengthen the social standing of individuals in the community. Finally, parma appears to represent a key safety net among women participants, who connected the practice with economic stability.
Another way in which empowerment is shaped by cultural differences is through marriage norms. In Kenya marriage norms differ across ethnic groups and regions. In Makueni County, most households consist of monogamous marriages. There is more variation in Kisumu County, where wife inheritance is sometimes practiced; in such cases, a man may sometimes have two wives. The number of wives, however, was not a centrally important to one's empowerment. Instead, participants emphasized the role of age at marriage as being more significant in determining one's empowerment. The Luo men's group described how the daughters of the disempowered man were often married at very young ages. The Kalenjin men explained that the least empowered women were often married before beginning secondary school.
In Kaffrine, many marriages are polygamous. While Islam allows for a man to have up to four wives, the number of women a man is married to is ultimately indicative of wealth and is, therefore, considered reflective of his overall empowerment status. Most of the CCD discussion groups in Senegal revealed that the empowered man was expected to have between three to four wives. All participant groups agreed that the intermediately empowered man usually has two wives. The least empowered man was described as having one wife. It was explained that a woman who is the third or fourth wife has married a wealthy man and, therefore, has access to resources which may ultimately empower her. In contrast, a woman who is the first wife is in one of two situations-she has either married a young man and who is not yet independently wealthy; or, she is married to a poor man who never attained enough income to be able to have a second wife. In both cases, this woman experiences economic hardship and does not have as many opportunities for empowerment.
Participants in the CCD discussion groups offered differing thoughts about how polygamous marriages affected women's empowerment. The young women's group discussion in Kaffrine illustrated the most empowered woman as being a fifth wifesomething that is very rare in this region because the man must be significantly wealthy and powerful in order to have so many wives. In contrast, the Malem Hodar young women did not believe that the number of wives a man had was an important indicator for empowerment. The young men, however, struggled to discern whether the first or second wife was most empowered; they deliberated about the roles and responsibilities of each wife within the household and ultimately resolved that the first wife was most empowered.
Overall, marriage structure is important for the empowerment of men in Senegal, often an indicator of his economic empowerment. Previously established wealth can allow for marriage opportunities that further reinforce one's status of empowerment-the ability to have more wives allows for opportunities for more children to contribute to household agricultural productivity and livelihoods. Investment in those children may also return resources to the broader family, which serves to further empower the man, his wife/wives, and his other children. Marriage structure is also an important indicator for women, but in a different way. The patrilocal traditions common in Kaffrine Department often tie women's wealth to the financial assets and stability of her husband and his family. As such, women's economic empowerment is greatly determined by her husband. Additionally, consequently, economic indicators of empowerment may not be determinative of a woman's overall empowerment-particularly because she may have little control over her husband's wealth. Instead, other indicators of empowerment may shed light on a woman's ability to overcome economic disempowerment and transition to alternative pathways to empowerment.

Understanding Empowerment in Transition
One of the more interesting themes that emerged from analysis of empowerment across these three countries was the discussion of transitional factors that led to higher levels of empowerment. This is the final step in CCD activity, where facilitators ask the discussion group what factors or conditions are needed for a person to move to the next level of empowerment (i.e., what is necessary to move the disempowered person to an intermediate level, and an intermediately empowered person to a high level of empowerment?). The original intention of this phase of CCD was to understand the structural conditions necessary for improved empowerment. However, when conducting CCD in each of these sites, we saw that the responses often turned to individual actions that could lead to personal improvement. This seemed to place the responsibility of one's empowerment condition very much on the individual's life choices rather than merely one's opportunities. Consequently, empowerment was viewed as dynamic, and no level of empowerment was depicted as an inevitably permanent condition.
In Makueni County, the conversation returned to the issue of social isolation, as participants discussed ways that the disempowered person could become re-engaged with the community through social groups or church gatherings. The participants also discussed ways in which a disempowered person could work with NGOs to develop skills (such as literacy) to not only improve their life, but also as a way of improving social connections. One woman in Makueni County explained, "Because there are lots of activities in the church which will help her to mingle with a lot of people and she can learn a lot . . . And at the adult education classes, she can learn some reading and writing as well as getting some other information". Individual initiative was also seen as critical to improving empowerment at the intermediate level. Participants discussed the importance of increasing savings and in proper investment in agricultural development to improve crop yields and overall agricultural productivity. In Kaffrine Department, this meant investing in "better seeds"; in Makueni County, this meant investing in improved livestock and farming equipment. For Luo women in Kisumu County, joining community microfinance groups was seen as a path to empowerment. In the Kalenjin community, empowerment could be increased by diversifying crops to include sugarcane. Indeed, in many of the discussion groups-across contexts-overall diversification of income generating activities was seen as a critical step in moving beyond an intermediate level of empowerment. A young man in Kaffrine Department explained that a person "must do farming and some business at the same time". The important message across all these cases is that one's empowerment is, at least in part, due to individual choices and actions. Structural barriers to empowerment were emphasized less.
Still there were some structural barriers that did resonate with communities through the CCD drawings. Across all contexts, educational access was a critical structural component that was considered essential to empowerment. An elderly woman in Kaffrine Department said, "If the children succeed and do well, that can change the situation of their mother". Another agreed, "If one of his children succeeds, he [the least empowered man] will quickly catch up to the [middle] empowered man". As a result, participants repeatedly acknowledged the importance of children's education. An elderly woman noted "Her children, all of them, should go to school, or two go to school, and the other learns a profession". The Malem Hodar participants described how one child's education can transform the empowerment status of the entire household and family. While education was considered a critical pathway to empowerment, in Kenya this was seen as both a structural and individual responsibility to support those who were less empowered. The Luo men's discussion group suggested that a "well-wisher," (i.e., a person intending to do good) might sponsor a child from a disempowered household so that they could complete their education and potentially help transform the empowerment experienced within that household. Development institutions have long set goals for women's education as a means of improving women's empowerment or the education status of their children. However, these findings suggest that some communities also view children's education as beneficial for the empowerment of everyone in the family. Our participants describe a collective process which takes place over generations, but that can be spurred by advancements among members of the family, particularly children. To draw once more on Porter's (2013) terminology, children are regarded as subjects of transformative agency, implicating them as key actors for facilitating mobility at lower and middle levels of empowerment [4].
In Nepal, participants described empowerment as a process that occurs across an individual's lifetime and across generations. Some emphasized that empowerment must occur gradually to be truly transformative. Capturing this idea beautifully, one man said, "By climbing the first step of the stair, you can't reach the top . . . you [would] fall down right away. If you move slowly, then you can arrive here," while motioning to the illustration of the most empowered man. Others described changes in empowerment as a collective process that occurs among families. One woman explained, "Earlier, others at home used to say 'Oh, this and that work is to be done, you cannot go'...Now, I can go somewhere after handing over household chores to people [at] home...". Within the CCD discussion, women and men noted the ways in which changes within her family contributed positive outcomes for her own empowerment.
Lastly, family planning emerged as an important factor when transforming one's empowerment status; however, this was only true in Kenya. In both Kenyan field sites, family planning was considered a critical need for disempowered men and women. In Makueni County, the young men's group explained that those who were seeking to improve their lives needed to start with careful planning around all aspects of life. They explained that, while this requires personal responsibility at some level, it also requires structural support including access to information (educational classes) and contraceptives through ante-natal doctors' appointments. This may reflect the relatively strong presence of healthbased NGOs in Kenya, compared to Senegal and Nepal field sites.

Post-Research Effects
Understanding the mechanisms at work in local contexts can support development programming that is tailored to local relationships and conceptions of power. By bringing local understandings to the fore of development work, the CCD method intentionally brings participants into the transformative process. It asks participants to think about how power can be transformed in their communities. Additionally, because participants often draw their discussions from their own lives or the lives they see around them in their communities, this abstract thought experiment about empowerment has the potential to become self-critical. For example, at the end of our group discussion in Makueni County, a young woman explained to our team, "I really liked this activity. I never thought that I could move to a different level of empowerment, but now I am thinking that it is possible." In parallel, we also witnessed collective engagement with the activity long after we had left the community in Nepal. Three months after the initial CCD activity was implemented, we returned to share results. We presented our findings from that initial activity and asked about current feelings of empowerment since our last discussion. We found that women and men who had participated in that first meeting had been discussing what they had shared, learned, and concluded about empowerment during CCD with others in their households and communities. We found that families were having discussions about empowerment during meals together, and that community members were sharing their thoughts about empowerment with each other while working in the fields. When they returned for this second meeting, many had re-evaluated their conceptualizations of empowerment and moved beyond physical attributes of empowerment (focused on health and physical well-being) to individual consciousness, describing self-confidence as being the key driver of power. The discussion that resulted from the empowerment activity not only gave ownership to the community to conceptualize the term, but also provided the space for participants to discuss what obstacles were preventing empowerment, and what opportunities or resources were present in the community to help them transition to that next level. While experiences of participants engaging in self-critical exploration following participation in CCD were not systematically evaluated-and recognizing these two experiences are insufficient to draw broad conclusions-they do indicate that research methods that engage local conceptualizations may not only yield results and data that allow for improved development programming, but that the process of research itself has the potential to engage participants in transformative process.

Discussion
Findings from CCD in communities in Senegal, Kenya, and Nepal yield interesting and diverse results. Participants were able to share how economic, social, and reproductive aspects of their lives both empowered and disempowered them. These findings, generated through use of CCD with the explicit aim of understanding local conceptions of empowerment, illustrate that differences exist in how empowerment is conceptualized across and within cultures. By engaging these results across contexts, three major findings emerge to inform our larger question about the plurality of conceptualizations of empowerment in understanding food security. These include the existence (or persistence?) of neoliberal conceptualizations of empowerment across communities; the essential cultural backdrop by which results or indicators of empowerment must be interpreted; the importance of understanding mechanisms for empowering change.
Why does this information matter? How does recognition that empowerment varies both across and within cultures shape sustainable development interventions? First and foremost, it illustrates the need for gmethodoloies that investigate local understandings of empowerment. Knowing the local conceptions of empowerment can show what mecha-nisms are inhibiting gender equitable access to resources in the community. Is it poverty? Is it cultural or religious norms regarding mobility? Or is it something else, yet unidentified with existing methods used to describe and measure empowerment? Given the close relationship between mothers' resources and child nutritional and household food security outcomes, understanding the gendered mechanisms that constrain access may lead to the development of more effective interventions. Second, through the process of drawing levels of empowerment, the CCD method can highlight and illustrate to practitioners and scholars defining characteristics and qualities of the most vulnerable sectors of the community who may benefit from development interventions. This is important because vulnerability, similar to empowerment, will vary across contexts and may not be identifiable to someone from outside the community. By describing locally conceptualized levels of empowerment and identifying characteristics of vulnerable populations participants shed light on sub-populations experiencing varying forms of disempowerment, which may or may not overlap, depending upon the defining characteristics of those groups. Third, through the process of identifying transitional mechanisms of empowerment-that which allows one to move from one level of empowerment to another-the community identifies the potential entry points for partnership with institutions. CCD, thus, allows researchers or other development partners to understand the community-identified mechanisms most likely to catalyze change in empowerment as locally conceptualized.
Empowerment, as understood through both the drawings and participant conversations about the drawings in the research sites, can be generally conceptualized as having the ability to sustain a comfortable life and to experience the benefits that accompany this condition: physical health and appearance, education, multiple livelihoods, respect, social networks, and political power, all to varying degrees across sites. These illustrations appear to resonate more with articulations of Chambers' conceptualization, which is rooted in livelihoods, economic welfare, and its accompanying social influence [10]. They also appear to corroborate Galie and Farnworth's (2019) observations of empowerment as relational, mediated by the support of others, and related to community members' assessment of and an individual's alignment with culturally accepted gender norms [34]. There are many conceptualizations of empowerment, some of which are more relevant to the illustrations and narratives provided by these communities than others, and CCD can serve as a means for identifying which theoretical bases are most effective in explaining the cases at hand. We turn to Lutz's (1995) argument that developing unifying theories serves to marginalize the already marginalized, in this case, women. Imposing foreign standards of empowerment on women in developing countries has the potential to further disenfranchise those women [35]. Giving voice to local understandings of concepts can help to inform the scoping parameters of the study and to apply theory that is reflective of women's and men's lived experiences.

Conclusions
The question of empowerment-how women's and men's status shape programmatic goals and outcomes for local communities-is worthy of inquiry for sustaibale development interventions, ranging from food security to reproductive health agendas. However, to develop true gender transformative approaches, it is essential for scholars and practitioners alike to recognize that how empowerment research is carried out, with whom, using what tools and measuring what outcomes, will not only influence the answers they get, but may support a more inclusive vision of empowerment overall. In this article, we identify one strategy that development researchers may take to seek out opportunities for better representing the ideas and interests of local communities. Our findings suggest that such opportunities can be found in prioritizing local knowledge from the onset of our research endeavors.