 Thanks for having us. So this presentation is on behalf of myself and Professor Lilly Sy, who unfortunately had to leave for a flight to Liberia, but wishes she could be here. So I'm going to talk a little bit about our research with Ms. Lendo and my society. But first, I want to situate this in where we're coming from as academic researchers. So I'm a third year PhD student at MIT in the political science department. And what Lilly has started about a year ago is something called MIT GovLab, which is a group of political science researchers trying to bridge the gap between citizens and government. So we've talked a little bit earlier today about the role of academic research. So I'm hoping to debunk some myths. So basically what we're trying to do is see how new methods and programs and technologies can make governments more accountable to their citizens. So we're coming at this from both ends of the spectrum, the citizen side and the government side. So what MIT GovLab tries to do is to work closely with practitioners and partner organizations and funders on an iterated experimentation approach. So we're not going in for year-long academic research agendas that we are bringing to the table. Instead, we're working on relatively short and small scale experiments that use rigorous social science techniques that build on the findings from previous rounds. And our goal is to test which operational and design choices are the most effective in particular contexts. So what we're really trying to do is show that impact research doesn't have to take a long time. And it doesn't really have to be stuck in the ivory tower. Instead, it can be really practitionersly useful for funders and practitioners in the field. So we're coming at this collaboration in terms of how it fits into our larger research agenda at GovLab around questions of civic technology. So specifically, we're interested in learning under what conditions do new and lower-cost ICD channels for citizen feedback and monitoring lead to increased participation. And secondly, we're looking at how useful do governments find this feedback provided through ICD channels compared to feedback through traditional channels. And then finally, when do ICD channels for citizen input lead to more government response of this and better service provision? And can they facilitate cooperation between citizens and government? So our research with Ms. Lendo falls under this first question. And so we were connected to my society and Ms. Lendo through the OMDR network to investigate this common question of interest. So specifically with our collaboration with Ms. Lendo, we wanted to answer three questions. The first is what presentations or what we call in the literature, framings of political news and information are most likely to stimulate online citizen engagement in Kenya. And then secondly, what types of actions are Kenyan citizens most likely to take online? And finally, does online social media alter the existing patterns of political engagement in Kenya? So again, as academic researchers, we're bringing to the table some political science theories of political behavior and participation, but also drawing on questions that Ms. Lendo really wants to know and my society wants to answer in seeing how we can fit these all together. So to do that, we created a randomized control experiment using Facebook advertising. So we used Facebook advertisements to recruit a sample of online citizens in Kenya who use Facebook. So the ad looked like this. And once someone clicked on the ad, they were randomly assigned to one of these six treatment pages. So again, I mentioned framing. So we were interested in looking at how different framings of the same information would potentially differentially affect engagement and mobilization and sort of desirability to take any sort of action. So here we're looking at an opportunity framing, which said something that I'll show in a couple of slides along the lines of you should take action now to gain something. And then the threat framing was situated in the American politics literature that suggests that people are very risk-abiding. So we suggested if you don't take action now, you might risk losing something. And then finally, we had a neutral or control treatment page that we could also compare. And then we were interested in adding this bandwagoning option, which I will also show in a couple of slides. But basically, it explained to people how many others are also taking action and how you should join the effort around this issue. And then to measure our outcomes, we had four potential actions that people could take. They could share the article on Facebook. They could share it on Twitter. They could write their own personal message to the Senate Majority Leader, or they could sign a petition that was let go within forward. So this is an example of one of the pages that we randomly assigned people to. This is the control or neutral page. So this is where something like this partnership couldn't have been possible with researchers sitting in the ivory tower and practitioners on the ground. This actually required traveling to Nairobi, meeting with Jessica, conducting focus groups around different issues that we thought might be viable options to show people online that would actually get them to take some sort of action. So here we decided that for the moment, it would be an issue around spending in counties in Kenya. And sadly, spending on development that didn't actually happen as it was supposed to. But we thought that this might be a good call to action for people on Facebook. So this is what the neutral page looked like. And then this is what the added opportunity page had at the bottom. And then this line is the bandwagoning line. So a concrete number about how many other people have joined the effort, encouraging others to also join. So some initial results of the full-scale experiment. So we ran this ad on Facebook for two weeks. And then the ad was served to about 1.4 million Kenyans on Facebook. And of those people who saw the ad, about 24,000 actually clicked on it. Of those, it was about 18,000 unique clicks or visitors. And then the conversion rate was a bit smaller, about 4%. 653 unique visitors from Facebook actually took at least one of those four actions. But in the relative scheme of things, this isn't a bad conversion rate for Facebook ads. To answer that first question of what types of framings of information are actually more engaging and mobilizing, it looks as if it didn't have much of a difference. So this is one of the nice things about randomized controlled experiments. We can actually get at that attribution question and that causation question. So in the question of does opportunity or threat lead to more action, the answer is no. If you look at this control, opportunity, and threat bars, it's all about the same. But what's interesting to note is this control plus, opportunity plus, threat plus actually does lead to more action compared to the relative threat or opportunity page. So this bandwagoning line telling people how many other people are actually taking action and encouraging them to join the effort did have an effect. So the second question we were interested in in terms of what types of actions are people taking online, somewhat unsurprisingly, we see that most people want to share the article on Facebook. We recruited a sample from Facebook, so that's probably not surprising. But somewhat surprising is that the petition was second most popular. And then the Senate one was arguably the highest cost since people had to write their own personal message. But still was not the last one, which is Twitter. Perhaps suggesting as some conversations earlier have alluded to that maybe there are different people on Facebook and Twitter in Kenya. This is something that Jessica and I have talked about investigating potentially in further iterations of this experiment. And then the last question we had. Does ICT tools, do they lead to greater engagement of the disengaged? So if we're looking here at men and women who are on Facebook and Kenya, the answer is no. So empirically we know that older men especially are the ones leading the groundwork in the political arena in Kenya. And here we see that men overall are doing much more of the action taking than women. So of the people that actually clicked and saw any one of the six pages, a higher percentage of men are taking any action than women. And the way we did the Facebook advertising we're able to disaggregate among men over 30 and men between 18 and 29 years old. And again we see that it's the older men or the men over 30 who are doing most of the action. And somewhat disappointingly young women are the least likely to be taking any sort of action. So in the future we would like to build on these findings and do some more iterative experiments, integrating political science theory into practice and design and somewhat trying to see if we can get these disadvantaged and generally disengaged people to take more action. So one of the things we've been thinking about is that in the political science theories they suggest that disengaged people generally care about one or two issues and that it's the privilege and already engaged who can sort of engage in politics in an abstract level and more generally. So maybe if we create a new AB testing experiment around health and education issues, can we get women and especially young women engaged a little bit more? So I will now turn it over to Jessica to explain more from the window side. Thank you Lea. Lea came to us at a time which was very opportune for us because I had just become executive director and document that I needed to start shaping the strategy for the site. And in many ways, she met my needs and it was natural that we'd sit and discuss. Was it early February or anything? January. Yeah, and it was just before the beginning of the last house sitting. Our sitting started in February, mid-February and end in December. And so that was very opportune, that it was just right. At least it built into my year and I could factor in from the beginning. So she didn't even get a blanket, no. And of course, as she came with our media support was already a good thing. And so the opportune presented possibilities for us in public participation. Kenya has a law that says the public have to participate in all levels of governance, both at the national and sub-national level. But the system is not in place yet within government and how to engage. And so for sure, government doesn't have a structure in place. So for us, this was an opportunity to experiment what's possible in this. Then at the time, Zalendo had about barely 800 users on Facebook. Yet we already had about 6,000 users on Twitter. And so it was an ideal time for us to try and grow our deals without having to spend any, you know, for it. And then it was also an opportunity to see how can we relate differently with the parliamentary secretariat and also the parliamentarians themselves in terms of building social capital, both for the individual MP or senator and the House seeing us as actually a support system for them, given that the House is still battling between the presidential system if you're borrowed from the American and the Westminster system we had before. So with that tradition and push and show between the system and within the system as to people wanting to move ahead and say, hey, let's open up a space for public involvement. But there's, of course, you know, change comes gradually as my society can act as in UK. Then also provided research expertise that we didn't have. For me, I'm more a practitioner. I'm not interested in doing too much qualitative, quantitative stuff. She met my needs with something I didn't have. Then practical challenges. So her coming to Nairobi was very useful because when she came, of course, I had my guard up. What's she gonna ask? And it's not in a bad way, but a lot of times academia comes with already a set template of what they want to be done. And so she came and she asked me, what did I need? And I told her, I could explain to her how the parliament works and the things that were most at the heart of the people and that the public discourse in the country and that in itself meant that assumptions are corrected and we're able to identify what issues we could take out to a focus group and so that we're able to figure out what we can focus on in the short term. So contextual reading and sharing was important. And then identifying a particular issue with public interest and low level risk for us was important because ultimately, she can do something from America and then up on the line will be mine and the founders. So she would have to make sure that we are protected and I won't start getting strange phone calls. Already I know my phone is stuck so I'm making sure that my phone was very important. And then she had also to be flexible to the ever-changing parliamentary calendar that a lot of times they can say these are the issues that are going to be discussed but because of the things that are happening in the local context they keep changing. Like for example, last week there was a bill that was supposed to have been done fast reading but it's been pushed on the side because they said they need to censure the speaker. So you have to decide who is going to respond to that need and how fast. And so having been flexible was useful for this. And then now with regard to my society it is not just about this particular exercise but just generally having a tech team which is located in a mature democracy. Sometimes they may not be very responsive to our needs because they don't realize that things in the democratic space in a country in transition are more fluid. And so that's something important for academia to note and also anybody wanting to support or do tech work in that context that they need to be flexible so that they are able to make their partner responsive in the context. And then there was a delay in sharing the final research findings and that meant that we could not give them to this leader of Senate, majority in the Senate as proposed and then for us we actually forgot to do a blog post around it. Just to showcase what the findings were interesting. And it was useful to know that people had particular ideas on what they liked done or the development money from their context. So the findings, what did that mean for us? Older men taking action versus the young. I think they work for you survey that was done a while back which also showed the same thing that older men tend to be very vocal and so indirectly this validates other civic research in the space. And then in Kenya women at many times vote according to the structure they get from the men in many ways. So you find that whatever they say kind of carries the day. Then the counter development issue was far removed from the youth as the older men see control the power. They mean instruments of power and work in Kenya. They control the land, they control the tea, cash crops, proceedings and things like that. Then the bandwagoning frame was likely to get in traction now context because people in as much as politics expressed at the individual level a lot of times people vote collectively. So if my ethnic community wants me to do this coalition many for me disagree with them on a personal level that likelihood that I will vote along that same line is very high. So there's no one that does not surprise that was not surprised to us. That Kenyans wanted more resources allocated development work at the county level was not in doubt. And then for the future information or the issue we pick has to be iterated at the level at the personal level. You know, it needs to sting the individual so that they're able to take action online. Because financial figures sound, you know, like pictures of pies in the sky you really want to get anyone to take action. So we need to bring them down to what does this mean to the course of paying a teacher, a policeman, you know because those are the basic a lot of people are lowly paid. So if that's the people the people are trying to influence at that level we need to make sure the money doesn't feel like it's too far away from their reality. Then unintended benefits were mainly so in that sense I am in love with Claire. So we had visibility of more than a million people and that was great for us. We gained 12,000 members who have stayed with us since we have had very little dropouts. And then the research findings were useful to Zalendo's strategic direction distilling the value of the value proposition can give parliamentarians. Actually right now we are exploring the possibility of growing that there's actually an MP who's reached out to ask can we do this public participation around a bill together. And the learnings I'm using I'll be using essentially is what I learned from Claire's experience. Then the parliamentary secretary commission is very open to us just talking and discussing how can public participation really be realized at the parliamentary level. Issue framing becomes critical. I can never, I didn't know about it before so harboring it up is now something I keep in mind in figuring out how do I attract attention of this particular group that has been and engaged before and then they were from the user survey that they ran as part of this exercise were able to gauge user interests and identify touch points that you can capitalize on. It was interesting that there's a reading shared before about the experience of the interested by standards in the US. And I would say the same is true because the information she was able to find was religion, family members and friends are very influential in determining the political results of people. So in that sense, we need to find ways of using those touch points to get action on the site and also offline. Yeah, thank you.