Expert insight: Gertrude Sai, Quantitative Manager, Girl Effect

Gertrude Sai is Quantitative Manager, Evidence and Insights, at Girl Effect.

Lorna: Can you start off by telling me a bit about your background and how you came to be where you are now?

Gertrude: Yes, sure. My background is quite varied but mainly analytical. My degree in Psychology was definitely centered around statistics and research. My third-year dissertation project, which was fairly quantitative, was published and that really cemented my love of research and data analysis. Related to that, I’ve always had an inherent interest of working with vulnerable groups and young people, so once I left university I gravitated towards roles within that sector – mainly not-for-profit charities and social enterprises.

I started out working as a project intern for a company that works to improve unemployed people’s lives and get them transitioning back into work. I also worked at The Children’s Society as an analyst looking at the social demographic data of their users. More recently I was at Young Enterprise for five years, responsible for managing and overseeing the monitoring and evaluation department. My role consisted of data analysis, reporting impact and evaluation necessary for  the wider operations of the charity, but also to assist in fundraising and communication objectives. I really did work within a multi-disciplinary approach. Simultaneously, I recently completed an MA in Education and International Development. Then just before Christmas, I left Young Enterprise and now I’m at Girl Effect where I’m Quantitative Manager.

Lorna: Can you tell me a bit about what Girl Effect does?

Gertrude: Yes, sure. In a nutshell we empower girls to navigate the difficult time of adolescence. We’re an international development charity. We use our in-depth understanding to make a positive change in girls’ lives. We’re quite big on technology, using artificial intelligence, interactive voice-response, chatbot and mobile technology to find new and creative ways to measure our impact and reach, and to engage girls more personally across quite a large scale globally.

We reach millions of girls across more than 50 countries, so our scope is quite large. My main role is to look at data within Africa to try and understand the impact of the programmes we deliver, mainly around sexual and reproductive health aimed at 9 to 14-year-old girls.

Lorna: Can you tell us a bit more about the type of data you’re working with?

Gertrude: Girl Effect has a team of eight within its evidence team and everyone has quite different specialties. My role is very much focused on quantitative data so, for instance, I’ll be analysing the results of surveys that Girl Effect has administered. One of the projects that I’m looking at is the take-up rate of the HPV vaccine in nine-year-old girls. They’re given a magazine that talks about the benefits of the HPV vaccine, so we’re looking at the pre and the post survey responses in identifying if take up rates are different amongst those who’ve seen the magazine compared to a control group who didn’t have the magazine.

Lorna: A lot of charities and social enterprises use donor data to try and work out who’s donating and who’s lapsed, that kind of thing, but this seems much more focused on the recipients rather than the donors. Is that right?

Gertrude: Yes, although it is a funding requirement that we evaluate, the good thing about Girl Effect is that within each country we have colleagues on the ground who deliver the relevant interventions pertaining to the context and also administer the survey responses. They’re the ones who are doing the interviews, translating them back and sending them on to my team and then to me to analyse what the quantitative impact is, so it’s very much in-house in terms of data-collection, program design and content analysis.

Lorna: What kind of tools are you using and what sort of analytics are you actually doing with the data?

Gertrude: At present, I’m using SPSS for descriptive and higher-level analytics and also Excel, mainly when showcasing to external audiences because that’s quite a recognisedway to illustrate data. Also, something that I hadn’t come across before, but I’m currently learning is a software called Reflect that’s quite similar to SPSS particularly around crosstabs.

Lorna: Is your analysis primarily focused on looking back and understanding what’s happened or is there a predictive element to what you’re doing?

Gertrude: Much of it is currently is retrospective and focused on evaluating what has taken place. It is only a two-point survey, so we’re only doing a pre and a post. There is a wider piece of research that looks at the behaviour change but that also involves going out to girls and surveying as opposed to having a hypothesis and trying to use the data that we have to predict what the change would be. We are beginning to place emphasis around predicting the expected behaviour change. By going out to girls who have engaged in our interventions in identifying the long term behaviour change.

Lorna: Do you think there’s scope for more unstructured data analytics?

Gertrude: Within Girl Effect, I think our strength is qualitative because we get to work with girls on the ground, they always have stories and insight from girls who have engaged with our content. As you know, in the charity and social enterprise sector there can be an over-reliance on quant and numbers.

Lorna: Can you say a bit more about that? Why is that do you think?

Gertrude: No matter what charity I’ve worked for, as much as we want to be able to give people case studies and let them see the impact of the charity through the telling of a person’s story, resource-wise that isn’t always practical. You can only talk to one person at a time. One person has to interview someone, perhaps that interview then needs to be translated, then it needs to be analysed and someone has to write a report. It’s very labour intensive and can take a lot of time to do properly.

In contrast, if you’ve conducted a questionnaire then you can get to the point of having publishable findings much more quickly. Take an example from Young Enterprise. We might conduct a survey looking at employability competencies to see the effects pre and post intervention. That then gives us a shiny number of 95% of the young people who took part in our programme increased their competence for instance. That number goes into a report and people can see it on the page and immediately grasp what it means, quickly and easily.

Lorna: Can you talk a little bit about Young enterprise and what kind of work you were doing there?

Gertrude: My main responsibility was analysing the impact of our programs on young people’s employability skills. We had a range of programs however, YE’s flagship company program which is up to an academic year during which students get to run their own business.

My main responsibility was analysing different aspects of those programs and seeing what the individual impact is. For instance, with YE’s day programs, we would administer a questionnaire at the end of the day and ask young people to reflect back on their learning. In terms of analysis, we’d conduct a paired t test to identify whether the difference in scores was statistically significant.

With the company program it was generally a large cohort so over 15,000 young people take part in the program every year. The size of the dataset allowed us to manipulate the data more. For instance, we’d look at whether or not there were differences between gender or age in terms of impact. Age is particularly interesting because participants could be aged from 14 to 25. We were able to see if the impact is more or less on younger participants compared to those who are older.

Lorna: Thinking over your whole career to date, are there particular projects that you’ve found to be particularly rewarding or interesting or that have been particular favourites?

Gertrude: Yes. There was one project in particular. With the company program for instance, we have a modified version for students with special educational needs. Students with mild to moderate special needs could participate in the programme. We would administer a questionnaire online, but we found that take-up was generally low because the needs of the students were so vastly different. Those were milder needs were responding to the questionnaire whereas those with more serious needs were not, meaning that the results were not really properly representative.

Lorna: A self-selecting sample?

Gertrude: Yes, exactly. We had inadvertently introduced a bias because we were only really getting results from students with milder needs or who were better able to access the survey. Those with more complex needs were not so easily able to taking part in the questionnaire. I spoke to many teachers who had been involved in the programme to try and understand what the barriers were. Actually, the students with special needs really took a lot out of the program but that wasn’t being captured by our survey. We did some qualitative research to better understand what the barriers were and discovered that it really came down to the fact that the survey was administered online.

At the beginning of the academic year, we actually changed the questionnaire, keeping it online for those who wanted to respond that way, but also giving out a hard copy in each student’s folder of program materials. We immediately saw quite a substantial take-up and teachers saying how much they prefer that paper version.

It’s not something that’s drastically different, although there is a visible resource cost that had to be taken into account, but it was just a way of making sure that we’re capturing that impact and making adjustments as the student group taking part in the team programme. That is definitely a project I really enjoyed because I had already heard the stories from students over the years but saw their comments weren’t reflected in the data. It was nice to see at least just a small change and just making a paper version really made a difference. Having worked with questionnaire data in the form of numbers, it was nice to work on a project that involved hearing people’s experiences in an interview format.

Lorna: What advice would you give to young people who might be thinking about entering an analytics career? Is it important to have a qualification, is it a mindset, what are the things that have helped you to succeed?

Gertrude: I think for me personally, it’s definitely been apparent that you need an attitude of learning, a willingness to learn and to be proactive. It’s also important when producing reports and tables and data visualisations that you need to be able to keep the intended stakeholder in mind. It’s not just about taking the output from SPSS and sticking it in a report but actually being quite creative in the way you present it so that the data tells a clear story. That has always been quite key in my role because I’ve been fortunate enough to work with senior management a lot and go into board meetings to present my findings. Ultimately, they don’t want to know about the technicalities of the research, they just want to know about the impact. It’s being able to translate it in a different way and that took a lot of learning. Being able to do that has been extremely valuable to me in my career.

Scroll to Top