How would you feel if you received the following text messages from your child's school?
- "Engaging in your child’s school life will enable him/her to enjoy school more and to work better in school."
- "Make a list with your child about 3 things that s/he likes to do in school and 3 things s/he does not like and ask him/her why."
- "Support and guidance from parents are fundamental for your child. Tell us how you participate in his/her school life."
- "School is a space for everyone, including families. You are invited to come to school next week. See you there!"
From high-income countries like the United States to middle-income countries like Brazil, messages like these, sent weekly over text messages (SMS), have been shown systematically increase parental engagement within primary and secondary education - improving student learning outcomes as a result. We know this thanks to randomized control trials that have assigned some parents, but not others, to receive these messages. These studies followed parents and students over the course of a school year, tracking their beliefs, behaviors and school outcomes.
These interventions are often classified under the umbrella of behavioral “nudges”—they do not really convey new information to parents; rather, they try to change how parental engagement in their children's education is perceived. Previous research has shown that these messages make school life more salient to these parents, inducing them to independently collect decision-relevant information. For example, a study in Brazil shows that targeted parents became more accurate about their child’s grades in math and language classes, even though messages did not communicate those directly.
Nudges are considered to be cost-effective—in fact, they were listed as a ‘great buy’ for cost-effective education approaches by the 2023 Global Education Evidence for Advisory Panel. First, their effect sizes can be meaningful: studies in Brazil and the U.S. document that such interventions can boost learning rates by about 25% over the course of one school year. Second, text messages are relatively cheap all over the world: In Brazil, the intervention implemented cost about USD 10/year per student and in Cote d’Ivoire, the program we implemented cost less than USD 7/year per student.
Could the same interventions help address the learning crisis more broadly, even in the very different context of rural Sub-Saharan Africa? The answer is not clear. On the one hand, these interventions have been shown to help vulnerable students the most, indicating they might be even more cost-effective in low-income countries. On the other hand, the latter are characterized by additional challenges. First, lower adult literacy rates suggest that parents might not be able to benefit from the intervention as much. Second, teachers' higher absenteeism and lower proficiency suggest that higher parental engagement alone might not suffice to improve learning outcomes.
To test this, we conducted a cluster-randomized controlled trial in Côte d'Ivoire in the 2017-2018 school year. The study, sponsored by the Transforming Education in Cocoa Regions (TRECC), was conducted in partnership with the Ivorian Ministry of Education, which enrolled schools, teachers, and parents directly. We were also interested in whether messages to teachers with pedagogical tips might improve children’s outcomes.
Working with 99 public primary schools (children typically aged 6-11), we randomized 25 schools to send messages to the parents directly, and 25 schools not to (the control group). To test for complementarities between efforts by parents and teachers, we assigned another 24 schools to send messages to teachers only – aiming to increase motivation, class time on task, and to teach more effectively. The final set of 25 schools sent messages to both parents and teachers. To test specifically whether literacy constraints are binding, we further split the parents-only treatment group into text (simplified French) or audio versions of the messages (French and six local dialects, which recipients could choose upon enrolling in the study). Our research partners on the ground, Innovations for Poverty Action (IPA), conducted surveys with parents and children (along with standardized tests in French and math for children) at the beginning and at the end of the 2017-18 school year.
We found that nudges to parents increased student learning outcomes by about the same magnitude as in higher-income settings, although effects were not as precisely estimated. Part of the reason for this might be the much higher degree of heterogeneity we saw in the Ivorian setting. Concretely, the benefits of the intervention were mostly concentrated in lower-performing students and boys.
While both are consistent with the idea that the intervention can decrease inequalities (since girls typically perform better than boys within primary and secondary education), it is disappointing that there were no detectable benefits for girls, who tend to drop out of school at higher rates than boys in Côte d’Ivoire. We believe an important factor might be that social expectations constrained the extent to which parents were willing to participate in their girls education in this context given lower expectations for girls' educational attainment compared with boys.
Interestingly, nudges to teachers were not effective in this setting, either by themselves or in combination with nudges to parents. If anything, in a companion paper we show evidence that targeting teachers might have even backfired in this context.
Also, literacy constraints do not seem key even in face of 40% illiteracy rates: text messages worked just as well as audio in this setting. The reason seems to be that parents came to school to ask teachers about the content of the messages. Showing up in school more often is already part of the theory of change of the intervention, in a setting where teachers perceive most parents to be disengaged from their children's school life. So there were likely some changes in parents’ behaviors around school engagement as a result.
One surprising finding was that the intervention seemed to have increased child labor as reported by children, especially when it comes to helping with house chores (with no clear gender differences). Since educational outcomes improved alongside children's engagement in these activities, these speaks to the challenges of understanding the intricate relationship between children's time use, parental engagement and learning outcomes. In another paper, we have also shown that measuring child labor in cocoa fields, an important force driving student dropouts in this setting, is plagued by reporting biases, especially in areas more commonly targeted by sensitization campaigns. Together, these findings suggest we still have much to learn about children’s engagement in child labor and its associated with schooling and learning outcomes.
All in all, SMS nudges to school parents are a promising cost-effective intervention to boost learning outcomes and educational inequalities even in low-income rural setting, but further research is needed to understand (1) how they could benefit a broader range of students, and girls in particular, and (2) additional interventions that could magnify their effects, by successfully motivating teachers and decreasing the forms of child labor that detract from the efforts to keep children in school beyond primary education.
Importantly, the effects we see are small, and insufficient to meet the needs to learners to truly address the learning crisis. To increase effects on learning, demand-side programs such as this one likely need to be complemented by supply-side programs (i.e., teacher professional development) to translate into meaningful gains for children’s learning.