Duncan Webb on Reducing Anti-Transgender Discrimination in India

In this episode, Shruti speaks with Duncan Webb about his research on reducing anti-transgender discrimination and other prejudice.

SHRUTI RAJAGOPALAN: Welcome to Ideas of India, a podcast where we examine academic ideas that can propel India forward. My name is Shruti Rajagopalan, and we are kicking off the 2023 job market series, where I speak with young scholars entering the academic job market about their latest research on India.

Our first scholar in the series is Duncan Webb who is a Ph.D. candidate in economics at the Paris School of Economics. He received his B.A. in PPE at Oxford University and his masters in economics at Paris School of Economics. His research lies at the intersection of development and behavioral economics. We spoke about his job market paper titled, “Silence to Solidarity: Using Group Dynamics to Reduce Anti-Transgender Discrimination in India.” We talked about deep-rooted prejudice versus marginal prejudice, preference falsification, behavior in group settings, endogamy, caste prejudice and more.

For a full transcript of this conversation, including helpful links of all the references mentioned, click the link in the show notes or visit mercatus.org/podcasts.

Hi, Duncan. Welcome to the show. It’s a pleasure to have you here.

DUNCAN WEBB: Thanks so much for having me on, Shruti.

RAJAGOPALAN: Yes. I’m actually really fascinated by what you’re doing because transgender community and both the kind of rights they’ve gained in India and the kind of discrimination they’ve had historically, there’s been a bit of a tipping point, thanks to the supreme court judgment in NALSA, a whole bunch of regulatory and legal interventions that have taken place.

You are looking at this space and you’re looking more specifically at how cisgender customers’ prejudice and their discrimination against members of the trans community plays out, in your particular instance, in the context of ordering groceries and getting them delivered at home. What you find is quite interesting and counterintuitive to at least what my priors were, which is when customers are making the choice in private, they actually discriminate against the trans delivery worker even when it’s costly to them.

This is quite a bit costly. They’re willing to forego almost twice their median daily food and grocery expenditure just to avoid 15 minutes of discussion or meeting with a trans delivery worker. After they’ve had a group discussion with their neighbors, they’ve had some chance to chat with them, they’ve all decided to collectively make the choice on what to order and who’s going to be the delivery person, you find that this prejudice against trans workers almost entirely disappears. First, can you just tell me what your priors were? Also, then, walk us through what was the big idea behind the experiment?

WEBB: Yes, I agree. It’s a very surprising result. We were examining a couple of things in this experiment. First, we were examining what’s the effect of this kind of group dynamic on discrimination. We are also examining what’s the effect of raising people’s awareness about the legal rights of transgenders. Actually, to start off with the experiment, we were really focusing on that rights piece and how that affected discrimination.

It ended up that the group discussion and the group dynamics were so strong and led to such a strong reduction in discrimination that that became the centerpiece of what we’re looking at. Concretely, what we show is that when you bring people together who were privately discriminatory, and they are just having a free-flowing discussion, they are talking to each other, they’re presented with opportunities to discuss and to hire workers, some of whom are transgender—we are not leading those discussions. We’re not forcing people to say nice stuff about transgenders. Once you bring those people together, even though they’re discriminatory in private, they start saying nice stuff about transgenders, and they start picking them much more often. They seem to persuade each other to discriminate much less, even when they’re making choices much later on.

RAJAGOPALAN: How does one figure this out in an experimental setting? If you can just walk us through that, because it’s very difficult to get insight into people’s private thoughts and private actions. Then the group discussion is, of course, you can observe as the fly on the wall, but what does your experimental setup look like?

WEBB: We wanted to have a real stakes’ measure of discrimination. The issue when you are measuring discrimination is if you just ask people, “Are you discriminatory?” Most people say no, surprisingly enough. We wanted to set up a field experiment where people were having an opportunity to select workers in a real stakes’ setting and see how much they discriminate against the transgender community.

Concretely, what we do is we go to people, and we say, “We will offer you some free groceries. You have to select which groceries you receive and also which worker carries out that grocery delivery. You will have to interact a bit with the grocery worker when they can carry out that delivery.”

People make a series of choices, and they know that that’s going to be a real choice that they’re facing. That means that they have an incentive to really tell the truth. If they really don’t want to interact with a transgender worker, they will have to select someone else and be obviously discriminatory. On the other hand, we are also varying the items we are offering to people. We have some random variations.

In some cases, one worker is offering three items and another is offering one item. That means that participants in this experiment have to face a tradeoff. Sometimes they might prefer the worker on the left when they’re facing this binary choice, but the worker on the left is also offering fewer items. They’re trading off a material benefit in terms of the groceries they receive with the worker that they’re going to have to interact with.

That allows us to get this economic or monetary measure of how much people are discriminating against the transgender community. As you said at the start, in the control group, people are willing to sacrifice almost twice their daily food expenditure in order to avoid interacting with a transgender worker. That’s quite an extreme level of discrimination that we see initially.

Then once we’ve engaged people in that group discussion, once they’re interacting with each other and they discuss who they would like to hire with two of their neighbors that we recruit at the same time, then that discrimination basically reduces to zero.

RAJAGOPALAN: I want to break this down. The first part of what you find, which is the control setting, it’s very Beckerian. It seems like it’s very much a deep-seated prejudice in that they are willing to pay for this particular preference or prejudice that they have to discriminate against someone from another group. Now, the way my prior was set is they do a group discussion, they’re likely to maybe virtue signal in the group, or they know it’s illegal to discriminate.

In a group setting, they don’t discriminate, but when they go back to their private setting, they once again indulge in the same behavior. Maybe there are some marginal differences, maybe the group discussion made a slight impact on their prejudice, but you would still expect it to be about the same. What you find is that after the group discussion setting, even when they go back and make these choices in private, the discrimination almost disappears.

To me, that is the super surprising part. What exactly is going on? Because it seems something far more than virtue signaling, which is what I would expect in the Beckerian traditional taste and preferences model of prejudice.

WEBB: Yes. I think that’s exactly the point. We wanted to measure not what behavior is going on during the discussion, but really how does that discussion impact people’s private choices afterward. Are people changing their private attitudes toward transgenders as a result of whatever communication goes on in that discussion?

What we find is that even after that discussion has finished, the people that were involved in a discussion discriminate much less when we’re measuring, say, 20 minutes later. On top of that, we also revisit lots of these participants. Around a month later, we go back and ask people a series of similar hiring questions where they’re choosing between different delivery workers. Even after that amount of time, people in the treatment group are discriminating significantly less.

They’re about five percentage points more likely to select transgender workers even after that medium-run lag. It does seem that something more significant than just virtue signaling is going on. People’s private attitudes are changing, something about their own personal norms, about what is right to do that is not just conditional on them being perceived by others to be virtue signaling. That also seems to be driving some of their behavior.

We were also really thinking about is it a virtue signaling story and is it a signaling story where people are just trying to show off that they’re nice people when they’re in a group? To that end, what we do is add a mechanism treatment where people aren’t involved in a discussion per se, but they’re making individual choices where they know that those choices will be revealed to their group later on.

This is exactly to test the theory that you just had. If people are virtue signaling, then when they know that their choices will be made public to someone, that implies that you should discriminate much less in that case as well. What we find, surprisingly enough, is that that doesn’t do much on its own. Just making your choices public is not enough to make you discriminate less.

RAJAGOPALAN: There is something about the discussion.

WEBB: Yes, exactly. There is some magic sauce in the discussion where that discussion is causing people to change their minds and change their behavior much more than a simple virtue signaling story would suggest. And in particular, we are thinking that there’s a group norm that emerges during the discussion that wasn’t there before. That might be driving what’s going on.

RAJAGOPALAN: The other hunch I had on this, and tell me what you think about this, is let’s say that there’s someone from an out-group where this need not be someone from the trans community, but anyone from another group. Could be another language, could be another caste, could be another gender. Let’s say there are two things going on. One is there is a certain awkwardness in interacting with any member from another group.

It’s awkward to talk to someone when you are not both native speakers of the same language, or it’s awkward for young women to interact with delivery men and so on. A second part of it is fear. What if my neighbors see me engaging with an out-group member for 15 minutes? Which is a reasonable amount of time.

This is not just someone leaving stuff at your doorstep and taking off and the fear of being judged. There are two parts to it. There’s awkwardness and fear of being judged. To me, when you see that drop even a month later, it seems like what the group discussion has done is the fear of being judged has gone away, but the effect has not completely gone away because it’s likely not deep-seated prejudice. It’s just awkwardness and some cost of engaging socially with people who are from another group.

Do you think this is what’s going on with cisgender people and transgender people? That there isn’t really a deep-seated prejudice. It’s just awkwardness. They don’t usually come across them very much. They don’t know how to act and how to behave and how to talk.

WEBB: Yes, that’s a plausible story. I think that’s not mutually exclusive with the more severe discrimination that might be going on at the same time. I guess what’s important for us, as economists, and for social science more broadly, is the fact that whatever it is that’s driving people to not select those transgender workers in this real stakes’ setting, that is causing de facto discrimination.

Even if it’s a lightly held personal attitude or awkwardness of not wanting to interact with someone, that does lead to these very consequential economic decisions that can have the equity and efficiency costs that we know that discrimination leads to. I do agree that there might be this kind of kernel of awkwardness of interacting with someone else that you aren’t familiar with.

That is exactly the kind of thing that might underpin discrimination in other settings. It’s an open question as to whether this type of intervention might translate to other groups. If we think about caste discrimination or religious-based discrimination, we don’t know very much about what are the effects of just getting people to talk about that other group.

Is that going to reduce either the social awkwardness or the more deep-seated negative attitudes toward that group? That’s something we really need to do a lot more research on. My hunch would be that the transgender community are probably in this space where lots of people’s attitudes are not severely negative. You do have this slight unwillingness to interact with, wanting to keep people at an arm’s length. That alone is sufficient to drive discrimination, but it also might be more amenable to changing as well.

RAJAGOPALAN: That’s exactly what I got from the paper, which brings me to the question of how does this differ from other kinds of discrimination which are much more rooted in deep-seated prejudice? Here caste is an important example as is religious discrimination. A lot of the anecdotal narrative we have, especially since we’ve had this explosion of digital revolution in India and WhatsApp groups and so on, is that there were privately held preferences that were discriminatory or bigoted against an out-group member.

People kept it quiet because it’s not done or not said usually in polite company, but the moment they saw on WhatsApp groups or family groups that, oh, everyone is okay to talk about this, you had the preference falsification ceased and you had this cascading. This is the standard model, and suddenly, you saw the level of discrimination go up when you went from a private setting to a public setting.

This is the broad narrative about what’s happening with minorities, especially Muslims in India right now. Are you worried that this is not a generalizable result unless we really know what kind of discrimination we’re talking about, and how deep-rooted the prejudice is, and whether there is any kind of preference falsification because those are things which are much harder to parse out through the experiment design?

WEBB: My sense is that this type of result is going to be generalizable in areas where people’s attitudes toward a minority are close to being marginal. People would be easily pushed to be positively discriminating but are, by default, negatively discriminating. When you are in a situation where people are very deep-rooted and very negative attitudes and very strong bigotry, just getting people together might even have the reverse effect, like you said.

There is evidence going back a long time to social psychology of the type of dynamics that can be generated in groups in which maybe what happens is you just get more extreme behavior. When you have people that are discriminating to start off with, they might reinforce that discriminatory attitude and reinforce that behavior. Whereas, if you start off with people that are more pro-trans or more pro-minority, you might get the reverse effect.

I’m thinking about this marginal space in the middle, where actually societal attitudes or the social norm that’s in place might be more progressive than the people’s negative attitudes, but the negative attitudes aren’t quite so discriminatory. I do think that there are other groups, it’s possible to think about, where there’s a norm of helping this group, but people still want to keep them at arm’s length.

In the U.S. or the U.K., for example, you might think about homeless people or people with disabilities, people of socioeconomic disadvantage like people living through poverty or the Roma community. Those are the kind of groups where I feel like this norm of helping and behaving correctly or positively toward that minority might be in place and might be sufficient to nudge people to help out. That’s partly because people’s attitudes aren’t quite so negative in that case.

RAJAGOPALAN: I completely understand the marginal part. Now, of course, it’s hard to know what prejudice is marginal versus what is deep-rooted unless we have a lot of evidence for it, or it’s very clearly observable. For me, one example I thought of on marginal prejudice might be in India you have these victims of acid; young women where spurned lovers or someone else who is a bully has done a lot of damage to their face by throwing acid on them.

They’re completely disfigured. They’re obviously victims of a horrific crime. Because they are so disfigured, there’s a fair amount of discrimination against them, but it’s not deep-rooted. It’s just a reality of the fact that they look so different from what one would expect, and no one knows how to quite engage with them. On the other hand, this may not bear out the same way for other things, such as caste and religion, for instance.

One super interesting paper that I had read was by Suanna Oh. This is really quite dramatic how deep-seated both privilege and discrimination, but also ideas of ritual purity and caste purity, how deep-seated they are. Because workers are willing to forego very large payments to not have to do a task which would conflict with their caste identity.

The difference is like 23 percentage points when it’s just the different task of another caste. The gap increases even more to 47 percentage points when the task that’s assigned is considered to be of a lower caste than what they got assigned at birth. That’s a really big impact. There were people that even when they were offered 10 times their daily wage, they refused to do the tasks that came from another caste. Some of this is obviously public setting.Even when they made it private, there’s questions of ritual purity, what happens to us when we die and rebirth and questions like that.

My sense of how we separate between marginal cases and deep-rooted cases, I thought one hint might be if they have anything to do with preserving endogamy. When it comes to religious discrimination and when it comes to caste discrimination, a very large part of the discrimination is the social norm where both the in-group and the out-group are policing their members to offer very costly signals to show that they won’t engage in exogamy because that’s how the group is maintained. My sense is those turn out to be very discriminatory, not because of just prejudice, but also because of social policing, which obviously won’t go away when you do an experiment of the sort you have.

It’s only going to get strengthened in a group setting. Where there are cases where endogamy may not be preserved, that doesn’t matter. My sense is the transgender community is not a threat to marriage endogamy in India because that is a different subgroup, especially when it comes to sexual orientation and marriage. That was the framework or the lens through which I was trying to understand discrimination against the trans community versus others. What do you think about this? How do you locate this in the general discrimination space in India?

WEBB: I think that’s super interesting. There is psychology literature going back a long way that basically proposes that feelings of sympathy and empathy toward an out-group might take place, but only when the minority group themselves don’t actually challenge the status quo, like you said. Possibly, what’s going on is that the transgender worker that might do one delivery in this experiment, they’re not challenging the status quo, like you said. In particular, one of the aspects of the status quo that you highlight in terms of endogamy, they’re seen as outside the circle of people that my participants would be interacting with in the first place.

It’s possible for the participants to externalize them if you like. Interestingly, there are other similar groups that do challenge the status quo. When we were starting out this project, we were very interested in the broader LGBT space, partly because it’s very, very under researched outside of the U.S. and Europe basically. We were doing a lot of qualitative work, a lot of focus group work, speaking to people about their attitudes toward transgenders and toward gay men mostly.

When you talk about transgenders, people are mostly pretty positive, at least in Tamil Nadu where we are working. However, if you talk about gay men, the idea of being homosexual, that seems to challenge the status quo much more and that’s where people become much more negative and have much more, obviously, negative attitudes. One narrative that surfaces quite frequently is, is it the case that if there are lots of people who are homosexual, the fertility rate of people in the state will be very low, and so we’ll run out of babies immediately. That’s something that clearly gets people’s hackles up. People will start saying, “This is challenging to the fabric of society in some way.” That is at least the stated driver of people’s negative attitudes when we spoke to them. Whereas for transgender people, it’s much more external.

RAJAGOPALAN: Yes. Also, one is the external. The other is the trans community in India is a little bit different from the movement that’s going on in the U.S. and so on. There they are the third gender, and a lot of it is assigned at birth, especially within the hijra community that you’re studying. These are not so often cases of people undergoing surgery, though they definitely also exist, but a very large part of the community is perceived as, “Oh, they didn’t choose this.” Whereas when it comes to the rest of the LGB community, it’s a question of “was this a choice?”

It is the fact that something is a choice that might threaten the status quo. In some sense, trans members announcing themselves is better to protect the status quo so that they don’t have problems in traditional marriages and fertility rates and so on. Is that another reason that this is not a very deep-seated prejudice, in the sense that they didn’t choose this, and they don’t threaten the status quo, in fact, them revealing themselves preserves whatever else we’re trying to preserve in the social fabric?

WEBB: I guess you could think of it like that. I guess one other aspect of the trans community, which is important for understanding the discrimination that they face versus, for example, LGB community members, is that trans people are so much more visually recognizable in this context. As you said, they have this traditional role within Indian society throughout history, and people can recognize them and they’re very aware of them.

That does make them more vulnerable to the hiring discrimination and economic discrimination that we see in the experiment. It also means that there’s this social norm that promotes helping them and prohibits discrimination against them. That is very different to the type of discrimination that the LGB community faces. I think the first order of concern for them is having to hide their identity for such a long time throughout their entire lives. Actually, the mental health costs potentially of that are enormous, and that’s something that I’m very interested in trying to do more research on in the future.

RAJAGOPALAN: This is also interesting to me because to go back to the legal question that you first started out your research with, right now, the marriage equality case is currently under consideration at the Indian supreme court. There is a genuine question of norm-setting. To me, it seems like a part of the group discussion, at least though not all of it, is responding to a certain legal or constitutional norm that is out there and that’s set right at the top, and people are responding to that.

How will norm-setting change their attitudes toward the LGBQ community? Not just the trans, if things like marriage equality go through in the Indian supreme court. I guess this is also an interesting question about policy implications of norm-setting, which seems like, oh, can norm-setting really do that much? What you find is, at least in some kinds of discrimination, they go quite a long way.

WEBB: Yes, indeed. There is a long literature in law and economics, which I’m sure you know very well, that talks about the expressive law hypothesis. Basically, the idea here is that when a law is enacted or the supreme court makes a judgment that can signal to people that society has certain values. In particular when we’re looking at a supreme court judgment, such as the NALSA judgment, in 2014, that affirmed the rights of transgender people in India, the fundamental rights according to the constitution, once people are made aware of that law, that they’re going to interpret that as a signal that actually, oh, society is becoming much more accepting toward transgenders than I thought.

That’s exactly the hypothesis we wanted to test in this study. Especially because lots of people in our context, in Tamil Nadu, they maybe had some hunches about the law, but they didn’t know about specific laws. You can change people’s beliefs about the law and then examine how that affects the levels of discrimination.

Concretely, what we do is tell people about this supreme court judgment, and then that also, similarly to the discussion that we mentioned earlier, that also reduces discrimination significantly in the short run. As you say, though, there is a question around how this affects whether something similar could occur for the LGB community. The Section 377 ruling that happened a few years ago, which basically decriminalized homosexuality, that potentially has the power to reduce discrimination against the LGB community.

Again, one of the issues is that lots of people don’t know about it. The awareness levels are very low. That’s a problem that is more concerning in places with slightly lower state capacity like India compared to in the U.S. where all the Supreme Court decisions are going to be so widely publicized that everyone knows about them.

I think the expressive power of supreme court judgments, especially when the supreme court is such a highly esteemed institution in India, despite some of the concerns about the independence of it, more recently. The people’s attitudes, when we’ve done qualitative work, people always say, “Yes, yes. It’s an amazingly trustworthy institution.” If people are hearing that this amazingly trustworthy institution is legalizing same-sex marriage, they might have second thoughts about their attitudes toward the gay community.

I think that potentially could have a really powerful effect. When you’re thinking about policy implications, the scalability of a law change is in some sense very easy. You just have to write something down on a piece of paper, and there’s no massive costs of running an intervention where you have to get everyone to talk to each other or running a big social program. If you can get a law enacted that signals this positive social value and raise awareness about that law, potentially through wider media channels, if you’re broadcasting on TV and on radio and on social media about the legal rights of minorities, that could be a very cost-effective way of reducing discrimination on a wider societal level.

RAJAGOPALAN: I think, again, just a caveat that only in marginal cases because we have outlawed caste discrimination. There are a whole bunch of laws, statutes, atrocities against Scheduled Caste, Scheduled Tribes, things like that, against hate speech. They’ve been on the books for about 75 years, and you don’t see the needle moving in a very big way in the statutory sense, or even in terms of those cases coming to court. I think I agree with you.

There’s a lot to be gained in cases where the prejudice or the discrimination is on the margin, like very marginal cases, but the deep discrimination or prejudice cases are clearly going to be slightly different, harder to solve with expressive voting. When I look at what you’ve written in the rest of the literature, and this is, like I said, work by Oh on caste discrimination in the job setting, by Matt Lowe, who’s looking at cricket teams and people working together versus in adversarial situations, Arkadev Ghosh, who’s looking at factory workers on a conveyor belt, either doing high-dependency jobs or low-dependency joint production and so on.

I find that it confirms the economic logic a lot, which is, now we can literally rank prejudices depending on how severe or deep-rooted they are, and you can literally price it. You can price it in the job setting, you can price it in terms of lags in production or inefficiencies in production, depending on how long it takes for the groups to start working together. You can check it out in team outcomes when it comes to Matt Lowe’s work and so on. What are the economists missing in this literature would be my question? Because all of this just confirms the economic view of how discrimination actually plays out.

WEBB: I think lots of the literature on discrimination has focused on people’s individual decisions and potentially how that’s impacted by interacting with another minority. Like Matt Lowe’s paper, we examine, if you’re talking to, or if you’re playing on a cricket team with someone else from a different caste, or if you’re working with someone from a different religion, how does that affect my own individual choice to discriminate or not?

That is usually modeled in economics as this prejudice parameter, so this deep-rooted thing where we don’t want to interact with the minority, or maybe it’s a statistical discrimination story in which I have a belief that someone from another caste or from another religion is going to be less reliable or perform less well on a job. In both of those pictures, typically we think that preferences are very hard to change. Typically, we think that beliefs will only change if you receive new information.

Maybe I interact with someone who is a super performer, and so I change my beliefs about the minority in general. But in my case, what I’m showing is that there is no information that’s being given to people in discussion, additional information that’s coming from the outside. They’re just talking to each other. People change how much they discriminate significantly after a very short interaction. They appear to be changing this whatever we would call it or whatever we modeled it as a deep-seated preference parameter.

It’s not that deep-seated because it changes dramatically in the space of 10 minutes, and it changes based on a discussion that naturally arises. I think what the typical models that we use in economics are missing is the notion of a group and the notion of being influenced by what happens around us. That means taking into account a couple of things.

One is social norms, as we’ve noted so far, that things like virtue signaling, things like the norm of not discriminating are going to be affecting people’s behavior, but also, that’s a very dynamic process where it seems like that norm emerges as part of the discussion. I think what we also need to do is incorporate ideas of mutual persuasion and how people are interacting with each other and communicating with each other.

It’s that nexus of both persuasion and norms where you get this big effect in this case. I think that’s the model that is useful to explain what’s going on in my case and isn’t the typical individualistic model that we see in the literature. Actually, so there’s one version of a model that could explain what’s going on, which exactly tries to think about the interaction between norms and persuasion that could explain all these results.

That would be that possibly people are both trying to match what’s going on in their group; they want to conform to their group. They think everyone else is discriminatory, initially, but they also want to signal that they’re a good person. Those are two very natural motivations that people might have. When you are just making a choice in private, then you’re just going to follow your private attitude and they’re negative. You discriminate in private.

When you’re making a choice in public, you have both these influences. You want to match other people in your group and you want to look like a nice person. On net, maybe that doesn’t make too much difference, and so you still don’t discriminate any less. However, if you can also persuade other people to change their mind, you have this third option, which is you can discriminate less, you can signal that you’re a good person, and you can also persuade everyone else to come with you.

In that case, you get both the benefits of conforming to your group because you’ve persuaded everyone to join forces and discriminate less, but also you get to signal that you’re a good person. This would explain some of the results, but it really requires both the norm story and the persuasion story to fit together and start together to explain all the data that we see.

RAJAGOPALAN: What else are you working on?

WEBB: The most related project I’m working on is a project in Madagascar thinking about the interaction between social norms and stigma and taboo. What we’re trying to do is understand how well can we change a social norm that harms teenage girls, and that social norm is the one that prohibits talking about menstruation.

Basically, in our context, it’s very taboo to talk about menstruation, and so people don’t know much about how to manage it. There’s also significant teasing and bullying that happens at school because of menstruation. That means that potentially, girls can end up not attending school because of the bullying or the teasing that they find. We partner with an NGO, CARE, to evaluate a program they’re looking at which is called the Jeune Femme Leader Program, the young girl leader program. Basically, what they do is they nominate ambassadors for menstrual hygiene and hygiene more broadly in each of the schools that we’re working in.

We evaluate to what extent are they able to change the social norm, promote positive open discussions around menstruation and how that impacts learning outcomes, absenteeism outcomes and other more psychosocial outcomes. What we find interestingly is that—this is very preliminary results—but that the girls in the treatment schools end up learning much more and progressing through the grades more quickly as well. Partly because of a positive narrative that’s built up around these ambassadors.

RAJAGOPALAN: That’s really cool. This is also a big problem in the rest of the Global South. There’s a big problem in other parts of Africa, there’s a big problem in South Asia, so it’s quite cool that you get that result. Hopefully, you’ll see more of this. Thanks so much for doing this. This was such a pleasure.

WEBB: Thanks so much, Shruti. It is a pleasure to be on.