Developing gender-transformative innovation packages for sustainable intensification: the case of maize leaf stripping in northern Ghana

Abstract

Sustainable agricultural intensification (SI) seeks to address multiple development objectives at the same time, among them social aspects of sustainability. However, interest in gender norms as potential catalyzers or obstacles for achieving these objectives has remained low. In this article, we use a SI assessment framework as an analytical lens for the identification of social aspects—in particular inequitable norms—that could be targeted through gender-transformative innovation packages. We evaluate social science data from a maize-livestock intervention in Ghana. In a mixed methods study, data were collected from 60 farmers involved in experimentation. The analysis unearthed a number of overlapping norms that shape men’s and women’s actions and interact with the SI package to produce tangible inequalities. To promote gender equity, these norms need to be targeted. But it will not suffice to simply add gender-transformative components to technical packages, if the technical components are not redesigned to fit the same equity goal. We showcase this using the results. An important conclusion is that assessments of both social and technical components of interventions must be more consciously designed in transdisciplinary processes—with equitable arrangements in mind.

Keywords:

• Gender norms
• sustainable intensification
• innovation assessment
• interdisciplinarity
• Ghana
• transdisciplinarity

Introduction

Sustainable agricultural intensification (SI) aims at accomplishing various development outcomes at the same time. Its aim is to raise productivity for food and feed without harming the environment, while paying attention to social, economic and nutritional objectives. Whether the two terms, sustainability and intensification, are actually compatible has been under debate since the inception of SI thinking. An initial emphasis on productivity and the environment was later supplemented by other objectives to embrace a broader concept of sustainability (Garnett & Godfray, 2012). However, consideration of gender issues and various other inequalities that SI interventions may (re)produce remains weak (Himmelstein et al., 2016). Loos et al. (2014) state that without specific concern for social justice agricultural intensification cannot claim to be sustainable. Looking at the diversity of SI objectives, one may ask how assessments of interventions and practices have integrated gender equity.

Gender equity is frequently defined using the analogy of a playing field. Measures to ensure a fair process level the field for women, men, girls and boys and constitute gender equity. Gender inequity (as the opposite) denotes social disadvantages that have historically evolved and prevent the different players from operating from the same starting positions. Inequality is a result. The successful removal of disadvantages leads to gender equality (Darmstadt et al., 2019). Gender norms as expectations of how men and women should behave in a particular society or group play an important role in the process of equity. Often, they restrict women’s access to resources and limit their power and voice. Challenging and transforming these norms—where they are harmful—can contribute toward greater gender equity (Cislaghi, 2018; Cislaghi & Heise, 2020). Interventions that engage women and men in critical analysis and renegotiations of inequitable norms have been termed “gender-transformative” in contrast to “gender-accommodating” interventions that acknowledge but do not challenge such norms and “gender exploitative” interventions that take advantage of them (IGWG, 2017).

To return to the question of how diverse SI objectives (including gender equity) can be integrated in evaluations: arguably the most comprehensive effort to move from measuring single indicators toward a holistic assessment of innovations has been made by Musumba et al. (2017a). For the smallholder context, they present the Sustainable Intensification Assessment Framework (SIAF), “a framework of objective-oriented SI indicators organized into five domains critical for sustainability: productivity, economic, environment, human condition, and social domains” (Musumba et al., 2017a, p. 3). An innovation’s objectives are identified and indicators linked to it to assess performance across several domains. This happens at a pre-adoption stage when farmers are involved in experimentation. A precondition for meaningful use of SIAF is interdisciplinary cooperation.

SIAF’s social domain includes the indicators “gender equity” and “equity (generally)”, the latter dealing with equity across specific social groups such as those differing in their livelihood strategies. The underlying “equity framework”, briefly sketched in the SIAF guide (Musumba et al., 2017a) and its accompanying manual (Musumba et al., 2017b), draws on Kabeer’s (1999) conceptualization of empowerment and its three dimensions (resources, agency and achievements). The authors translate Kabeer’s empowerment dimensions into “equity metrics” that measure differential access to resources (land access, livestock ownership), differential levels of control over resources as agency (management control, market participation, time allocation), and gendered differences in benefits from agriculture as achievements (income, food security, health status). Access to information as capacity constitutes an additional metric. Overall, Musumba et al. (2017a, 2017b) leave open how the concepts of empowerment and equity relate to each other. Kabeer (1999) sees empowerment as a “process by which those who have been denied the ability to make strategic life choices acquire such an ability” (p. 435), a process which is characterized by an interdependence of structural and individual change. The comprehensive character of SIAF would be further strengthened if the influence of structures (for instance, norms and policies) on individual preconditions for and outcomes of agricultural experimentation could be incorporated into assessments. In this paper, we take up this last point. We adapt SIAF and its social domain to bring into focus the underlying causes of gender inequality, the inequitable norms that shape the performance of innovations and their sustainable benefits. With regard to agricultural research for development programs in general, Kantor et al. (2015, p. 297) write: “Not addressing the barriers created by existing norms and attitudes (…) limits the enduring nature of development outcomes.” Innovation packages tend to consist principally of technical measures with few or no consciously designed social components. By contrast, pioneering gender-transformative studies (Cole et al., 2020; Kantor et al., 2015; Mulema et al., 2020) attest to how changes toward more equity amplify the positive outcomes of concomitantly offered technical components. Much earlier and on a broader base, Leeuwis (2004) had written about the need to deal simultaneously with the technical and socio-organizational dimensions of innovations to arrive at viable solutions. This view was echoed in a recent expert panel report (Barrett et al., 2020) on bundling social and technical innovations to increase synergies in development outcomes (although with no reference to Leeuwis).

SIAF was co-developed by the Africa Research in Sustainable Intensification for the Next Generation (Africa RISING) program. The study presented in this paper was conducted under its West Africa project. Africa RISING works with a multi-stakeholder, farming systems and action research approach. A variety of SI technology packages are developed with farmers through experimentation and assessment. Since 2018, the use of SIAF has become mandatory for all Africa RISING research activities.

In this article, we use an adapted gendered version of SIAF that is in line with its holistic, participatory and interdisciplinary approach. In the first step, the framework serves as an analytical lens for the identification of social aspects (in relation to technical aspects) that constitute catalyzers or obstacles for the objectives of the selected innovation. A specific focus is on gender norms. Building upon this in the second step, we develop recommendations for the redesign and combination of technical and social components in gender-transformative socio-technical innovation packages. In a third step, the results will inform development practitioners who intend to work on the same or similar innovations. It needs to be emphasized that analysis, recommendations and results relate to the particular socio-cultural context of the case study with its specific gender norms.

We evaluate social science data from the assessment of a SI package in 12 communities of the Northern, Upper East, and Upper West regions in northern Ghana. The package was arranged around a crop-livestock innovation (maize leaf stripping) and did not contain consciously designed gender innovations. Some of the biophysical aspects of the intervention have been outlined in earlier separate publications (Hoeschle-Zeledon, 2019; also, Komarek et al., 2021 with a limited regional focus; Abdul Rahman et al., 2022). They will be included in this presentation wherever relevant. Bundling the results of the biophysical assessments with the results of this social science inquiry at the preadoption stage shall inform the scaling of context-specific gender-transformative SI packages.

Materials and methods

Conceptual framework

Participatory action research and farming systems research, which underlie the Africa RISING program, share a concern for social justice, critical reflection and collaboration with development participants. For that reason, they are especially suited for further integration with approaches that seek to transform inequitable gender relations. However, interventionist studies of this kind also require researchers to engage in ethical reflection on their own work to improve their accountability to farmers (Maguire, 2005; Wong et al., 2019). Accordingly, this article describes a review and learning process. We take the absence of consciously designed gender components in the leaf stripping intervention as a starting point to ask how an adapted version of SIAF can facilitate the future development of gender-transformative innovation packages. Following Leeuwis (2004, p. 12), we define innovations as “not only composed of novel technical devices or procedures, but also of new or adapted human practices.” Technical aspects of packages are never void of social dimensions but tightly interwoven with them. Gender norms play an important role in shaping practices and innovation processes. They can be understood as a sub-set of social norms that outline what a specific group expects of a woman or man. Notions of appropriate gender behavior are internalized, embedded in institutions and produce constraints (Cislaghi, 2018), often depending on how gender interacts with other social criteria. Gender-transformative approaches challenge norms that create poverty and inequalities (Kantor et al., 2015) and seek to innovate social arrangements.

SIAF (Musumba et al., 2017a) provides a basis for comparison between innovations and common practices in five domains (Table 1). Interdisciplinary teams have used the framework for investigations in various contexts with emphasis on different SI objectives (Abdul Rahman et al., 2020; Fischer et al., 2018; Snapp et al., 2018). For the leaf stripping intervention outlined in this paper, the research and author team consisted of social scientists and biophysicists.

Table 1. SIAF framework: examples of indicators related to each domain (Musumba et al., 2017a).

Environment	Productivity	Human domain	Social domain	Economic domain
Vegetative cover Plant biodiversity Erosion Soil chemical quality	Crop productivity Animal productivity Input use efficiency Yield gap	Nutrition Food security Human health Capacity to experiment	Gender equity Equity (generally) Social cohesion Collective action	Profitability Income diversification Returns to land, labor and inputs Labor requirement

To identify social and more specifically gender-transformative components of innovation packages, we have developed a conceptual framework (Figure 1) for the adapted SIAF. It is marked by the following features:

Figure 1. Conceptual framework.

First, the restriction of equity issues to the social domain (Table 1) is lifted. Gender analysis, and more broadly equity analysis, is woven through all domains. Gendered institutions underly men and women’s opportunities to produce, attain food security and sustainably manage natural resources (Badstue et al., 2020; Meinzen-Dick et al., 2014). For a livestock innovation, Fischer et al. (2018) show how all domains can serve as entry points for gender analysis. They state that insufficient gender integration into the entire framework puts researchers at risk of pursuing a multi-disciplinary instead of an interdisciplinary approach. In Figure 1, gender analysis is removed from the social domain and forms a separate assessment level for all domains. Furthermore, each domain serves as a specific lens for the investigation of the whole, counteracting as much as possible a fragmentation that demarcated domains may invite.

Second, the conceptual framework (Figure 1) distinguishes between the immediate interaction between the core innovation (in a SI package) and its users and the enabling or constraining context. For instance, the interest of farmers in a forage chopper may be limited not only by its design (immediate interaction) but also by the lack of adequate storage facilities for chopped feed (constraining context). This distinction creates space to carve out how the core innovation can be adapted and bundled with supportive innovations in the context. Immediate interactions between an innovation and gender relations often manifest themselves in imbalanced participation of men and women in interventions, inequitable benefits from innovation use, and sometimes in their differential preferences. However, these are merely “symptoms” of underlying gender norms that constitute part of the enabling or disabling context. Access to resources, labor arrangements and household decision-making (among others) are shaped by norms (Badstue et al., 2020; Kantor et al., 2015). Equitable participation in co-design and experimentation, as well as sustainable and fair benefits from innovation processes, can only be achieved if restrictive norms are transformed. In Figure 1 we distinguish between two levels of general assessment that correspond to two levels of gender analysis.

Third, the conceptual framework is grounded in a mixed methods approach with close attention to farmers’ ways of knowing and assessing, as well as their demands. Investigations of innovations bundles from a gender lens should consider measurable as well as less measurable aspects of sustainability. In terms of the latter, qualitative methods and their epistemological focus on exploring respondents’ realities play a prominent role.

If farmers’ preferences are captured numerically and linked to a narrative description of how respondents themselves evaluate their experimentation (following criteria potentially different from those scientists use), both emic (insider, local) and etic (scientific) interpretations are taken into account to arrive at context-relevant solutions. Leeuwis (2004) and Hecht (2018) have discussed the complementarity of emic and etic approaches in innovation processes. With respect to sustainability indicators, Scerri and James (2010) recognize the importance of quantitative indicators for measuring progress over time, but view qualitative approaches as equally necessary to account for structures of power, values, norms and interests in targeted communities. Community members’ negotiations on what needs to be changed, and to whose benefit, are left out where the achievement of sustainability is seen as being only a technical issue (based on statistical information). The limitations of purely quantitative methods and the importance of qualitative methods have also been acknowledged for identifying and understanding gender norms (Badstue et al., 2018; Behrman et al., 2014). Actors’ understandings of empowerment and their diagnosis of technical and social constraints are fundamental for shaping effective transformation (Galié & Kantor, 2016).

Small ruminant keeping, maize farming, and the core innovation

Approximately 70% of Ghana’s local goat and sheep meat production comes from its northern region, where smallholders tend to combine crop cultivation with small ruminant husbandry (Amankwah et al., 2012). For medium- and low-endowed households, small ruminant keeping forms part of livelihood diversification strategies aimed at reducing risk and increasing resilience. Animals are mostly kept in low-cost/low-input systems. For many farmers market-oriented production with higher input costs is not a viable option. Crop farming remains the mainstay of livelihoods, with maize as the dominant staple crop. Important cash crops are cowpea, groundnuts and soya beans (Amankwah et al., 2012; Assan et al., 2018; Konlan et al., 2014; Kuivanen et al., 2016).

A participatory assessment with farmers identified feed scarcity as one of the main constraints to raising livestock productivity (Konlan et al., 2014). Studies attribute a decline of natural pasture to urbanization, growing livestock numbers (Konlan et al., 2018), shifts from communal to individual land tenure (Avornyo et al., 2020), bush and forest burning (Amankwah et al., 2012; Assan et al., 2018), harvesting of firewood for sale, climate change (Assan et al., 2018), and intensified land use during the rainy season as a shift from a fallow system to annual cultivation (Kuivanen et al., 2016). Next to pasture, crop residues form an important component of the animals’ diet. These residues are also sold in emerging feed markets, especially in the dry season (Konlan et al., 2018). The intervention evaluated in this paper was aimed at addressing feed scarcity through the increased use of maize as a dual-purpose crop, more specifically through an innovation called maize leaf stripping.

Maize leaf stripping is defined as the removal of the lower leaves of the maize plant to provide feed for livestock during the cropping season without compromising on grain yield. Stripping can be done at two stages: 50% tasseling, when half of the plants in a plot have their tassels (male flowers) open, and 50% silking, when half of the plants in a plot have their silks (female flowers) open. The fresh leaves supplement feed in a period in which the animals are mainly tethered or herded in order to keep them from invading cultivated plots (Lukuyu, 2015). The innovation assumes that maize leaves are underutilized as feed (left in the fields after harvest) or fed only late in the cropping season. An evaluation of biophysical studies on maize leaf stripping is contained in Komarek et al. (2021). To our knowledge, no gender evaluation of this innovation exists.

Methodology and sampling

In the context of Africa RISING’s action research approach, collaboration with farmers plays an important role. To ensure participating farmers’ interest in collaborative experimentation the project partly uses self-selection sampling (as a non-probability sampling method) for its work in target communities in northern Ghana. At sensitization meetings, scientists and technicians introduce farmer groups to agricultural innovations. Names of farmers who express an interest in and fulfill preconditions for experimentation with specific innovations are listed. Thereafter, groups determine trial participants in a ballot. The described sampling process was applied for the leaf stripping study. The 90 group-selected participants not only keep livestock and farm maize but also showed commitment to a process of training, knowledge exchange and trial evaluations on their land. The paper focuses on the experiences of these farmers from 12 communities in Savelugu, Tolon, Kassena, Bongo, Nadowli and Wa West districts in the Northern, Upper East, and Upper West regions of Ghana who experimented with the maize leaf stripping package for one or two cropping seasons between 2017 and 2019.

Collaborative research was initiated by biophysicists and led by an interest in: 1. the effect of leaf stripping on the grain and fodder yield of maize, and 2. the use of stripped maize leaves as livestock feed supplement in the cultivation season (Abdul Rahman et al., 2022). Leaf stripping was offered as a SI package consisting of the innovation itself, improved seeds and fertilizer, and good agricultural practices (spacing during planting, band placement and covering with top soil as fertilizer application method, and two weeding regimes). Using a technology park and farmer field approach, farmers, agricultural extensionists and project technicians established and maintained the experiment. It consisted of a randomized completed block design with three treatments and farmers’ fields as replicates. The treatments were: 1. no maize leaf stripping, 2. maize leaf stripping at tasseling, and 3. maize leaf stripping at silking (Abdul Rahman et al., 2022). The cost of inputs used for the trial was shared between farmers and the project. Farmers provided labor and land for the establishment of the trial with a minimum land size of one acre. The project provided improved seed, pesticide for fall armyworm control and fertilizer. Members of the interdisciplinary research team appraised input provision by the project differently: the biophysicists regarded it as an important measure to acquaint farmers with SI packages. The social scientists, who joined the team at a stage when the biophysical trials were already ongoing, saw it as potentially amplifying biases in discussions with farmers. For this reason, the social scientists suggested an evaluation in two rounds: the first round after a season in which support was still given, the second after the following season in which input provision would have ceased (after a limited two-year experimentation period).

The first gendered SIAF evaluation of the SI package took place in 2019. From the overall intervention sample of 90 farmers (67 men and 23 women experimenting with leaf stripping) we purposively chose a sub-sample of 21 women and 39 men (60 respondents) for the gender assessment. Sampling of the latter was guided by efforts to maximize the number of women respondents and to gather sufficiently large groups for focus group discussions (FDGs). The gender sample covers all but the two districts of Bongo and Wa West, which were excluded for their low number of women participants. In a strict sense, the results apply to selected groups only and are non-representative. The study combined quantitative and qualitative approaches: in gender-separate FGDs (three with women, five with men), farmers discussed the innovation package in relation to the five SIAF domains. Additionally, they established activity profiles for the division of labor in maize cultivation and livestock feeding. A short concomitant survey with the same participants captured demographic information and general farming systems, as well as gendered household arrangements. In 2020, the same 60 farmers were revisited for a second round of FGDs (four with women, seven with men). Key informant interviews were held with five project technicians (all men) who had worked closely with the farmers during and after the trials. Both rounds of data collection resulted in a total of 19 FGDs. Audio-recordings from FGDs and key informant interviews were transcribed verbatim and translated into English. Using Atlas.ti, transcriptions were coded in a qualitative content analysis process. Descriptive statistics (mainly frequency distributions) were generated from the survey data set with STATA.

Description of gender assessment sample

Overall, more men than women were involved in the leaf stripping intervention. This is also reflected in the gender assessment sample (39 men versus 21 women), which is a sub-sample of the overall intervention sample (see “Methodology and sampling”). In what follows, all data refer to the gender assessment sample.

In terms of gender proportions, 64% of all men and only 14% of all women came from the Northern region (Table 2) which is predominantly Muslim. In contrast, there is a larger Christian population in the Upper West and Upper East regions. Restrictions on women farming in their own right are more pronounced in the Northern region as compared to the other regions and associated with socio-religious norms (Apusigah, 2009).

Table 2. Sample description (n = 60).

		Men	Women	Total
Region	Northern	25 (64%)	3 (14%)	28 (47%)
	Upper West	7 (18%)	9 (43%)	16 (27%)
	Upper East	7 (18%)	9 (43%)	16 (27%)
Religion	Muslim	23 (59%)	3 (14%)	26 (43%)
	Christian	9 (23%)	16 (76%)	25 (42%)
	Traditional	7 (18%)	2 (10%)	9 (15%)
Marital status	Married	38 (97%)	14 (67%)	52 (87%)
	Divorced	1 (3%)	0 (0%)	1 (2%)
	Widowed	0 (0%)	7 (33%)	7 (12%)
Seasons participated	One season	5 (13%)	2 (10%)	7 (12%)
	Two seasons	34 (87%)	19 (90%)	53 (88%)

The mean age in the sample is 47 years (for both women and men). Only seven farmers were 35 years or below. These were four married men from the Northern region and three married women from the Upper West with otherwise no conspicuous characteristics compared to the rest of the sample. In terms of formal school attendance, 25% of the respondents had an average of eight years of education with no pronounced gender differences. While almost all of the men (97%) were married, one third of the women (33%) indicated that they were widowed. The sample is made up of 43 respondents (72%) from men-headed monogamous households, 11 respondents (18%) from men-headed polygamous households, and six respondents (10%) from women-headed households (all of the latter in the Upper West and Upper East regions). All men (39) in the sample indicated that they were heads. On the part of the women, 14 reported that they were the spouse of a head, six women headed their household, and one was the mother of a male household head. The average household size was 10 with a minimum of two and a maximum of 40 persons. Polygamous households were on average larger (mean 16 members) than monogamous and women-headed households (mean nine and seven members respectively).

Most of the farmers were involved in the intervention for two seasons (87% of the men and 90% of the women). They cultivated and stripped maize on a one-acre farm. Average land sizes available to households ranged from one to 12 acres (average five acres). This availability must not be read as general accessibility of land for women’s farming activities. With predominantly male land ownership, farmers in the study regions cultivate on two types of plots: (1) So-called household plots are mainly under the control of the man head. These plots often cover the demand for staples and cash crops. (2) Smaller plots, on which wives or other household members farm, have more pronounced individual control of yield and income. The allocation of the second type is often temporary and depends on the man head’s decisions (Michalscheck et al., 2020). Respondents most commonly kept small ruminants. Goats were reared by 82% of the men and 76% of the women. Sheep, often having a higher value in the market, were kept by 87% of the men and 43% of the women. Numbers for other livestock (Table 3) and more detailed information are available in the results subsection “Economic domain.” The focus of this paper is on small ruminants.

Table 3. Livestock ownership.

	Men (n = 39)	Women (n = 21)	Total (n = 60)
Local cows	12 (31%)	2 (10%)	14 (23%)
Draught cattle	2 (5%)	0 (0%)	2 (3%)
Sheep	34 (87%)	9 (43%)	43 (72%)
Goats	32 (82%)	16 (76%)	48 (80%)
Pigs	6 (15%)	9 (43%)	15 (25%)

Results

This presentation of the results deals with the SIAF domains one at a time. Domain-related aspects served as broad entry points in the first research round (Table 4) and were consistently shaped by what respondents revealed during discussions.

Table 4. Overview: SIAF domains and aspects dealt with in the study.

Environment	Productivity	Human domain	Social domain	Economic domain
Feed components Soil management	Livestock productivity Maize productivity	Food utilization Food availability Capacity to experiment	Collective action Natural resource conflicts	Resource access, ownership and decision making Labor relations, roles and burdens Perceived profitability

Environment

In this domain, we investigate feed components and inquire whether farmers perceive maize leaf stripping as an alternative to open grazing or cutting vegetative matter as feed. This relates to the broader question whether a shift toward feed cultivation could enhance adequate vegetative cover in rangelands. We also ask how leaf stripping interacts with the use of crop residues for other purposes (such as for soil management).

Feed components

For the dry season, farmers reported a large variety of feed sources including crop residues, browse shrubs and processing waste. However, in most cases these only moderately supplement the feed requirements of small ruminants that are otherwise roaming free. In two FGDs (one among men in the Northern region and one among women in the Upper East), respondents perceived a conflict between their wish to keep animals closer to home to avoid theft or loss, and the financial and labor constraints in respect of stall-feeding animals or building appropriate pens. Open grazing of small ruminants was also associated with the increased risk of ingesting plastic waste. Several women from the Upper East shared that they had lost livestock due to plastic ingestion, a problem that was confirmed in its extensiveness by their project technician.

In the rainy season, confinement of animals or keeping them under close supervision is necessary to prevent them from invading cultivated areas. They are herded or tethered for grazing, partly on community land. This is a task often given to young boys who also cut grass and wild browse plants for feed. Respondents characterized their feeding system in the cultivation period as “mainly grazing with some stall feeding” (35%), “zero grazing” (27%), or “mainly stall feeding with some grazing” (22%). “Grazing only” (tethered or herded) was least common (15%). There were no pronounced gender differences, except for the fact that eight out of nine “grazing only” responses were by men. More than half of the men respondents (54%) and 38% of the women respondents said they purchased some feed during the rainy season. In men-headed households (MHHs), the heads most often (75%) paid for feed. Only in one community were fodder trees cultivated. Especially men spoke of walking long distances to find sufficient grass and suitable trees in the wild (see labor allocation in Table 5). They framed this as labor-intensive social competition for feed, and made no reference to the potential pressure the reduction of vegetative cover would cause on the environment.

Table 5. Activity profile for maize cultivation and livestock feeding activities (MHHs only).

	Household labor
	Strong involvement	Weak involvement	Hired labor
Maize cultivation
Dibbling (creating holes)	Men	Other household members
Seeding	Women	Other household members	Women, girls
Fertilization	All household members
Weeding	Men, boys	Women, girls	Men, boys
Harvesting: breaking cobs	Women, men		Women
Harvesting: gathering cobs	Boys, girls	Women	Women
Livestock feeding
Cutting pasture	Boys	Men
Stripping maize leaves	Men	Boys, women
Herding	Boys	Men
Feeding (in stalls, pens)	Women, men	Boys
Watering (in stalls, pens)	Women	Other household members

Collated from survey data, participatory activity profiles, focus group transcriptions.

Stripped maize leaves were regarded as a reliable additional option but not as an alternative to other feed sources in the rainy season. 85% of the men and 71% of the women fed the leaves exclusively to small ruminants. In the trial seasons, only one man (2%) had completely relied on the innovation to supplement grazing. Others declared that they mostly (37%), partly (43%) or marginally (19%) depended on leaf stripping (with no pronounced gender differences). On average, less than half of the feed requirement for livestock in the rainy season was supplied by stripped maize leaves for both men (44%) and women (45%) farmers. One man had sold leaves as an additional source of income. After the trial seasons, a project technician from the Upper East observed: “Some saw that there is abundance of pasture. Some of them also allow their fields to fallow so that they can put their animals. So, they see it is a waste of time to go and strip leaves.” A man farmer from the Northern region discontinued the innovation for the same reason. Several men (from all three regions) expressed a preference for herding or tethering their small ruminants because they did not have adequate transport for feed. Transport of maize leaves was confirmed as a challenge by a project technician in the Northern region. Men respondents who found the innovation useful often had farms close to their homestead and appreciated not having to go out in search of pasture as a project technician noted: “Especially those with maize around the backyard stripped to feed their animals.” Tradeoffs between herding, cutting pasture or stripping maize leaves were less prominent among women, since the first two activities are often assigned to men and boys (Table 5). In a FGD among men in the Northern region, increasing land pressure and the decreasing space for pasture and roaming animals were identified as creating a future potential for leaf stripping. On top of that, the early stripping of lower maize leaves was seen as an efficient use of the whole plant, since commonly only the upper part is fed late in the cropping period. However, the lower maize leaves often get soiled and have to be cleaned before animals will consume them. For this reason, shorter maize varieties were assessed as unsuitable for leaf stripping.

Soil management

How does cutting maize leaves for feed interact with other uses of residues, such as for soil management? In seven FGDs, men and women farmers said that they leave maize crop residues in the fields without the straightforward intention of improving the soil. Roaming animals consume the “left overs” and leave their droppings, which respondents find beneficial. Some women collect maize stalks as firewood, as a man from the Northern region explained. In discussions in the Upper East, several men declared their wish to use all maize residues as livestock feed, but women said that carrying the residues home was too labor-intensive. In the Northern and Upper East regions, men described transport and storage facilities as lacking and expressed a desire for improvement. They said that with leaf stripping more of their residues were consumed by their own livestock, instead of by roaming animals that belong to other households. Overall, respondents did not establish a tradeoff between residues for soil improvement and residues for feed. In three FGDs (one with women and one with men in the Upper West and one with men in the Northern region) respondents indicated that more stall feeding with stripped leaves had led to better manure collection, a result that is confirmed by Komarek et al. (2021). The authors not only speak of more efficient manure collection through longer confinement in pens, but also of potentially better manure quality if ruminants are fed with leaves from Quality Protein Maize.

Productivity

In this domain, we ask how in the respondents’ eyes leaf stripping affects small ruminant and maize productivity and compare their assessments with the biophysical results.

Livestock productivity

In the FGDs, there was consensus both among women and among men that providing stripped leaves in the rainy season had raised their animals’ productivity. This was attributed to a higher rate of feed supplementation and more feed variety. With regard to the first reason, leaf stripping prompted farmers to shift from less to more feed supplementation. “The animals fed with fresh maize leaves look healthier than when you just tie them up in the field to graze,” a woman from the Upper East reported. This shift was associated with the second reason, a greater variety of feed sources. Variation makes “animals consume more than what they normally eat,” is how a man farmer put it. A concern about achieving a good mixture of feed ingredients to maintain ruminants’ appetite was a recurring theme of discussions. None of the respondents assessed the nutrient content of maize leaves as being superior to pasture. In fact, one man participant warned against giving maize leaves only as causing diarrhea. Both higher rates of supplementation and the diversification of feed sources were seen as resulting in healthier small ruminants: they maintained their weight or even gained weight (in a period of feed scarcity). Some women and men who had sold animals said they fetched better prices in the market because of their body size. The biophysical results confirm weight gains among sheep and increased digestibility of feed if stripped maize leaves are added to pasture. For the average daily liveweight gain, Komarek et al. (2021) indicate an estimated marginal mean of 29.3 g for sheep fed with maize leaves as compared to 10.9 g for those fed without.

Maize productivity

In three FGDs with men in the Northern region, farmers described how they had initially feared that leaf stripping would negatively impact on their maize productivity. A farmer recollected: “When they said at first that we should strip the leaves for the animals, I said they were going to spoil our maize because if you strip the leaves, it would not yield.” However, the majority of respondents attested that when the harvest came, yields were better in comparison to the common practice. Stripped maize plants were described as producing harder and larger cobs with bigger grain sizes. This result was explained in terms of the plants’ better utilization of fertilizer and water. One woman respondent from the Upper West attributed good yields to the improved seeds and fertilizer the project had supplied, rather than to stripping. Another declared that even without adding fertilizer her maize had developed more grain after stripping. In one Upper East community, leaf stripping was associated with past practices: “This practice confirms what our forefathers used to do those days. They were used to stripping the leaves of millet, guinea corn and sorghum which resulted in higher yields,” one man said. In spite of the widespread perception of raised productivity, there were also other voices. “There is not much difference [in maize productivity], but with the stripping you get the benefit of feeding your animals,” said a man from the Upper West. A woman saw some of her fellow farmers’ “excitement” over the innovation as being due to a tendency to value new experiences more than what you are used to.

Biophysicists compared the productivity of leaf-stripped plots with non-leaf-stripped control plots. Improved seeds, inorganic fertilizer and good agricultural practices were used on both. In general, these on-farm evaluations did not show any significant effect of leaf stripping on grain and leaf yields for both men-managed and women-managed trials (Abdul Rahman et al., 2022; Hoeschle-Zeledon, 2019). Where yield penalties were observed, these were explained with farmers’ potential inclusion of physiologically active leaves in stripping (Hoeschle-Zeledon, 2019, pp. 16–17). Komarek et al.’s (2021) in-depth study of seven households in the Northern region showed lower maize grain production (−12%) and higher forage availability (leaf and stover, +69%). The authors view this as a relevant tradeoff for crop-livestock management systems, without taking farmers’ views and their production goals into consideration.

Unlike the biophysicists, the men and women farmers in the FGDs compared the productivity of leaf-stripped plots (with inputs provided by the project) to their conventional practice (with fewer inputs and differing management practices). The gains in grain productivity farmers perceived must therefore be attributed less to stripping and more to increased input use and good agricultural practices. A project technician from the Northern region said that the support from the project improved residue productivity, but that farmers could not adequately make use of it: “If we could teach farmers how to preserve leaves and improve the transport system, stripping could be intensified.” Accounts of a couple of men and women farmers confirm the technician’s statement.

Human domain

The human domain is concerned with food security and the capacity to experiment. If livestock and maize productivity were to increase due to leaf stripping, how would it affect food utilization (consumption of animal products) and food availability in terms of maize? It is important to note that these aspects depend on gendered household decisions on what proportions of produce to sell or to consume. Both men and women from the Northern region emphasized the key role men heads play in these decisions. Respondents from the Upper West and Upper East confirmed similar patterns for their households, but also described more leeway for negotiations between husbands and wives. Decision-making on sales and income will be further elaborated under the economic domain.

Food utilization

In the communities investigated, the consumption of small ruminant products is limited. Sheep’s milk and goat’s milk are not part of the diet. Meat consumption tends to be restricted to special events. Animals are slaughtered for important visitors or to mark traditional festive seasons and Christian and Muslim holidays. Asked about household consumption patterns, many respondents reported that fairly equal amounts of meat are given to all gender and age groups on these occasions. This view was shared by women and men in eight FGDs in all three regions. Overall, livestock keeping was not depicted as being important for feeding one’s own family, but for earning money from the sale of animals to pay for school and health expenses, as well as other emergencies, and for raising or maintaining one’s own status as a livestock keeper. Higher animal productivity or better animal health due to leaf stripping can help farmers to get better prices in the market or increase their status, but will not necessarily lead to more meat consumption.

Food availability

Qualitative data on leaf stripping and maize grain yields revealed a high level of concern with food availability. “If we get enough yield, we may sell. But if we harvest and it is not enough … no one will sell because you may end up buying additional food,” a man explained. Several women and men (from all regions) reported that maize stocks had lasted for more months than before, due to the innovation. Two women heads and a woman cultivating maize on an individually controlled plot continued to use the innovation package even after project support had ended, in order to become food secure, as they said. This may suggest a relationship between women’s power to make decisions, adoption and food security. Income from maize sales was described as being primarily directed toward solving pressing needs (such as school fees, emergencies in the community), and only in the second place toward the purchase of inputs and livestock. There was no indication that income from maize or livestock sales would be used to improve household dietary diversity.

An additional aspect in the human domain is the capacity to experiment with the innovation at the household level. Women’s room for experimentation with leaf stripping is limited by patrilineal land inheritance and the cultural conception of men as “prime farmers” and “livestock owners.” We expand on this in the social and economic domains.

Social domain

This domain focuses on collective action and natural resource conflicts at the community level. Gender issues are considered as much as in the other domains (see conceptual framework).

Collective action

Men and women experimented with leaf stripping in groups, some groups building upon the work of previous ones. Asked about group composition and entry requirements, the majority of respondents stressed that age, gender and ethnicity did not matter for participation. Regular attendance at meetings and the ability to work hard were mentioned as the most important selection criteria. This claim contradicts the low number of women and young people in the sample, which is most likely due to their lack of control over land. This result links up with the human domain. If lack of access to land limits women and young people in their capacity to experiment at the household level, it also limits their potential for collective action (in the social domain), namely their participation in agricultural groups. Furthermore, less wealthy community members may be deterred from joining groups, if activities include the exchange of labor for a free meal. “You are unable to find money to feed them [after work on your farm]”, one man from the Upper East stated. Although relations with nonparticipating farmers were described as for the most part cordial, one woman from the Upper West reported envy by some community members who had not received free fertilizer.

Benefits of collective action were perceived as arising from discussing the innovation package with participants of the same trial, but also with those taking part in different trials, and even with community members in general. Moreover, mutual labor support was highly valued (especially in times of sickness or where hired workers could not be afforded). Good agricultural practices promoted with the innovation (such as spacing during planting) were seen as requiring more labor, which groups would efficiently help to tackle. One man from the Upper East doubted that groups would continue to perform joint labor after completion of the trial, particularly when it comes to planting: “The problem we have here is the erratic nature of the rainfall. You go to help somebody, by the time they will be coming to your own farm, the rain has stopped. When it rains you will use your family to do the work.” Two women’s groups spoke of collective labor as empowering. In spite of the above benefits, respondents also mentioned group conflicts related to decision-making, labor avoidance, and lack of attendance.

Natural resource conflicts

How does leaf stripping interact with livestock infringement on cultivated fields and competition for pasture? Field invasion causing damage to crops was reported as a widespread problem in the rainy season. It was described as resulting from three factors: First, the inability of boys (who do most of the herding) to control animals when driving them past planted areas. This may have serious consequences for the child (including physical punishment). Second, animals invade fields because they are tethered too close to cultivated plots, or tethered only loosely and break free. Third, some livestock keepers do not tie up their animals in the rainy season. Respondents regarded them as having no time to attend to their livestock, or as generally not being concerned about their well-being. Leaf stripping was assessed as helpful in terms of keeping one’s own animals well fed and preventing them from trespassing on others’ fields. But it would only reduce livestock infringement if adopted on a larger scale. Specific mention was made of Fulani pastoralists, who were described by a man as “trooping into this area with their animals to destroy our farm lands.” Data from three FGDs in the Northern region point to the pervasiveness of such conflicts between pastoralists and cultivators. Competition for pasture among farmers who cut feed emerged as another issue. “There is real competition. If you cut pasture and do not hurry to collect them to your house, another might go for it and store for their livestock instead,” said a woman from the Upper East. In this respect, leaf stripping was regarded as bringing relief. A man respondent from the Northern region: “Leaf stripping has now become one of the means by which we feed our livestock because even these days, you can’t get pasture in the bush. If you want pasture to cut, you must go far away to look for it.”

Economic domain

In the economic domain we explore how leaf stripping interacts with: first, resource access, ownership and decision-making; second, labor relations, roles and burdens; and third, respondents’ perceptions of the profitability of maize and livestock production.

Resource access and ownership, and decision-making

In the survey we requested respondents to indicate who in their household would usually have a maize farm (providing room to identify several maize farm managers per household). For MHHs, 84% of the respondents indicated that men heads produce maize on household plots. 24% reported that women spouses, and 11% that other household members (such as adult children or the head’s parents) manage individual maize farms. Although maize can be cultivated by different individuals, larger production is most often on household fields. Following this, we asked who makes decisions regarding the income from maize. 65% of the respondents mentioned sole decisions by men heads, 23% joint decision-making involving a spouse and/or another household member, and 9% sole decision-making by a woman spouse. 4% declared that they did not sell maize at all, all of them women. Deliberations between heads and their older sons at times precede decisions on how to use the income from maize sales from household plots. In the Upper East and Upper West, respondents said there was room for women to be consulted in these processes.

Plot allocation for household or individual farms (including crop selection) may change on a seasonal basis. In the context of patrilineal land inheritance in northern Ghana, men heads (as land owners) play a major role in these decisions. This was mirrored in the leaf stripping trial: in the FGDs the majority of respondents spoke of experimentation with the innovation package on a household (and not individual) plot. In four FGDs with men, some justified this with the argument that they were better able monitor the inputs provided by the project and the potentially higher yield. Should the innovation package be taken up on a large scale, inequitable benefit sharing among household members could intensify.

Turning to livestock ownership, respondents most commonly keep sheep and goats (Table 3). Except for pigs, women indicated less often than men that they owned livestock. However, this result has to be read with care. Gender norms see livestock ownership as underscoring men’s masculinity and social standing, and prevent women from openly declaring their holdings. As a woman from the Northern Region explained, individual ownership is known among household members but not communicated to “outsiders.” As a result, animals that women have bought or bred are often reported as being owned by the husband or a male relative. The same gender norms make women’s engagement in livestock business appear “disreputable.” “A woman can’t take an animal to the market to sell. It is a taboo in our culture as Dagombas,” said a man from the Northern region. “Butchers who buy livestock in the villages do not enter houses. Likewise, women do not visit markets to sell to butchers,” stated another man, establishing a distinction between the domestic and the market sphere for women and men respectively. Consequently, men heads have to endorse all livestock sales and effect market transactions for their wives. In MHHs, 57.5% of the respondents identified men heads as sole decision-makers for small ruminant sales. Joint decision-making, involving several household members, received 28.5%, sole decision-making by wives 8%, and by other individual household members 6%. Some women said they receive only part of the proceeds when their husbands sell animals on their behalf (and some men confirmed “hidden deductions”). In spite of this, married women claimed they had a good level of income from their own livestock. Women heading their households were seen as having the right and responsibility to take their own decisions. Yet, they still need to request male relatives to sell livestock on their behalf. If leaf stripping contributes to better livestock productivity and income through sales, women’s benefits depend on broader social consent to women’s engagement in market transactions, income control and equitable decision-making.

Labor relations, roles and burdens

Men’s land ownership goes hand in hand with the authority to determine who in the household should implement which task in maize production and when. This relates to both household and individual plots for two reasons. First, household plots are given priority in terms of labor investments by all members. “I cannot just come out and say: I want to go and do something on my farm, unless your household head tells you. Or you go and tell him: Now that we are done with your field, what about mine?” explained a woman farmer. Second, when permitted to attend to their individual fields, some women rely on the support of their husbands for certain tasks such as dibbling. These arrangements may lead to considerable delays in the cultivation of women’s fields. Independently from the fact that household members may own individual animals, livestock tasks were conceptualized as a joint responsibility.

Table 5 captures tendencies only and is limited to MHHs. The division of labor may vary strongly among households. The profile shows a potential involvement of all household members in most labor steps of maize cultivation. For livestock a more pronounced division of tasks emerges, with low engagement of girls (who are more called on for domestic chores). The prominent allocation of weeding, cutting pasture and herding to men and boys was explained by the need to manage nature. Recurrent comments in FGDs were that women are less able to weed “bushy fields”, collect long dense grass, or climb trees to cut branches for feed. In the absence of male support, however, women do engage in these activities. The fact that feed collection is carried out far from home was also seen as interfering with women’s cooking and food processing obligations. On the other hand, feeding and watering animals close to the homestead were classified as tasks women can include in their housework, while men devote themselves to extra-domestic activities. Hired labor was reported as being more common for maize cultivation than for livestock work (where respondents indicated a preference for purchasing feed, see results subsection “Environment”).

What are the potential interactions of the innovation package with labor relations, roles and burdens? In terms of good agricultural practices for maize cultivation, some women and men saw spacing as increasing the time for dibbling and seeding. At the same time, harvesting was assessed as less burdensome, as leaf removal would make air circulate and cobs more visible. In the survey, we asked respondents who in their household would strip leaves and allowed for the identification of several household members. In MHHs, 63% of the responses went to men heads, 20% to boys, 15% to women spouses, and 2% to girls. Although pasture cutting and livestock herding are often delegated to boys, leaf stripping was seen as less suitable for child labor. Particularly for stripping at the tasseling stage, training is needed. If the innovation is wrongly applied, cobs may not emerge – a risk many participants of the trial were not willing to take, as project technician explained. In their separate evaluation of seven households in the Northern region, Komarek et al. (2021) found a grain yield penalty of 12% that could have been due to wrong application, namely the stripping of physiologically active leaves. Our survey data did not reveal a preference of farmers for either of the stripping stages (tasseling versus silking). Qualitative data suggest that the silking stage is preferred in order to avoid risk and better involve boys (see also results of farmer field day evaluations in Hoeschle-Zeledon, 2019, p. 18). Some participants taught their spouses or older children how to strip leaves. But for the most part the one household member trained by the project (more often a man than a woman) remained responsible for it. Hiring labor for this task was limited for three reasons: financial constraints, the necessity to provide training, and the unattractiveness of the task, namely physical discomfort caused by contact with maize leaves such as skin irritation and itching.

Where farmers replaced some of their cutting of pasture with leaf stripping, this had two results: first, a shift from child to adult labor; and second, new chances for women to take part in feed collection. The following examples show how respondents defined leaf stripping as a new labor role in a variety of ways. Having felt restricted by the belief that women should not engage in pasture collection in the wild, some women heads from the Upper East and Upper West embraced leaf stripping as a new opportunity to provide for their livestock (partly assisted by their sons). A number of married men intentionally took up the new labor role, as they saw their wives as already carrying a high domestic labor burden. A project technician from the Northern region said about men in their new role: “There are instances where you go to the farm. The man farmer is alone doing the stripping. It is tedious for the bending and getting up if there are no women or children to help.” A few men found the shift from child to adult labor too burdensome and left stripping to their wives, arguing that this activity is not in the wild and therefore less “dangerous” than pasture collection. Farmers who cultivated maize close to the homestead were more likely to make continuous use of the innovation than those who had to walk a long way. Labor relief was assessed by comparing the walking distances to maize farms with those to places where pasture could be found. Those who abandoned the innovation explained this with the fact that it cannot be easily delegated to children or hired labor at a time of other urgent farming activities. If leaves are to be stripped before they turn yellow, these respondents stated, this coincides with other time-bound tasks of higher priority, such as weeding Bambara nuts for women, or tiger nut farming for men. Many men and women farmers from all three regions also said that sending children to school has caused a general shift from child to adult labor, which further aggravates the labor bottleneck in the cultivation period.

Perceived profitability

Farmers tended to be primarily motivated by subsistence production and less by commercial interests (see results subsection “Human domain”). This was also reflected in the comments of key informants who spoke of the low profitability of livestock production, since farmers predominantly sell to cope with emergencies (and not for highest profitability), or they sell at times when market supply is high (market days), thus creating a temporal and spatial glut. Still, some women and men farmers reported that they got better prices for ruminants that had increased in weight due to the consumption of stripped leaves (see “Productivity”). Assessments of maize profitability (including prices in output markets) featured lower in farmers’ discussions than the affordability of improved seeds and inorganic fertilizer for household production. Evaluations varied widely: some respondents found improved maize seeds not affordable at all, and had – through project support – experimented with them for the first time. Others said that they consistently buy improved seeds. Some farmers could afford certain amounts of inorganic fertilizer, others not. These differing indications could be found among men and women from all regions. A number of women said that they sell livestock, process shea butter, or save in groups in order to be able to buy fertilizer. The new government fertilizer subsidy program (Planting for Food and Jobs, PFJ) was regarded by some as helpful, by others as still not alleviating input costs sufficiently. Without fertilizer, maize leaves do not develop so well, so why should you strip them? This question was raised in a FGD among men in the Northern region. Irrespective of gender, respondents varied in their perceptions on maize profitability: Some viewed increased market production of maize as hardly profitable, considering the prices of inputs, while others spoke of high prices or came out against even higher maize prices, seeing them as a challenge for those who need to buy (or buy back) for consumption.

Discussion

This section asks what the conceptual framework has revealed in terms of its assessment levels (Figure 1), and what recommendations can be derived for gender-transformative leaf stripping packages. The investigation unearthed a number of overlapping norms that shape men’s and women’s actions in maize farming and livestock keeping and interact with the innovation package to produce tangible inequalities across all SIAF domains. In the following analysis of our own results (Figure 2) in relation to other scientific findings it is important to consider that norms are diverse, fluid, contested, and at times contradictory (Pearse & Connell, 2016).

Figure 2. Gender norms and tangible inequalities.

Our investigation shows that patrilineal land inheritance in the study region constitutes the foundational institution from which multiple other gender norms arise: it provides men with ownership rights and the authority to allocate land for household and individual production; it undergirds men’s voice in decisions on land use, labor, innovation adoption, and the use of produce and income. These findings confirm the results of other investigations in northern Ghana (Fischer et al., 2021); Michalscheck et al., 2020; Ragsdale et al., 2018; and Apusigah’s (2009) proposition that male community leaders and household heads are shown high respect due to their decision-making powers and resource control. Examples of tangible inequalities that can be seen in SIAF domains are a gendered productivity gap (productivity; cf. Adzawla et al., 2020), women and young people’s lower capacity to experiment and engage in collective action (human and social domains), women’s potentially lower income from livestock sales (economic domain) and their limited feed options (environment).

Separate from land inheritance but equally important for masculinity and social standing are norms associated with livestock ownership (most visibly shaping outcomes in the economic domain). Women’s ownership of small ruminants appears “blurred” and fragile as compared to men’s ultimate ownership, which at times is expressed as including wives and their property (cf. Fischer et al., 2021). Norms expect men to endorse the decisions women take to sell their own livestock and to accomplish the actual market transactions, a point that confirms Assan et al. (2018). However, (more or less hidden) production activities are not the only niche women may occupy in small ruminant value chains, as Konlan et al. (2018) and Avornyo et al. (2020) show. They found an active involvement of women as feed traders in the otherwise male-dominated small ruminant value chain in the Upper East, Upper West and Northern regions.

Norms that restrict women’s activities to the domestic sphere and prevent their engagement in pasture collection and livestock marketing (as well as participation in other specific nodes) seem to be particularly complex and contradictory. While men hold land titles, women are responsible for gathering natural resources such as wood, shea nut and dawadawa on male-owned lands and communal holdings. These collection and processing tasks are treated as unproductive or reproductive, and as mere extensions of women’s domestic responsibilities. The construction of farms as masculine sites explains why women have to rely on trees for income and consumption (Apusigah, 2009). On the other hand, respondents in our sample considered pasture collection and herding as “dangerous” and arduous work in the wild, and therefore unsuitable for women and girls. Leaf stripping on maize farms appeared as a safer task in a domesticated space. At this point, residue collection on farms is associated with women’s domestic feeding chores (similar to the collection of maize stalks as firewood). Equally contradictory and product-specific are norms regarding female market participation for crops and livestock. Apusigah (2009) writes that it is a near taboo for women to sell household staples without their husband’s authorization, but men (especially in the Northern region) do send their wives to market produce on their behalf. At the same time, women have some room to independently sell the harvest from their individual plots. By contrast, women’s involvement in direct livestock sales was regarded as a disreputable move from the domestic into the market sphere—a norm with the most salient repercussions in the economic domain (Figure 2). The public appearance of women as crop or livestock traders seems to question masculinity in different ways.

The norms outlined above restrict women’s potential benefits from maize and livestock production in general, and more specifically from the leaf stripping package. In the process of scaling leaf stripping, these restrictions should be targeted by integrating gender-transformative components into the overall innovation package. When tailoring transformative components to context, evaluations of interventions (similar in topic or geography) should be taken into account. For instance, for livestock keepers, Mulema et al. (2020) have piloted a community conversation approach in Ethiopia that blends technical components with reflections on, and changes in, gender norms. Fischer et al. (2021) have captured what various actors in the study area in Ghana view as entry points for shifting inequitable land tenure institutions. Their results can be supplemented by an evaluation of practical approaches used for addressing gendered land conflicts elsewhere (for instance, Stiem-Bhatia & Koudougou, 2018). In terms of social components, Galié and Kantor (2016) emphasize the need to engage with local conceptions of empowerment and to distinguish between a “language of gender equity” (that farmers may easily adopt) and more in-depth shifts in attitudes and behavior. Analyzing a small ruminant intervention in Tanzania, they write that gender sensitization has its limitations if it does not result in manifest changes, for instance in property rights. In our conceptual framework this corresponds to monitoring the effect of modified norms on tangible outcomes.

To promote gender equity, it will not suffice (and could even be counterproductive) to simply add a gender-transformative component to a technical package, if the technical components in the package are not examined and redesigned to fit the same equity goal. We use the general assessment levels of the conceptual framework (Figure 1) to showcase this. Table 6 summarizes the performance of the leaf stripping package, as well as the constraining or enabling aspects in the context, as outlined for each SIAF domain in the results. Additional components to the package are proposed that can enhance benefits and further crop-livestock integration. In the design of these components gender issues need to be considered in relation to the norms presented above and in relation to other norms. For instance, the increased feed variety and supplementation farmers reported was most beneficial to those who had maize farms near their homestead. Lack of transport was a reason to discontinue the innovation. An additional transport component in the package could address farmers’ challenges in relation to moving fresh maize leaves, dry residue, manure/compost and pasture between various sites and reduce labor times and burdens. Bellwood-Howard (2012) provides insights into an intervention in northern Ghana, in which farmers tested individually and group-owned vehicles for agricultural transport, such as donkey carts and bicycles. If norms that favor male livestock ownership and equipment (for instance, donkeys) or male income control to effect major household expenditure (such as for bicycles and carts) are not concomitantly worked upon, women’s participation in transport groups, ownership of and access to carriage facilities and benefits from them could be restricted. Table 6 is not exhaustive. For instance, measures to increase feed efficiency could be added. Due to the focus of this paper, these additional aspects need to be dealt with elsewhere.

Table 6. Recommendations for additional bundle components and gender integration.

Bundle performance, immediate interaction	Enabling or restricting context	Recommended additional components in bundle	Gender equity consideration
Increased feed variety and supplementation through fresh maize leaves	+ Maize farms close to homestead − Lack of adequate transport	Improvement of transport facilities (e.g., donkey carts, bicycles) Formation of transport groups Improvement of animal housing, enclosures	Ensure women’s access to and ownership of transport facilities as well as meaningful participation in groups
Better manure collection through stall feeding	+ Enclosed feeding sites for collection  + Short distances between collection sites and farms		Ensure women’s access to manure for individual plots and ownership of animal by-products
Increased maize residue productivity (for off-season use)	− Lack of adequate transport, drying and storage facilities − Alternative use of maize stalks as firewood	Improvement of drying and storage facilities by using locally available low-cost materials Improvement of transport facilities (see above)	Explore potentially gendered preferences of residue use (feed versus firewood) and ensure consideration of women’s needs
Risk of lower maize grain productivity due to wrong application	− High involvement of boys in feed collection − Training of one adult per intervention household only	Training of several household members and wider community to increase labor flexibility and capacity for hired labor Development of ToT model for community-organized trainings	Ensure equitable participation of women and young people
Restricted seasonal window for stripping	− Coincidence with other time-bound agricultural activities − Delegation to children or hired laborers restricted
Increased grain productivity if fertilizer and improved seeds are used	+ Government fertilizer subsidy program − Restricted affordability of inputs − Low market orientation of maize production	Experimentation with alternative fertilizers (e.g., compost in combination with increased volumes of manure collected)	Ensure women’s access to manure for individual plots and ownership of animal by-products
Increased grain productivity if good agricultural practices are used; increased labor requirements	+ Collective action, labor exchange − Erratic rainfall patterns restrict collective action − Necessity to compensate labor groups − Group conflicts	Better dissemination of climate information to enable timely collective action Development of innovative group constitutions	Ensure gender-responsive selection of dissemination channels and co-develop group constitutions
Animals weight gains, better health	+ Animal weight influences market prices − Generally low livestock profitability	Improvement of farmers’ marketing skills	Ensure women’s market participation and income control

Overall, the results of this study and their discussion underline the importance of building gender assessment into experimentation with innovation packages. The conceptual framework with its five broad domains, two assessment levels and mixed methods approach (Figure 1) lent itself for such a process. Referring back to the three steps outlined in the introduction, the framework supported the identification of enabling and constraining norms for equitable investments in and benefits from leaf stripping (Step 1) and the development of recommendations for additional package components with a particular focus on the consideration of gender equity (Step 2). The last step (Step 3), namely uptake by development practitioners, should be accompanied by further research to enable iterative processes of learning on concepts and tools.

Conclusion

In future, farmers’ participation and the integration of practical social solutions with scientific progress could be enhanced, based on Bergmann and Jahn (2017) model of the transdisciplinary research process. In their model, nonscientific and scientific actors develop a joint problem framing at the very onset, produce solution-oriented knowledge from both experience and evidence, and finally evaluate as a team how efficiently new knowledge and practices contribute to solving the identified problem. For the SIAF methodology, this would require expanding its application from the evaluation stage (where it is currently anchored) to include the design and action stages of the action research cycle. Joint problem framing with equitable participation by men and women would be followed by experimentation with various package compositions, accompanied by gender-transformative measures. Team evaluation would result in the co-adaptation and co-development of package components that are again reflected upon in terms of their gender equity. In the whole process, the five SIAF domains can help to maintain a broad perspective on intended and unintended outcomes comprising both emic and etic indicators. Limitations on the scalability of the developed packages can be expected and are due to the large diversity of households and differential demands of men and women living in them. This requires careful targeting. A discussion of household typologies as a framework for innovation targeting in our study context in northern Ghana can be found in Kuivanen et al. (2016). Limitations in respect of enduring gender transformation may arise from processes and institutions beyond the household and community levels and need to be considered. All in all, “design should not be looked at as a straightforward and controllable process” but as the development of new social and technical arrangements through iterations (Leeuwis, 2004, p. 145). More equitable gender relations should be part of this.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded by United State Agency for International Development with [grant number ARG#: AID-BFS-G-11-00002].

Notes on contributors

Gundula Fischer

Gundula Fischer, a social anthropologist and gender expert, worked for the International Institute of Tropical Agriculture (IITA). She is now an independent researcher and consultant in agricultural research and development.

Kipo Jimah

Kipo Jimah is a gender specialist at the International Institute of Tropical Agriculture (IITA) in Tamale, Ghana.

Eliasu Mumuni

Eliasu Mumuni is a senior lecturer in the Department of Communication, Innovation and Technology at the University of Development Studies in Tamale, Ghana.

Abdul Rahman Nurudeen

Abdul Rahman Nurudeen works an agronomist at the International Institute of Tropical Agriculture (IITA) in Tamale, Ghana.

Kofi Glover

Kofi Glover is a lecturer in the Department of Behavioural Sciences at the University of Development Studies in Tamale, Ghana.

Addah Weseh

Addah Weseh is an associate professor in the department of animal science at the University of Development Studies in Tamale, Ghana.