Assessing the application of gender perspectives in land restoration studies in Ethiopia using text mining

Restoration of degraded land is key to enhancing land productivity and farmers' wellbeing in subSaharan Africa. Evidence shows that the benefits of land restoration are tremendous, ranging from biophysical benefits in soil health, agricultural productivity, ecosystem services, to socioeconomic dimensions such as improving farmers' income and livelihoods. Yet one issue that is rarely considered is how the outcomes of restoration initiatives affect different social groups, specifically women, men, and the marginalized. Our study aims to understand the extent of gender and social inclusion in land restoration studies in Ethiopia. Through analyzing 314 peerreviewed scientific articles on land restoration initiatives and studies for Ethiopia published between 1985 and 2020, we aimed to: i) develop a novel approach that would allow for the semiautomatization of analyzing gender aspects in a large set of scientific documents using machine learning techniques, ii) assessed the current evidence on gender integration in peer-reviewed studies related to RDL, and iii) identified what gender issues are discussed in RDL literature. We employed text mining techniques to analyze the literature for keywords and to classify articles into three main classes: gender blind, gender sensitive, and gender transformative, based on the extent to which gender and social inclusion keywords appeared in the studies. Our analysis showed that gender and social inclusion are rarely considered in landscape restoration studies in Ethiopia, as most studies and interventions are heavily focused on the biophysical and environmental aspects of restoration. Despite an exponential increase in the use of gender and social inclusion keywords after 2005, most of the research (58 percent of articles) remains gender blind. The most frequently used related terms remained generic, e.g. community, access, participatory, and tenure. Only 38% of the studies could be classified as gender sensitive to a certain degree, and a strikingly low number of articles (4 percent) - which consider the systematic challenges women face - could be defined as gender transformative. This implies that scientific evidence about how land restoration has benefited men and women, and to what extent it could transform gender and power dynamics, is lacking. Our findings demonstrate the urgent need to consider gender and social inclusion to ensure associated benefits of land restoration practices.