Gender-based inequality in access to education is an issue of global concern. The use of educational technology is often cited as a potential way to help close educational gaps and promote girls' education. However, the existing evidence base in relation to girls' learning outcomes when using educational technology in low-income countries is limited. The evidence base was recently boosted by a study in which findings from classic educational development studies were revisited and disaggregated by gender. In this paper, we present a secondary analysis of this dataset, focusing specifically on the educational technology-focused interventions, and sourcing additional data. The analysis comprises 35 interventions, reported across 15 publications, published between 2003 and 2019. We discuss the relative efficacy of different types of educational technology interventions by comparing effect sizes of learning outcomes for girls. The findings suggest that interventions which focus on distributing hardware alone have mixed - and sometimes negative - effects on learning outcomes for girls. The impact of software-focused interventions is more positive, particularly personalised learning applications. Furthermore, we consider characteristics of the studies included in the analysis, and identify gaps in the literature which will help shape research in this field in the future.