Pathways to Empowerment:Case Studies of Positive Deviances in Gender Relations in Ethiopia

Development efforts have increased women’s perceived empowerment and freedom, yet have failed to sustainably alter gender norms. There is a lack of research investigating reasons for this anomaly. This study, departing from the conventional approach, tries to fill this gap by employing an interpretative phenomenological approach to assess how women have managed to achieve expanded agency while living within a constraining normative environment. We argue that women have the capacity to deviate and the intentions that lead to new behaviors emerge not only from individuals’ attitudes, subjective norms, and perceived behavioral control, as suggested by the Theory of Planned Behavior, but also in combination with demographic and economic factors. Individuals need to make decisions in three areas ―self-conviction (attitude and perceived behavioral control), subjective norms (within household and community), and structures (state and non-state institutions). The results shed light on alternative empowerment pathways that could potentially inform the design of transformational interventions.


Introduction
Achieving gender equality at all levels is at the heart of development programs and one of the United Nations' Sustainable Development Goals [1]. A handful of research evidence shows that women's empowerment leads to gender equality and helps increase household productivity, which in turn results in improved household nutrition and health, with the ultimate outcome of improved household well-being. Thus, women's empowerment is increasingly at the center of research and development objectives [2][3][4][5][6].
There is a growing recognition among researchers, policy makers and development practitioners that it is the social norms that act as structural barriers to progress toward achieving gender equality worldwide [7][8][9][10][11]. Development literature argues that in many countries, while women have increased their perceived empowerment and freedom, these do not always result in altering the constraining social norms [12]. Studies looking into the reasons behind this are generally lacking. In the past, development approaches that were designed and implemented with the goal to transform the institutional landscape-constraining social norms-were partially successful, failed, or, at the extreme, resulted in reinforcing the existing gender relations [10]. While most empowerment studies look into factors affecting women's empowerment, examinations of the positive side-how a few agents, particularly women, manage to achieve expanded agency while living within constraining normative environmentare generally absent in empowerment research. This study tries to address this gap within the context of mixed and livestock-based systems in Ethiopia.
The gender dimensions of social structures have a critical place in Ethiopian agriculture. Ethiopian rural societies have definitive gender roles that form part of their norms. Previous studies on gender issues in livestock in Ethiopia have revealed that women often have less access than men to agricultural inputs, technologies, market information, extension services, labor, rural associations and economic opportunities [13][14][15][16]. Gender norms related to systems of ownership discourage women from owning assets through which they might acquire more agency and empowerment [14], even though women are primarily responsible for major agricultural activities [15].
Nevertheless, anecdotal reports, for instance Lemma et al. [17], show that constraining norms can be reshaped and possibly ultimately done away with. In Ethiopia, although research on changed gender relations is generally lacking, custom-based gender relations and their associated gender roles in mixed and livestock-based systems are slowly beginning to change [18]. Using qualitative assessments of case studies, the present research contributes to the debate on women's empowerment by shedding some light on existing knowledge of gender contexts, how positive deviance in gender relations occurs, and what the associated consequences are in mixed and livestock-based systems in Ethiopia. This is vital for generating evidence that could be used for the design and implementation of transformative approaches to help change constraining gender norms and bring about long-lasting social transformation. The study seeks to answer the following research questions: • What are the positive deviant behaviors in gender relations and the associated social (gender) contexts in the mixed and livestock-based agricultural systems in Ethiopia? • Why and how do people deviate from the normalized but harmful gender norms and what are the associated impacts in terms of gains in the empowerment dimensions? • What are negative influencing factors in the course of making decisions to deviate?
The remaining parts of the article are structured as follows. The next section presents a review of literature used to construct the study's theoretical framework. Section 2 outlines methodology, beginning with the theoretical framework, before the results are presented in Sect. 3 and discussed in Sect. 4. Section 5 concludes, drawing out policy implications and highlighting limitations of the study.

Literature Review
The present gender bias in economic and social life can be traced back to the accounts of both science and religion as sources for the gender-biased arguments and inequalities between men and women. For example, although years of scientific research on the role of sperm and egg in fertilization appear to support a far more active role for the egg in fertilization than is typically recognized, anthropologist Martin [20] argued that accounts of human fertilization in medical textbooks often pragmatically use gendered language and stereotypes to anthropomorphize representations of the sperm and egg. "Masculine" sperm were depicted as strong adventurers, heroes, and conquerors, actively swimming towards the egg and "penetrating" its defences […].
Likewise, gender language is used in depicting the egg: "Feminine" oocytes, […], are described as passive and fragile, being "swept" along the fallopian tubes to await the arrival of the sperm [emphasis original], [21, p. 1].
Trogen [21] reported this to be just as true decades later. He showed that the same distinction persists in popular representations of fertilization and traces back the origins of these metaphors to ancient times-Aristotle's account of fertilization-that provides a model of social influences underlying modern accounts of fertility.
In fact, culture plays a key role in shaping how we, as biological scientists, describe what we discover about the natural world [22]. One can realize that the picture of egg and sperm drawn in scientific accounts of reproductive biology depends on stereotypes essential to our cultural definitions of men and women rather than being based in biological reality. These depictions reflect broadly held views on the expected roles of men and women in society, and are the result of science's early conceptions of egg and sperm, with the same thinking being reinforced, intentionally or unintentionally, in the modern sciences.
The religious perspective also generally favors men. In the Christian world men are assumed to be the head of the household while women are expected to be submis-sive. This version 1 , by interpreting biblical verses on the husband-wife relationship, is used to maintain the headship role of men in family and society [23]. The headship theory states that distinct gender roles are complementary: men and women are in some ways equal but different, meaning that a man who takes the role of heading household is assumed to be a leader, whereas a woman, who is said to be nonetheless equal, is relegated to a passive role. Equality in this model is only referenced to cover up the real state of inequality and to keep the counterpart in a state of unconsciousness [24] about the fact. Among those church denominations that embrace headship theory, gender-based violence and abuse is very common [23]. Similarly, the gender relations in the Islamic world do not differ much. Although ssociety-oriented Islamists believe that Islam assigns women special rights and that individual women and men have equal value, domestic responsibilities (the unpaid care-work) is assumed primarily as the women's role while the role of generating the family's income is ascribed to men (20). Patriarchy and Islamic dogma are highly linked and often reinforce each other. Structure of authority and command has a central place in Islamic doctrine [25].
Feminist theory suggests that gender hierarchy which reinforces women's submission often results in their sexual objectification [26], which in turn leads women to adopt a submissive role and promotes beliefs in men's supremacy among the whole society. Although the causes of men's motivation for dominance over women are not yet conclusively established, social learning theory [27] and cognitive development theory 2 [28] suggest that it is a learned behavior adopted through an intellectual process via socialization. Socialization processes are guided by social norms. Social norms largely form the context within which women and men pursue their interests [29]. Although functioning independently, they are often interlocked with societies' formal institutions including the institutions of the state, the market, and intra-household bargaining dynamics that shape the day-to-day choices of men and women [11,30]. Together they form the opportunity structure of an agent within which women and men pursue their empowerment pathways.
The opportunity structure according to Narayan [30] is influenced by three factors-the permeability of institutions, the unity and behavior of powerful groups, and the state's implementation capacity [30]. The permeability of institutions (both formal and informal) refers to the degree to which an agent is able to influence these institutions (government policies and social norms) in their favour. The unity and behavior of powerful groups, the second dimension of opportunity structure, refers to the extent to which the strength and ideology of elite groups can hinder or facilitate the exercise of agency by subordinate groups. For example, if there is fragmentation 1 '[…] as women were striving for equality, Christian men were afraid of losing their power. They came up with a theory to keep women in submission: employ religion for the purpose. In order to make this plausible to believers, headship theory was described as a biblical model' [23, p2]. 2 Social learning theory explains some men's need for sexual domination of women through the premise that individuals develop notions of gender and associated behaviors by watching others and mimicking them. This learning is then reinforced vicariously through the experiences of others. Cognitive development theory suggests that gender-related behavior is an adoption of gender identity through an intellectual process, implying that misogynistic behaviors can be identified, remembered, and mimicked by subsequent generations of males [28]. and weakness on the side of the dominant group in terms of their ability to oppose, repress, or neutralize the claims of challengers, agents can exercise their agency to pursue their goals. The third dimension "[t]he state's implementation capacity refers to the effectiveness with which government authorities carry out policies that have been adopted" [30, p 48]. These factors are what constitute contextual conditions and opportunities, within which actors pursue their interests, while being shaped in a way that is in line with the ideology of the dominant group in the society. The World Bank report on norms and agency revealed that persistent gender gaps lie in the set of factors embedded in social and gender norms [6].
Constraining structures are made up of such social and cultural norms, including the mechanisms, rules, and procedures that regulate social interactions outside the functioning of the state [6]. That is why, in a practical sense, the emphasis of the empowerment framework is not only on supporting agents to have greater access to a stock of resources and opportunities, but also to address structural obstacles so that more equitable gendered outcomes become possible. It is the rules and normsintangible enabling or disabling sources-which are important to exercise power, and which serve to demarcate the boundaries of choice for different categories of individuals by forming their opportunity structure [31].
The gender norms which form the basis for gender relations, comprising the "differential rules of conduct for women and men" [32, p35] are shaped by individual behavior as well as social institutions [35]. They set roles to be played by women and men in agriculture, and more importantly, give rise to gender-differentiated capacities to access, own, and manage assets [34] and are largely responsible for the gender gap in use of opportunities between men and women in agriculture [35]. Moreover, they expose women and men to different levels of risks [36]. They define relational patterns within households and communities, and they impact allocation of decision-making responsibilities, as well as access to information and other important resources [37]. Systems of access, ownership, and control of resources vary greatly across contexts, influenced by gender norms that determine the meaning and dynamics of resource allocation [38]. Their contextual and fluid nature poses great measurement challenges in social studies [32,39].
Gender relations are often varied and constantly fluid, but there are almost always hegemonic notions of femininity [40] within a given cultural context that shape women's roles and relative positions. Inequality in power emanating from attributed roles and embedded power relations is sustained through day-to-day practices. At its most extreme, it is sustained through gender-related violence. Nevertheless, these power relations and hegemonic notions of femininity and women's place are always challenged and transformed by counter-hegemonic gender relations and behaviors [41].
Similar to most developing countries, Ethiopian women have low levels of empowerment across all domains. As a result, their participation and roles across different dimensions of sustainable development is by far minimal [8]. Despite the government implementing reforms at the macro-institutional level, traditional norms still rule over most aspects of life in both rural and urban areas, meaning that the average, poor, Ethiopian woman is disadvantaged and in a low social position [11]. These traditional norms are strong, particularly in relation to gender, and are embedded in the day-to-day lives of men and women, thus, limiting women's abilities so that they cannot take development opportunities as easily or frequently as men. Limited government capacity to enforce the laws that are in place to protect women sustain the problem. The strong nature of the existing informal rules and norms, which are also embedded within formal institutions, present severe obstacles to ongoing efforts to change women's status within Ethiopian society [11,42].
The gender norms which govern how men and women behave, have a detrimental effect on women's capacity to act (agency). It is one of the primary factors that creates important differences in the ability of women and men to learn, adapt, and make use of innovations in livestock agriculture in Ethiopia. Specifically, gender norms set roles to be played by women and men in agriculture and, more importantly, such norms give rise to gender-differentiated capacities to access, control, and manage assets [34]. For example, a recent study in Ethiopia demonstrated gender inequalities in the division of labour in livestock production, with women being more involved in labour-intensive activities, while men are more involved in income management and decision-making aspects [43,44]. However, the bulk of currently available evidence on the interplay of agency and gender norms in Ethiopia comes from research relating to crop-based farming systems among traditional societies [45]. Hence, much is unknown about the dynamics of the interplay between agency, gender norms and livestock related innovation processes in the non-crop-based system in Ethiopia. Globally, the World Bank has conducted extensive work on norms and agency, also with little emphases on livestock [12].

Theoretical Framework
The most commonly used socio-psychological model to understand how human intentions transform into specific behaviors is the theory of planned behaviour (TPB). The theory stresses the importance of attitudinal components -attitude, perceived behavioral control, subjective norms -in explaining and predicting behavior. To this effect, many authors believe that the use of the TPB as a framework for studying behavior can explain much of the intention as well as the future behavior [19]. In the PTB, intention refers to the determinant or immediate act of the specific behavior of an individual. In the current study, we define women's (men's) intention as her(his) desire to engage in positive deviant gender relations. As mentioned above, the intention consists of three psychological constructs: Attitude refers to the degree to which execution of a behavior is evaluated positively or negatively by the agent him/herself. Subjective norms, also known as referents, are the perceived social pressures from significant others to engage (or not engage) in the behavior. Perceived behavioral control refers to the individual's own perceived capability to successfully perform the behavior. Together, these lead to a positive or negative intention to perform a given behavior. A positive or negative intention towards a given behavior is formed when these three constructs were combined and commonly understood to be when they are in more favorable conditions [19].A rich body of literature has examined the relationship between one's economic conditions and gender attitudes. For exam-ple, increased feelings of one's relative economic deprivation make both men and women significantly more likely to support women's paid employment, implying that relative economic insecurity can essentially trigger support for women's economic participation [46]. While economic growth can empower women [47,49), women's empowerment may at the same time prompt economic growth as well [48]. It is also argued that gender gaps in employment can shrink among the poor during economic recessions, as poor women enter the labor force while poor men's labor force participation remains unchanged [50].
In Ethiopia, suggested drivers of change in gender relations in mixed livestockbased system include changing practices in rural markets, political interventions coupled with awareness, and changes in socio-cultural settings [18]. Similarly, evidence elsewhere suggests several factors as driving forces for one's engagement in challenging and reshaping existing norms. These include exposure to women leaders [51], serving with women in the military [52], participation in paid employment, mother's participation in the labor force [53], having sisters and daughters [54,55], exposure to urban life and television programs that present deviations from traditional gender norms [56,57], interventions that raise the overall aspirations of both women and men [46], and open discussions about gender equality among adolescents [58].
In the present research, by integrating feminist theory and contemporary research on norms within the social psychology literature-particularly the Theory of Planned Behavior [19] -we derive two hypotheses. Firstly, we hypothesize that even living within a constraining normative culture, social contexts that hinder positive deviances, women (and men) possess the capacity to deviate from existing norms as a result of internal and external driving factors. Secondly, as depicted in Fig. 1, we hypothesize that the intentions leading to a given behavior emerge not only from one's attitudes, subjective norms, and perceived behavioral control but also in combination with demographic and economic factors, when adopting positive deviant practices in gender relations.  10 The Council of Europe defines sexism as any expression (act, word, image, gesture) based on the idea that some persons, most often women, are inferior because of their sex [59].

Context of the Study Areas
The study was conducted in four target sites of the CGIAR (Consultative Group on International Agricultural Research) Research Program on Livestock (Livestock CRP) in Ethiopia across three main regions, namely Oromia, the Southern Nations, Nationalities and Peoples' Region (SNNP), and Amhara regional states, located in the west and south west, south, and north of Ethiopia respectively ( Table 1). The agroecology and production system characteristics of the study areas range from dryland to wet highland with pastoral, agropastoral, and mixed crop-livestock systems. Both Elweya and Abergele districts fall in the lowland agroecology and pastoralagropastoral production system [43,60] where farmers are largely engaged in rearing livestock in rangelands and producing crops on fertile land at a small scale [61]. Whereas, Horo and Adiyo districts are located in the highland agroecology which is characterized by a mixed crop-livestock system where livestock husbandry and rainfed cropping are closely interlinked -crops and livestock support each other to produce optimum output and spread risks [62].
Similarly, the study districts differ in their socio-cultural and economic profiles in several ways. On the one hand the population in Horo and Adiyo is composed of different ethnic groups with different religions mainly Orthodox Christianity, Protestant Christianity, and Islam. Marriage systems in these districts include both monogamy and polygamy. Largely, although the status of women in Ethiopia is lower than that of men-meaning women are less educated, poorer, earn less; their contribution to agriculture is less valued; and they do not have equal decision-making power [63] -women in this region decide quite autonomously about their earnings while their husbands dominate decision-making in all other areas of life. On the other hand, the populations in Abergele and Elweya districts are homogeneous and belong to Amhara and Oromo ethnic groups, whose religion is mainly Orthodox Christianity and traditional belief, respectively. The marriage system in these areas is relativity monogamous. Household decisions are often taken jointly with a relative dominance of men and properties are shared within the household [64].

Method and Data Source
The current paper is developed based on a qualitative study approach, snowball sampling technique -a network research that presents numerous advantages in registering "hidden populations" such as people with deviant behaviors [65]. Data was generated in 2019 through step-wise stages addressing the gendered contexts of empowerment resources 3 within which positive deviant gender relations are practiced, and why and how individuals or households decide to deviate from the normalized but harmful gender norms in rural Ethiopia. Firstly, the gender research team conducted a preliminary desk review in order to develop tools and checklists for focus group discussions (FGDs), and key informant interviews (KIIs) with the various research participants. Secondly, stakeholders 4 consultative workshop was held using an ongoing training workshop event in order to identify community representatives for FGDs. Thirdly, following the consultation meetings, community-level FGDs were held with representative men and women in separate groups. Apart from collecting information on gender contexts, the FGDs were also used as a platform to identify individual community members who are engaged in desired behaviors in gender relations that are assumed inappropriate according to community norms (i.e., positive gender deviant behaviors). Finally, in-depth individual interviews were held with these women and men in order to generate data on positive deviance in gender relations and the associated decision factors, challenges, and empowerment outcomes.
For the current study data generated from selected community members and Livestock CRP partners through the above processes was accessed from ICARDA 5 and analyzed in order to understand the gender contexts of empowerment resources and positive deviant behaviors in gender relations. The FGDs with community representatives collected information on the gendered contexts of issues such as access, ownership, control of resources, participation, and decision-making in the study areas. Case studies with 18 women and men were conducted to explore the driving factors for adopting the changed gender relations, their decision factors, and the obstacles to their engagement in such practices.
To explore the subjective experiences and views of women and men in positive deviant gender relations, the study draws on an interpretive phenomenological approach. This qualitative methodological approach is designed to "focus on personal meaning and sense making in a particular context, for people who share a particular experience" [66, p45]. The study relied on women and men community members and key informants for exploring and understanding their experiences in positive deviant gender relations, particularly why and how individuals decide to deviate from the normalized gender norms. The rich contextualization and description of their experi-3 Empowerment resources refers to all the socio-economic endowments that women (men) value as useful assets in the process of their empowerment pathways. 4 These stakeholders were kebele administrative officers; CBBP facilitators; staff of Regional Research Centres; staff of District Cooperative Promotion Offices; staff of District Livestock Agencies; staff of District Women, Children and Youth Affairs Offices; and staff of District Offices of Agriculture and Natural Resource Management across the study areas. 5 ICARDA is the International Center for Agricultural Research in the Dry Areas. It is one of the CGIAR centers and participant in the Livestock CRP in Ethiopia. ences in positive deviance in gender relations was the focus in adopting a phenomenological perspective for this study.
Generally, literature suggests that between 5 and 25 individuals who have experienced the same phenomenon should be interviewed in a phenomenological study [67]. Accordingly, through purposive sampling technique, 18 women and men community members across the study areas were identified for in-depth interviews. In phenomenological research a purposive sampling technique is recommended to provide insight into the specific experiences of respondents [65]. The primary inclusion criterion for identifying respondents for the in-depth individual interview was that they were engaged in practicing positive deviance in gender relations. Moreover, the 6 FGDs and 12 KIIs (with 7 community leaders and 5 experts) were also used to generate information on gendered contexts. For identifying participants in the FGDs and KIIs, their community representation in terms of gender, marital status, wealth status, social class, and duration of stay in the community were considered.
Using open-ended questions, in-depth individual interviews were conducted to capture the experiences of these women (and men) community members, as well as with community elders as key informants. The process of data collection allowed us to discuss the interviewees' experiences in depth and the interpretations they provided with regards to the research questions [68]. Although varied in length, the interviews took 1-2 hours on average, and all were audio recorded with the oral consent of the respondents. The study was approved as per ICARDA's Code of Conduct Policy for research. The study methodology is depicted in Fig. 2.

Analytical Techniques and Variables
Drawing on the Theory of Planned Behavior [19], the study adopted a framework for decision-making as an analytical approach. Before analyzing the qualitative data, the recorded interviews were translated and then transcribed. Thematic analysis was used to analyze the data [69]. After familiarizing ourselves with the content by repeated reading of the transcribed data, meaning and patterns were searched and initial codes generated describing what was said by the interview participants. In the process, we noted the ways in which language was used and performed conceptual interpretation of the data. We searched for patterns throughout the data and linked the emergent themes together, which were again reviewed, defined, and renamed. Hierarchical thematic structures such as subordinate themes, subthemes, and emergent themes were developed through this strategy [4], guided by NVivo software.
The framework for decision-making to adopt changed gender relations assisted the organization of results with respect to how decisions are made, using a three-tier design which includes decision areas, decision factors, and decision attributes [70]. Factors from prior research findings were considered to guide the identification of factors and attributes for the current research. Each factor and attribute were examined to determine its appropriateness for the adoption of changed gender relations ( Table 5). The preliminary findings were developed into a poster, and shared with stakeholders and the wider CGIAR research community, and feedback was received and synthesized 6 .

Findings
We first highlight the gender context across the study communities covering systems of access, ownership, control of resources, decision-making, and valuations of gender roles. We then present findings related to positive deviance in gender relations with a particular emphasis on the nature and types of deviant behaviors, and decision factors for adopting deviant but desired gender relations. This is followed by an analysis of influencing factors in the course of adopting the deviant behaviors in gender relations. The section concludes with a presentation of the associated empowerment outcomes realized as a result of engagement in the deviant practices. 6 The poster of preliminary findings was shared at the CGIAR's 2020 virtual annual research review and planning conference and can be accessed at https://cgspace.cgiar.org/handle/10568/109206.

Access, Ownership, and Control of Resources
Household and community resources that constitute the preconditions for empowerment are not equally accessed, owned, and controlled by men and women in the study areas. The gender aspects of social norms play key roles in mediating women's and men's rights to these resources. In general, community members believe that being part of a given household and community provides the right to have access to resources belonging to the household and community at large. However, detailed probing reveals that this is not always the case. There are household and community properties for which gender mediates the right to access, own, and control resources. At the household level, for example, cash savings and specific farm tools such as shafts, axes, and wooden boxes are owned and controlled by men spouses. This was evident in the women-only FGDs: I cannot cultivate my farmyard using oxen even if I wanted. Because my husband will not allow me to use his farm tools. Because he does not trust me, assuming that I do not have the required skills.
Similarly, community properties such as protected areas are often out of the reach of women. Community elders and kebele administration, usually composed of men only, construct informal or formal bylaws which restrict women's access to such resources.
In recent years, marginal lands are being protected in order to restore vegetations, believed destroyed through improper use by some community members [referring to the women community members, who most often go to these areas for the collection of fire-wood and other forest products]. Nowadays, we [referring to the men] have developed bylaws to ensure that.
-Men-only FGD participant, Horo district, 2019. Often, as in this case, the process of setting up such community-based institutions excludes women who are the primary users of the resources. As a result, their interests and priorities remain unaddressed.
When community members are asked about their awareness of their legal entitlement to acquire, administer, control, use, and transfer properties, both men and women suggest that they are aware of their legal rights. However, what is being practiced in reality is that men control these bundles of rights, although women often state that they have equal rights with men. Women also suggest that these rights are exercised through their husbands.
I do not worry much about who owns and controls resources as long as we [meaning her husband and herself] are in love relationship. I always feel represented in community associations through my husband.
-women-only FGD participant, Adiyo district, 2019. This is a part of community norms and implies that, in practice, social norms outweigh what is stated in the legal code regarding property rights for women and men. Although the Government of Ethiopia has taken positive steps towards achieving gender equality with several laws enacted governing asset sharing in marriage, giving stronger inheritance and property control rights to women [71], in reality, what is governing these bundles of rights is the norms. This can be seen clearly in the examples provided in Table 2.
Key gender issues in agriculture Access to household resources De jure Women have the right to acquire and use property. They have equal rights to men with respect to use of land. Women shall also enjoy equal treatment in the inheritance of property [72, Art. 35 (7)].

De facto
The right to access some household assets such as finance (cash) and certain farm tools is limited for women spouses. Similarly, the right to access community resources like protected communal forests is limited for women. The processes of crafting informal and customary laws to protect communal resources often exclude women and is generally gender blind. The right to own assets is seemingly the same for men and women spouses and often this right is described as 'joint ownership' in economic studies. Yet FGDs with famers revealed the masked dominance of men in systems of 'joint ownership'. Moreover, in most cases women spouses are not allowed to accumulate their own assets above certain thresholds.

Control of household resources De jure
An owner spouse (men or women) can freely dispose of his/her personal property-which includes the power of selling, exchanging, or donating the property to somebody else by one's free will and volition (73, Article 59].

De facto
Although most household assets are said to be jointly owned and controlled, the right to transfer, change, or market is dominated by the men spouses, who hold the final say.

Decision-making on key household resources De jure
Women have equal rights to men with respect to administration and control of properties [72, Art. 35 [7]]. De facto It appears that decisions affecting the household are made jointly; however, without men's consent women spouses do not have the right to make the final say, even on their own assets.

Decision-Making and Participation Status
Decision-making at the household level was consistently reported as a joint task across the study sites. Both men-and women-only FGDs repeatedly suggested that key decisions that affect the household are taken jointly based on discussions or consultations. Further probing, however, revealed that whenever there may be disagreements, the men hold the final say. In this case, for both gender groups, joint decision-making does not refer to anything beyond the fact of being informed, which gives one the sense of recognition, or being consulted. This involves not only recognition but also the sense that one's ideas are valued as positive contributions to the issue at hand by the head of the household, who is usually the man. However, the results of the FGDs with both genders suggest that men, because of their relative advantage in terms of exposure to information sources, assume themselves to be better decision-makers. For example, in the pastoralist community in Elweya district, seeking information on livestock management, marketing, health, and the like is culturally designated as a man's role.
The men generally assumed women to be incapable of making major decisions affecting the household. When asked why women do not have the final say one of the participants in men-only FGD participant at Elweya asserted saying: "[…] women are not good in decision-making because they lack patience and the ability to articulate things". The men described women as passive, caring more about emotional and relational issues whereas they described themselves as aggressive and leaders by nature. In pastoralist communities, these are believed as truths and generally understood as given by God. When women assumed any of the culturally designated roles of men-ploughing, participating in producer associations alongside men spouses, assuming headship/leadership roles, taking up public duties, giving final word in decision-making, and owning property-they were often mistreated by their spouses and men in their work-place. Men's responses against women, for instance, in leadership manifested in many ways-including harassment, beatings, and abusive sexual relationships-and this was widely noticed in work-places and public spaces.
When women are promoted to leadership positions, the men colleagues generally do not welcome them happily because these men think that they are placed in positions subordinate to these women at the work-place and their dominance is challenged. They manifest this perceived behavior in many ways including the use of words that undermine women's status, by engaging in the act of harassment, and at extreme level by counteracting her work to undermine and finally remove her from her position.
Such behaviors are manifested in response to perceived gender equality by men at all levels. In Elweya district, a men-only FGD participant asserted the perception using a local expression in the Oromiffa language: "walqixxummaan dubartii halkan geesse hanqatti", which roughly translates to "gender equality becomes short of equality at night". In elaborating the saying, he said that naturally the man was made to be a leader, and thus even if equality is happening in the day-time, at night the man's upper hand is naturally restored.
Social norms generally limit public spaces for women. Across the study sites, women's mobility is highly constrained. Social norms do not allow free movement of women such as visiting relatives, market places, and community events unless accompanied by men spouses or elder men in the family. If a woman is found doing such activities without the consent of her husband or men family members, she will be assumed to be behaving inappropriately or considered as arrogant. Similarly, participating in community associations and speaking in public gatherings, which are traditionally assumed as men's, especially married men's, roles, are taken as unacceptable behavior and often met with nicknaming that discourages women from such practices.
The division of labor along gender lines exists widely across socio-economic and farming systems with respect to domestic work and agricultural practices in the study areas. While domestic activities are entirely considered as women's roles, major agricultural activities such as ploughing, house construction, marketing of high-valued commodities and livestock, and community affairs are considered masculine; women's involvement in such practices is systematically discouraged. Similarly, men's participation in domestic activities is met with nicknaming and considered shameful. FGDs revealed, nevertheless, that women's contribution in agriculture is immense and takes much of their time. They are usually involved in the most labor-intensive aspects of farm operations from land preparation (hoeing and clearing) to transportation of farm products, yet this involvement goes unrecognized and unvalued.

Sexism and Valuations of Gender Roles
Sexism manifests in all aspects of social life, including language and communication. Examples include, as often noticed during FGD discussions, the generic use of masculine terms such as "he/him/his" by speakers to refer to an unspecified person. Moreover, the naming of a woman for her strength using words such as "wendi nat" -meaning "mannish" -is evidence of biased against women. These are statements used consciously or unconsciously to reinforce the dominance of men in a society.
The findings from the qualitative assessment assert this fact. A women-only FGD participant from Elweya district in south west Ethiopia reflected "we [the women] are created to assist the men". During the discussion, the women FGD participants argued that although they are over-burdened with domestic, livestock, and agricultural works compared to their men counterparts, they think that the most labor-intensive tasks-digging the water well and trekking livestock to watering points, which are relatively labor intensive-are carried out by the men. They tend to value such activities more than activities in the women's domain-cooking, child care, fetching water and fire-wood, managing small, pregnant, and weak animals left behind, and all other household chores-not because these activities are less labor intensive, but because societal norms assume them to be so.
We cannot ask men spouses to share work from us even when he is sitting and we are over busy with activities within the home.
-Women-only FGD participant, Elweya district, 2019. The women believe that what are traditionally assumed to be appropriate roles for women are their responsibilities to deliver in order to conform to the prevailing beliefs and prove themselves as a "good woman" in the eyes of men and society at large.

Background Characteristics
The majority of the 18 individual livestock keepers who were identified by community as practicing some forms of changed gender roles in agriculture and other aspects of their livelihoods are married or widowed women ( Table 3). Most of them are literate, with an average age of 39.8 years. The family size, on average, is 5.7. The majority of them are heads of their households, implying that even most of the married women are assuming the role of leadership in their families, which is uncommon. According to the community's wealth status rating, all of these respondents are classified as rich or medium, and they are members of at least one or more social institutions such as cooperatives, iddir 7 , and saving associations. Almost all reported experiencing an increase in access to services, information, asset ownership, and participation in decision-making as a result of adopting changed gender relations opposed to the normative culture. Seemingly, this is what maintained their motivation to continue practicing such activities in spite of discouraging factors, whether internal, family-related, or external peer pressure.

Nature and Types of Deviant Behaviors Identified
Women's subordinate status in Ethiopia generally forces them to avoid open selfinitiated actions. Nevertheless, the study identified a number of deviant but desired behaviors in gender relations. They are being practiced by a few individual women (men) in the study areas. The behaviors being practiced against normative gender relations, along with the driving factors and descriptions are summarized in Table 4. Table 5 summarizes the key decision areas that an agent needs to make in order to engage in positive deviant gender relations. In connection to that it also presents the factors and attributes associated to the key decision areas that motivate women and men to engage in the positive deviant practices reported in Table 4. A three-tier analysis for decision-making resulted in three major decision areas -self-conviction, subjective norms, and structures -appear to be necessary for individuals in order to deviate ( Table 5).

Influencing Factors in the Course of Adopting Deviant but Desired Behaviors in Gender Relations
Individuals who chose to follow a different course of livelihood activities against nor-   Loss of husbands to death and divorce as a result of various factors, and prior exposure to urban life, leading women to start engaging in activities traditionally ascribed to men, such as house construction, cultivating commercial crops, and any other income generating activities regardless of constraining gender norms in order to support their families and make a living. The burden of raising children after divorce, often left to women in most communities studied. Women speaking in public arenas Demographic factors (better schooling and progressive family) Autonomy and a supportive family environment at an early age. Growing up in a family where female children are encouraged to actively participate in family matters. A relatively longer stay in school compared to the rest of the community providing the courage and ability to speak out and participate in community affairs affecting one's own life. Shared decision-making and participation in social associations Demographic factors (schooling) A higher schooling and literacy level on the side of husbands. Husbands and wives consulting and jointly making decisions that affect the household. Added value of the wives' involvement in decisions affecting the family, leading the husbands to encourage wives to take part in producer associations together with themselves. Women participating in public duties (community management activities)

Combination of intrinsic motivation, economic, and demographic factors
Aspiration for a better wage as a result of some level of schooling combined with loss of a husband and a need to overcome life challenges. Resulting active participation in community and school administration, hoping that this would pave the way to a civil service position.
Women exhibiting pro-active information-seeking behavior Combination of intrinsic motivation, economic, and demographic factors The desire for higher income through use of technologies. Better social capital as a result of some levels of schooling, enabling openness to technologies, innovations, and new practices. Frequent experience of being used as a model and compared with men by fellow women and men farmers and extension agents.

Men participating in domestic activities
Combination of exposure and demographic factors Removal from one's comfort zone and favored social environment through becoming a soldier. Exposure to new practices such as cooking combined with better literacy. Resulting lack of shame in actively taking part in domestic activities. Women playing a role of community elder Combination of intrinsic motivation, economic, and demographic factors Expectation of getting hired at community level. Formal schooling leading to a search for formal work in the community. As a result, willful engagement in community management activities such as school administration and governance expecting that one day this will pave the way to a formal job.
Women involving in income-generating activities regardless of their nature, for example, house construction Economic factors (deprivation) Extreme poverty leading to family break-up, which imposed the burden of raising children alone due to divorce. Need to be involved in any income-generating activities for the sake of survival and to raise-up children. Change of location to escape cultural restrictions and negative reactions from relatives, peers, and community members and be able to engage in activities considered culturally inappropriate for women.

Decision areas Decision factors
Decision attributes Description for decision factors Self-conviction (attitude and perceived behavioral control) Economic freedom Economic advantage Struggle for economic equality through engagement in more productive activities that are usually outside women's reach because of gender norms. Striving for survival Engagement in a new sphere of practices considered masculine in order to sustain life, particularly after loss of men spouses.

Related gain Prestige
To achieve desired-for recognition-a sense of being respected and proud for being able to appear in public with men. Agency Act of revealing one's capability of engagement in highly rewarding jobs (such as commercial activities, for example livestock trading and house construction), although culturally inappropriate for women.

Self-determination
Desire to attain the power of control over economically important assets. Visionary Passion for change through one's own efforts and a strong belief in oneself. Perceived behavioral control

Manageability
The degree to which new practices are simple to implement, particularly for practices that require specific skills. Subjective norms (within household and community)

Family and community readiness
Readiness for change Family and community openness to accepting and encouraging changes that are against established but constraining norms. Progressiveness Community progressiveness due to gendered awareness interventions and other factors.

Relationship with others
Level of social capital and engagement in services that one has in one's community. Family and community support

Normative relations
The capacity or the extent to which reactions from internal and externals relations (family members, peer pressure, and social networks) as a result of deviation from existing norms are tolerable. Self-control Restraint from norm-based dictates (for example nicknaming) that discourages others from deviating and engaging in self-driven practices.

Commitment
Family and community commitment toward supporting changes. Collective agency General progressiveness of the public/community members as a result of education, Christianity, etc. Table 5 Decision areas, factors, and attributes associated to deviant behaviors in gender relations, quantitative assessment in 2019, rural Ethiopia them especially during the early stages of their experience with engagement in positive deviant behaviors. These influences included refusal from family members, peer pressure, and risks of being isolated from neighborhood and social networks, ranging from informal to formal associations. Almost all of the individuals interviewed were ridiculed and nicknamed for practicing such activities because these were believed to be against the existing norms. However, all suggested that these kinds of attitudes do change gradually when people observe the positive results of such behaviors on the lives of practicing individuals and their families. Nevertheless, they argue that drawing others along after themselves is never an easy task, as people tend to respect existing social norms more than the evident benefits of these practices. "Because, the benefits often take a longer time to be realized", said a woman from Horo district. Another woman shared a similar experience. She is one of the women farmers who actively participate in public meetings, extension events, and community management activities. Whenever I go away from home to participate in community events, most women think that I am a lazy woman or do not have work at home. They used to ridicule me saying arrogant.
-Woman, Adiyo district, 2019. In Abergele, a woman reflected on how she does the work of house construction which is normally considered to be masculine. Both men and women used to discourage her by saying "chikashin aboki", which means "do your muddy work". When she would climb to the roof of the house for assembly, both men and women used to say "imsishin indaywegash" meaning "be careful not to damage your reproductive organ while climbing up and down the wall". These are examples of words of discouragement mainly from the women according to the woman from Abergele district in 2021. Another woman also said that she was often nicknamed by her fellow women whenever she would speak in public during her early days of life in the community as a newcomer upon marriage. She would be tagged with different names and, at times, even insulted. They used to attach these insults to her marriage status, as she is the wife of a rural primary school teacher. Women often nickname her saying "wondiye" -characterizing her with a masculine identity assumed inappropriate for a woman. She also shared her experience as a member of a credit and saving association: Autonomy-supportive environment, for example, due to increased exposure to media that support more gender-equal attitudes and practices. Community incentives Words of encouragement from educated segments of society and the general tendency of young generations to be more liberal.

3
Since I often took part in community-level meetings and spoke up in public spaces, my iddir excluded me from membership reasoning that I failed to attend regularly, although I know that their main reason was my deviation from their way of thinking, existing norm, and practices.

Empowerment Outcomes
Individuals who shared their stories for the current study have witnessed positive gains in all dimensions of their empowerment-economically, socially, and politically. Their engagement in the new behaviors enabled them to have better access to information services, asset accumulation, and active participation in decision-making. Women who managed to take part in the "roles of men" such as plowing, harvesting, and threshing also developed the capacity to pro-actively seek information on improved agricultural technologies and market services. They are able to break social barriers and regularly contact extension agents in spite of their gender to access extension services. Life challenges provide the courage to overcome constraining norms. When a participant in a women-only FGD was asked "what makes you different from the rest of the women in your community?", she replied: In our community it is a norm to believe that if a woman ploughs land using oxen, the land will not be productive. This discourages us [the women] to engage in such activities considered masculine. However, I do not bind by such sayings anymore, because I want to change my life by engaging in any activities that bring me more money. I think this is what made me different from the other women in this community.
-Woman, Horo district, 2019. By actively participating in the farming of cash crops and other income-generating activities such as house construction, these individuals increased their income and were able to reinvest in more productive farm businesses. Their financial status and exposure to service providers further capacitated them to recognize their rights within the household and community. They are actively taking part in decision-making processes and participating together with their men spouses or elder men in social associations.
The social benefits participants realized from practicing changed gender relations were immense. For example, engagement of men in domestic activities not only significantly reduced women's domestic burden but also enhanced trust, respect, and love between couples and increased family harmony. Beyond that, women feel proud, more confident and valuable when they see their spouses engage in domestic chores and when they themselves are able to engage in men's roles as well. Apart from that, the extra time saved as a result was used to take part in rural associations which in turn expanded their connections (networks) and exposure to valuable information.
Finally, women (and men) who were able to overcome the effects of social norms were found to be active in local politics. They spoke in public, representing fellow women (men) members of the community, and actively engaged in community affairs. They developed the capacity to articulate and present their interests and priorities as well as those of other women. Their active engagement in public spaces helped them to take part in local administration and school councils. Moreover, community reputations developed as a result of these processes enabled such women to play the role of community elders, performing the work of mediation to resolve disputes.
These days, I am a respected woman among, at least, literate people and community administration council members as a result of my active engagement in school and community administrative activities.

Discussion
In this article, we explored the positive deviant behaviors in gender relations, including why and how people deviate from the normalized but harmful gender norms, along with the associated impacts in terms of gains in the empowerment dimensions. We integrated feminist theory and contemporary research on norms within the social psychology literature to guide and examine the data obtained from ICARDA along two main axes, (1) gender contexts of access to resources, and (2) agency. We also examined the outcomes as a result of engagement in positive deviant gender relations. In this study we show the gender context within which women (men) exercise their agency. We argue that the intentions that lead to given behavior/s emerge not only from one's attitude, subjective norms, and perceived behavioral control according to the TPB [19], but also in combination with demographic and economic factors that play an important role in adopting desired behaviors in gender relations. Three decision areas-self-conviction (attitude and perceived behavioral control), subjective norms (within household and community), and structures (state and non-state institutions) -were found to be linked to the practices of desired behaviors in gender relations. Systems of access, ownership, and control of resources appear to be similar for community members across the study areas. Although, an individual's status with regard to empowerment resources at the household and community levels is generally determined by one's belongingness to a given household or community, gender was found to be mediating the right to access, ownership, and control of resources. The gender relations and aspects of social norms generally favor men [15] and outweigh the legal codes with respect to these bundles of rights. This could be partly as a result of the nature of gender norms-interlocked with the formal institutions [11]. Nevertheless, it is argued that property rights give women greater status and leverage [74].
Similarly, decision-making, an important aspect of empowerment, is in turn found to be determined by an individual's status with respect to these rights. The more one gains property rights, the more they control decision-making power. Increasing evidence suggests that ownership status-who owns the assets within householdsaffects a range of household decision outcomes [75]. Gender-biased social norms are also reported as the primary reasons for the exclusion of women from the processes of crafting institutions meant to manage communal properties [13]. Across the study sites, particularly in the livestock-based systems, men's privileges as a result of the existing norms in accessing information that help them make better decisions com-pared to their women counterparts are further presented as a proof of, and leveraged to reinforce, men's existing dominance. Men's apparent superiority in making better decisions is actually linked to their access to information because existing norms favor them. Similar research has shown that who within a household receives information affects the outcomes of household decisions [75].
Conversely, this study has shown that in cases where women are able to achieve some level of empowerment, men often engage in repressive behaviors that further discourage women from pursuing more empowerment pathways. Men's desire for maintaining the status quo within the household, community, and institutions not only serves as a blockage to progress toward gender equality but also a cause for gender-based violence against women. Feminist theory suggests that gender hierarchy that reinforces women's submission often results in sexual objectification of women [76] which in turn leads women to adopt a submissive role and promotes belief in the supremacy of men across society. Social learning theory [27] and cognitive development theory [28] suggest that men's motivation for dominance over women is a learned behavior adopted through an intellectual process and could be labeled as cultural misogyny 8 that men develop unconsciously. Individuals develop notions of gender and the associated behaviors by watching others and imitating them. This process and learning are then reinforced vicariously through the experiences of others [27] and intellectual processes through which misogynistic behaviors can be recognized, extracted, and mimicked by subsequent generations [28]. Therefore, breaking the chains by discouraging negative stereotypes and encouraging desired behaviors is essential to improve women's overall condition in the study areas.

Nature or Nurture?
The common beliefs among communities that men are naturally endowed with good decision-making abilities and are natural leaders proved to be incorrect upon close investigation. Such beliefs, rather, stem from women's masked inability to make effective decisions which in reality is attributable to the gender roles that society ascribes to them (Kinati et al., forthcoming). The gender norms burden women with roles and do not give them the opportunity to develop their decision-making skills, while they also diminish their sense of competence [77]. Women's minds are tied up with fulfilling societal expectations and have no time to collect and process information in order to make effective decisions. This masked ineffectiveness of women in decision-making is often taken as evidence to prove that men are better decisionmakers and generally assume productive roles with higher social values [78]. The same argument is also used to prove that women prefer and are happy taking on domestic and less valued roles. Nevertheless, evidence shows that women do this in order to fulfil the roles that society expects them to play, implying that the account of Freudian women masochism 9 is wrong [79].

Gender Division of Labor Devalue Women's Roles
It is evident that gender roles systematically devalue women's roles in agriculture. The reasons why the works that are performed by women are often less valued is that they are not among the major household income sources. Crop and livestock farming, in which men dominate, is recognized as the household backbone upon which the entire family depends, although women contribute much to these. The norm is that men are believed to be breadwinners simply because they hold the plough and till the land [80]. On the other hand, the less valued activities are predominantly performed by women and the argument that these occupations are less valued only because they are assumed easier, require less physical strength and skill, and often remain unremunerated [81] completely disappears upon closer scrutiny. For instance, house construction is a role of men in the mixed farming system but of women in the pastoralist system. In Horo, constructing houses is a prestigious job, but in Elweya they view it differently. Being a house builder (carpenter) in Elweya is stereotyped as a domestic responsibility and assumed culturally as a post that suits women. In this area it is not a highly valued role. This implies that there is a direct relation between valuations of roles, who does them, and related gender stereotypes.

Gender Roles Dictate Choices
The other common beliefs among communities are that women are assumed to prefer domestic activities whereas men are assumed to prefer productive roles and are also assumed to be better performers than women. These gender-based division of labor are then used to prove a point that men and women are good at different things and have different preferences. For example, the assertion that men are better at and prefer productive activities done outside the home affirms the idea that men are created for these careers and women have an inbuilt preference for staying at home and nurturing families. Nevertheless, that is simply a social construct and untrue, despite being a common perception among communities. Gender roles are adopted and nurtured as a result of learned behaviors. The differences that we observe between men and women in agricultural practices and social lives are the result of nothing more than gender roles dictated by the gender aspects of social norms [32][33][34]. People learn from society at an early age that some behaviors will be rewarded with praise, and others punished with ridicule as they grow up [28].
Gender roles pose great challenges to women's empowerment. The problem is not simply that these roles are often unequal; rather, they are actively harmful and oppressive to certain social groups, particularly to women. An example from the findings in this study is the stereotype that women are less competent in making major household decisions. Research has shown that it is their work load that diminishes women's sense of competence [77] because of the fact that women spend, on average, up to five hours more on work per day than their men counterparts in the study areas [15]. Yet it is the prevailing belief that women are incompetent decisionmakers, and this discourages women from pursuing a path that might empower them simply because they get the message early in life that they are not appropriate for that role. As a result, women live a life of subordination and bear the burden of gender bias while being systematically forced to remain unconscious about it [24]. Women are still vastly underrepresented in productive activities with empowering effects. This has disastrous consequences, because not only are women excluded from paths that empower, but also as a society we cannot reach our full potential for innovation in agriculture while systematically excluding half of the population.
Nevertheless, although rural Ethiopia is highly patriarchal-men dominate public spheres while women are assumed to be subordinate and are expected to stay at home and be responsible for domestic activities-and this continues to shape women's opportunities because of the persistent values that influence their decisions. However, women may develop different strategies for personal fulfillment and autonomy despite existing gender-based constraints [82]. We found that women are able to accommodate the power of social forces and retain the capacity to act decidedly against constraining structures.

Overcoming Gender Identity
Even when the normative climate is not supportive of autonomy, several factors can be drivers of one's engagement in positive deviant behaviors to overcome the forces of social norms. Apart from attitude, subjective norms, and perceived behavioral control [19], and formal structural factors, one's demographic and economic conditions were found to be driving forces of an intention to deviate.

Economic Factors
When economic dependency is lost as a result of the break-up of a marriage, the loss of a partner to death, and/or a household's extreme poverty, women (and men) jump over boundaries of social norms and engage in what they find to be productive. Evidence also suggests that increased perceptions of one's relative economic status affect gender attitudes. A study conducted at the household and community level on how perceptions of one's relative economic status affect gender attitudes has shown that increased feelings of relative deprivation make both genders significantly more likely to support women's participation in social and economic activities [83]. This result suggests that relative economic insecurity of individuals can actually prompt support for women's participation in productive roles such as paid employment.

Demographic Factors
Family background and progressiveness of households, combined with being educated, were found to be driving factors for women (and men) engaging in positive deviant behaviors in gender relations while living in a patriarchal society. Autonomysupportive family, both before and after marriage, could be another important driving factor for individuals to engage in positive deviant behaviors. Formal schooling and increased exposure to media in recent years in Ethiopia could play further vital roles in the adoption and cultivation of desired gender attitudes by "breaking through the walls of patriarchy" [84, p18]. Men spouses with some level of education and exposure to 'foreign' practices often do not strictly abide by prevailing patriarchal norms and customs [83]. Spouses of such men actively participate in community affairs by assuming roles, such as community leadership and school administration positions, as well as the roles of an elder, which are commonly considered masculine according to community norms.
Couples who have been able to enjoy autonomy supportive family background are able to share every household decision and participate together in social associations and community affairs. Women spouses are able to exercise their agency by, for example, speaking in public spaces whether their men spouses are present or not. When women achieved some levels of education and their family environment is autonomy-supportive, they can exercise their agency and eventually challenge existing constraining social norms [85]. Progressive families, therefore, could serve as the basis for supporting individuals to develop such behaviors. Individuals can also develop notions of gender and the desired behaviors by watching and mimicking those with whom they interact [27].

Intrinsic Motivation Related Factors
Having the capacity to aspire combined with the desire for betterment also motivated individuals to engage in positive deviance in gender relations in spite of internal and external constraining factors. The degree to which execution of the behavior is evaluated positively before individuals engage in the desired practices against the normative context in which they live might play an important role [19]. Aspirations and the desire for higher income forces individuals to take up the roles of the other gender when they can expect these to provide them with better gains. For example, women were attracted to take up the house construction role, which is culturally assumed to be a man's role, as the business was highly paying in their area.
Individuals who have experienced changed but desired gender relations often took measures in order to avoid the negative impacts arising from one's depart from a subjective norm. As a coping mechanism they often left their locality and migrated to other areas to escape people in their social circle after they decided to start new desired behaviors that were not socially ascribed to their gender. By doing so, they avoided perceived social pressure from significant family members and others with whom they had grown up within a particular social structure. This implies that pressure from family or peers within their social circle, in the form of misogyny, was considered less tolerable in comparison to similar pressures from people in the destination areas.
Perceived own capability to perform successfully the planned new behavior, in most of the cases studied, is backed by their own economic and demographic conditions-for example, the struggle for survival combined with educational attainment and exposure. Apparently, the intentions that lead to a given behavior emerge not only from one's attitudes, subjective norms, and perceived behavioral control, as formulated by the Theory of Planned Behavior [19], but also in combination with demographic and economic factors that play an important role in adopting deviant but desired behaviors in gender relations. Individuals, being motivated by their demographic factors-better schooling and autonomy-supportive family environment-and economic conditions-sudden loss of spouse, fear of extreme poverty, and exposure to foreign practices-were sometimes guided to overlook attributes of planned behavior with regard to whether to take up the roles of the other gender. Therefore, as depicted in Fig. 3, these demographic and socio-economic factors are important constructs that need to be added to the socio-psychological factors proposed by TPB when studying an intention to deviate in gender relations.
The perceived locus of causality for engagement in the desired behaviors studied appeared to be mixed-self-determined and controlled (external to self). Behaviors such as women speaking up in public arenas; participation in producer associations, assuming public duties, decision-making; pro-active information-seeking behavior; and playing a role of community elder are self-determined and the individual persons perceive the locus of causality as internal to themselves. Whereas behaviors such as women's taking part in the "roles of men" in agriculture and house construction activities are externally motivated and thus perceived as external to self because individuals were forced externally to make such engagements.

Conclusions
The gender context across the study areas appears to be generally non-supportive of women's autonomy. Systems of access, ownership, and control of resources are dominated by men. The existing gender roles dictate choices and devalue women's roles, posing great challenges to the efforts being made to realize women's empowerment. Nevertheless, interviewed women (men) found to have the capacity to overcome the forces of social norms under various conditions. It appears that the intentions that lead to given behavior/s emerge not only from one's attitude, subjective norms, and perceived behavioral control according to the TBP, but also in combination with demographic and economic factors that play an important role in adopting desired behaviors in gender relations. Individuals or households need to make decisions in three areas-self-conviction (attitude and perceived behavioral control), subjective norms (within household and community), and structures (state and non-state institutions) -in order to engage in the practices of desired behaviors in gender relations.
Practices of changed gender relations are mostly challenged by external factors which include peer pressure, risks of being isolated from their neighborhood by both men and women, and loss of social networks, ranging from informal to formal associations. These are often exhibited in the forms of discouraging words received from Fig. 3 Factors that lead to positive deviant behaviors in gender relations peers (women and men alike), neighbors, and members of social associations, mostly experienced during the initial stages of new practices. The relative fluidity of the normative environments across the study areas set the context for women to exercise their agency to take important decisions and engage with deviant but desired behaviors and other opportunities for the betterment of their lives. Polices that promote autonomy supportive family environments, expansions of adult education and advocacy programs against misogyny through existing rural associations could strengthen and encourage community members engage in positive deviant gender relations.
This study is different from others on women's empowerment in that it attempts to fill the knowledge gaps in empowerment by looking at deviant behaviors in gender relations. It sheds light on the various empowerment pathways that could be used as inputs to the design and interventions of transformative approaches. The selfdetermined behaviors identified could potentially be used for the design of transformational interventions. Those individuals engaged in these behaviors might serve as role models and efforts to support such champions could play an important role in unlocking new approaches to gender transformation. When a critical mass of such individuals is achieved, sustainable social transformation can be realized. Hence, it is recommended that a public extension approach should consider ways to include them.
Some limitations are apparent in this study. First, the study works with different ethnic groups with different religious backgrounds. However, this is not well captured in the dataset on which this study was based and thus detailed analysis on how ethnicity and religion interact to shape context specific norms is not analysed. Second, the interpretive ability of the researchers might be somehow limited due to lack of physical access to the emotional state of the research participants as the analysis was based on available datasets. Interpretative phenomenological research is more robust when the researcher be able to make connections between people's talk (linguistic) and their thinking (cognitive) and emotional (physical) state. Lastly, the information presented on how men performed deviant behaviours and their own experiences in positive deviance gender relations is limited in the dataset used for this study and thus not discussed in detail in the paper.

Statements and Declarations
Ethics Statement This study is based on data collected from men and women farmers targeted by a small ruminant breeding program led by ICARDA and although ICARDA does not have ethics committee, the study considered all the potential ethical issues associated with conducting field study with human beings. Oral permission was granted from the respective local administrative bodies of the relevant district government offices. Following that, before undertaking the study, an informed consent from the people to be researched was obtained which involves an oral consent to participate in this study. In all the study sites, informed consent from the prospective informants and group discussants was obtained only after in-depth discussions with the prospective participants were done. Likewise, permission was also obtained from participating individuals and groups in order to take notes during the case studies and for taking audio recordings of conversations in order to comply with the local cultural beliefs. All the research participants were assured that any private issues and information gathered would be confidential and not be disclosed without their consent to any individuals, including their spouses, or organisations.