Data, Government, and a Little Bit of Cake – My Internship Story
Issued: Tue, 31 Oct 2017 15:34:00 GMT
Unit PhD student Megan McMinn is in her third year of her postgraduate studies and recently worked at the Scottish Government as part of an internship organised through the Scottish Graduate School of Social Science. Megan explains her experience working alongside analysts working at the heart of government.
In September I returned to my PhD following a 13 week internship within the Children and Families Directorate of the Scottish Government. I was placed within a team of statisticians, although the division is also made up of social and policy researchers and economists.
The research project
The team I was placed with collects, analyses and produces an annual publication of Children Looked After Statistics (CLAS) in Scotland. Under the provisions of the Children (Scotland) Act 1995, ‘Looked After Children’ are defined as those in the care of their local authority. Children are either looked after at home through a supervision requirement, or are looked after and accommodated away from home. The data are collected from each of the 32 local authorities in Scotland, and have been available at the individual level since 2009. The annual publication is based on only the previous year’s returns, and therefore at the country level there was little understanding of the pathways children take through the care system, and no way of knowing whether children were re-entering the care system at a later date.
The internship was advertised as further development of a longitudinal dataset, making use of the Looked After Children (LAC) Data Strategy to identify areas of research to cover during the internship, including the creation of a meaningful measure of permanence. However, once I started we decided that the longitudinal dataset that had been created, but not yet formally used, had a number of issues with it. Therefore, the main task of the internship became the creation of this dataset.
The yearly individual level returns contain demographics on the children entering, exiting or continuing care within a local authority area, details of their episodes of care, as well as the placements undertaken within each episode. When combining eight years of these returns, several issues with the data were identified. These included changing dates of birth, genders, conflicting dates, as well as children who did not appear in subsequent returns.
Producing new data
The final dataset contained information on over 50,000 LAC, which collectively experienced nearly 120,000 placements within approximately 56,000 episodes of care. After the dataset was finalised, I performed various analyses using it, so that the capabilities could be illustrated to future data users.
Firstly, the analyses reported in the annual publications were repeated to understand any differences the longitudinal dataset may have to the yearly ‘snapshots’. I also used survival analysis as a way of measuring the time spent in care required to achieve ‘permanence’ (either a positive destination once leaving care, or those still in care, but in a long-term, stable placement), and differences by sex, age, type of care placements, as well as differences by local authority. These were repeated for the 8 possible years of entering care, allowing for the impact of any policy changes to be evaluated.
The LAC data strategy highlighted a need to understand the placement pathways, and the stability of the various placement types (Looked after at home with parents, looked after by a friend or family, in foster care, residential schools etc). For this, I used a pathway mapping tool to visualise the placements over time, allowing for changes in the usage of different placement types to be seen.
During the creation of the dataset, I maintained a list of queries for each local authority on any inconsistencies found, so that these can be rectified at later returns.
Working within the government
The internship provided the opportunity to not only work on something completely different to my PhD, but to also experience working within the Scottish Government. I was treated as a regular member of staff, encouraged to attend team meetings, and Directorate ‘Huddles’ in order to learn more about the work going on within the government. Within the first month of the internship I had met with Paul Johnston, the Director-General of Education, Communities and Justice, who provided me with an introduction to the history of the directorate, and a discussion of the work I’d be producing during my placement. I also joined the analysts’ cake club which made Monday mornings a little bit better with home-made cakes!
During the internship I attended meetings with a local authority to gain some background on LAC; how children become looked after, where they are placed, how the data are generated, and how local authorities make use of their own data. The meetings highlighted the limitations of the current data collections, and differences between the practices of each local authority.
I also had the chance to be on the other side of a data-sharing application process. My longitudinal dataset was used to produce a smaller dataset requested by a university. Through this I was able to meet with the researchers, understand how they would be using the data, and discuss any limitations we had identified.
In my final weeks I was able to take on other tasks to support other members of the team, which included summarising data and drafting a briefing for policy analysts. I also had the opportunity to discuss my own PhD project with a policy analyst, and will be sharing the results once my thesis is finished.
At the end of November, I’ll be returning to the Scottish Government to attend the annual meeting of Local Authority Data Providers to present on my activities over the summer.
I would encourage social science students in Scotland to look at the opportunities the Scottish Graduate School of Social Science offers and other internships with policy makers and institutions.
Disclaimer: The views expressed in this blog are those of the author.
Megan McMinn is funded by a Medical Research Council studentship. The Social and Public Health Sciences Unit is funded by the Medical Research Council and Scottish Government Chief Scientist Office. The internship was funded by the Scottish Graduate School of Social Sciences.