We’ve asked our #DataImpactFellows to share a day in their life.
Over the past few months, Ben Brindle has been spending around half of each week working as a research assistant for the labour market economist Jonathan Wadsworth at the LSE’s Centre for Economic Performance (CEP). He takes us through a typical day.
On those one or two days a week when I make the journey to the London School of Economics, my alarm greets me only a little earlier than usual, but early enough for me to thank the heavens that I don’t have to reach the office in Central London until 10am.
“Who on earth are these people that commute into London from Brighton every day for 9am?” I think to myself, before remembering that in just a few weeks I’ll start working at the Department for Work and Pensions in Whitehall and will be doing exactly that.
A beautiful spot to work
Any Southern Rail delays aside, I arrive at my desk in the wonderful Lincoln’s Inn Field building and have up to an hour before Jonathan arrives. This gives me some time to make some last minute changes and clarifications to the code I’ve been writing and amending for a study that uses British Household Panel Survey and Understanding Society data to investigate the consumption effects of an increase in the UK minimum wage on minimum wage households.
Image: The building where the Centre for Economic Performance at the LSE is based
One last time to check that the minimum wage variables and the controls are correct – thanks George Osborne for making it that little bit more complicated and creating a new minimum wage rate for 21-24 year olds.
Like going through your homework with your teacher
Soon enough Jonathan arrives and we head off to his office to run through the code that I’ve spent the last few weeks working on. It almost feels like going through your homework with your teacher.
As Jonathan clicks to run the code an error pops up and panic begins to set in. Oh no, it’s fine. It’s only because the USB is assigned to a different drive on his computer. File paths changed, we go again with more success. That’s more like it!
A few files (and a couple of amendments) later and the raw data has been cleaned, prepared, and is now ready for the analysis, but first we need to check that the variables look consistent over time. I don’t think there was anything I overlooked, was there? Jonathan spends rather a long time staring intently at the expenditure variables before declaring “looks good, we’re ready to go”.
Success! I won’t see the outcome of the analysis just yet, however – it will have to run overnight because the dataset is so large.
Even the best economists…
After a quick break for lunch we turn to the project which has taken up most of my time at the CEP, a study that uses Labour Force Survey data to analyse the relationship between austerity and workplace stress.
Much of the time working on this study has been spent looking for variables that can be used to reflect austerity, but today we’re seeing if we can address the fact that the occupation classifications in the data changed in 2010 by mapping the old classifications to the new ones. Given that austerity kicked off just as the classifications changed, this is really quite an important task.
We run the code which takes the mapping weights from the Office for National Statistics and apply it to our data, but we have a problem: the population of the mapped occupation cells are around five times the size of the population of the cells in the other years. Is it my code that’s the issue?
We follow up this possibility but it appears instead to be something to do with the mapping. “It’s not the end of the world”, Jonathan says, “so long as the proportions are consistent over time”.
However, while the proportions aren’t miles apart another issue emerges instead, with the variable values quite erratic from one year to the next. “Perhaps a three-digit classification is the way to go”. It’s reassuring to know that even the best economists run into issues.
For the next couple of hours I take my code and amend it to prepare the data at the more aggregated three-digit level. Although upon reviewing the data I can see that the disparity between the cell populations remains, the variable proportions are far less erratic than before.
It’s a step in the right direction at least.
Taking advantage of opportunities
As Jonathan has gone home by this time, I spend the next 45 minutes or so on a historical study looking at commuting patterns. This requires using records of bus routes from 1931 to plot the coordinates of bus stops. I guess that research isn’t all glitz and glam, but somebody’s got to do it!
From there I head across the campus to watch a lecture by Hassan Damluji titled “What Citizens of the World Can Learn from Nationalism”.
After all, why bother working at LSE and not take advantage of their ability to attract some of the best scholars and thinkers from around the world?
About the author
Ben Brindle is one of the UK Data Service Data Impact Fellows 2019 and is in the second year of his economics PhD at the University of Brighton’s Business School.
His research, which is funded by the ESRC’s South Coast DTP, examines how the labour market responds to immigration-induced supply shocks; through either technology mix changes, where firms alter their production techniques or output mix changes, where firms that use the abundant labour type intensively grow in size. To do this, he uses the Quarterly Labour Force Survey and the Annual Respondent’s Database.