Introducing Faculty Features: Assistant Professor Shari Eli
December 7, 2017
Faculty Features is SPPG’s new series highlighting our faculty and research community. Our first edition features Shari Eli, an assistant professor in economic history cross-appointed to SPPG who has recently begun a project to digitize the Canadian census.
Shari Eli is an economic historian cross-appointed to SPPG who explores the effects of cash transfer programs on health and socioeconomic status. Now, she’s working on The Canadian Peoples (TCP), a colossal project that digitizes complete-count Canadian census records from 1861 to 1921. These digital records will enable Canadian researchers to inform policy-making through more and more accurate research on social and economic issues like inequality, immigration, regional development, urbanization, and more. TCP, which Eli is undertaking alongside Kris Inwood, Peter Baskerville, Geoff Cunfer, Lisa Dillon, Herbert Emery, Chad Gaffield, Alain Gagnon, Ian Keay and Marc St-Hillaire, has been awarded $1,465,825 by the Canadian Foundation for Innovation.
What research are you working on right now?
The main goal of TCP is to investigate questions about the way in which Canada developed from the period around Confederation into the 20th century. What are the circumstances under which immigrants settled here? What kinds of choices did they make upon arrival to Canada and why? How did they move around in the years after immigrating here? How did they interact with native populations?
We can also answer many questions about intergenerational mobility. For example, to what extent are the poor, and their descendants, upwardly mobile? Because we have census records that span from 1861 to 1921, and can make links to records that contain individual-level information for folks today, we can learn quite a lot about families over time and their circumstances leading all the way up to the present day.
What led you to working on TCP?
I first became interested in projects that look at the long-term effects of early life conditions when I was in graduate school in the US. Unlike Canadian census records, US census records have been available to academic researchers for quite a few years now. So it has been easier to study these types of questions in the context of American history.
Together with my co-authors Anna Aizer, Joseph Ferrie and Adriana Lleras-Muney, one research project that I undertook with US data was a study of the effect of welfare programs on how long one lives, the income earned in adulthood, and educational attainment. I wanted to study the same question in Canada, but Canadian researchers don’t yet have access to the complete count censuses. So that’s why being part of the TCP team has been so exciting: the potential to learn more about Canadian populations is so large.
What are you most excited about in your research?
I’m most excited about using history to answer questions that are of real policy relevance. The outcomes that I study are mortality, health in adulthood, level of education – if you look at historical datasets, you have the benefit of hindsight and knowing the outcomes of all sorts of events.
I’m excited to be newly cross-appointed with SPPG and I’m excited to work on something that is important for policy. I also hope people find this work really accessible – these kinds of topics affect people today.
What has been the biggest challenge in putting together TCP?
The older literature in economic history relies more on narrative evidence or a few important statistics, because collecting large data sets has historically been a great challenge.
Now, gathering and cleaning the data is a challenge. Matching individuals over time, and therefore across censuses, is important, for example, when trying study intergenerational mobility. Since literacy rates were low in the late 19th and early 20th century, census enumerators would come by and ask for a person’s name and the person often might not be able to spell it. Then the next enumerator would come by and ask for the same person’s name and spell it differently, and the actual person wouldn’t be able to fix it because he or she was illiterate. This creates a challenge for a historical researcher because one has to figure out if “John Browne” in the 1911 census is the same person as “Jon Brown” in the 1921 census.
Now that we have these opportunities to put these data together, we also face new challenges with regard to drawing conclusions that actually establish causality between the x circumstance and the y outcome.
TCP is an opportunity to marry qualitative and quantitative together – the quantitative being the digital data, and the qualitative being the whole body of work in the existing literature by scholars in the past who didn’t have these tools and relied more on narrative evidence.
We’re also collaborating quite a bit with folks in computer science. Computer scientists are really interested in the best strategies to get a link that has a high probability of being a true match, and they are interested in what to do with multiple matches. Historians are interested in this, sociologists are interested in this, and it really makes for interesting challenges with regards to writing new code and matching algorithms. It’s exciting to be able to be a part of such interdisciplinary work.
What’s one thing about your findings that you wish more people took into consideration?
Policy-makers sometimes become hesitant to provide cash transfers to the poor as opposed to implementing voucher programs or providing other types of in-kind transfers. But there’s great evidence showing that just giving people cash does lead to better outcomes. In my co-authored work on mothers’ pensions, we compared mothers who had received cash transfers to those who hadn’t, and then compared the difference in outcomes for their children. What you see is that mothers who received the cash transfers likely spent it on their children, as intended. While we can’t make comments about which is better between cash transfers and in-kind ones, we do show that cash transfers, which are often more feasible to implement than in-kind programs, are very beneficial to recipients.
What advice can you give students as they refine their own research interests?
I often tell students that they should take some time and think about the questions that really interest them. Write them down as they come to mind and then go through those ideas. Then, pick out the ones that seem like gems. After that, see if there’s any data that can corroborate what you think might be true, or if you can relate what your thoughts are to an existing literature in the research.
An idea might come to you as you’re walking around, or as you’re describing your work to a friend. One of my best ideas came as I was moving house – something about packing, unpacking, and sorting spurred on new thoughts and ideas.
But research is better when it’s not motivated by thoughts like “I could gather all this data from this website… Surely there’s something interesting to do with that.” That strategy is probably not going to result in research you’re going to cherish. It’s probably not going to be the thing you’re going to talk about and continue to build upon for the next decade.
Our last question is for fun! What’s a song you can confidently say you know all the words to?
“Across the Universe” by the Beatles. What’s great about that song is that it’s about thoughts and ideas flowing freely out of us and the joy that comes along with those creative moments.
Find out more about Shari on her website.