Longitudinal Education Outcomes Synthetic Data: Version 1, 1995-2021

<p>The 'skill bar' required to use<span style="font-style: italic;"> Longitudinal Education Outcomes (LEO) </span>data is high, with users having to have a good understanding of the content of the data before knowing what it can be used for and how to use it effectively.</p>
<p>DfE have therefore created low fidelity synthetic data (also known as "dummy data") which replicates the format of the LEO data. This enables potential users to gain a better understanding of the structure and content of these highly complex data prior to making a time and cost intensive project application. It also facilitates the preparation of code while waiting for applications to use the LEO data to be processed by the Office for National Statistics (ONS).</p>
<p>Access to such low fidelity synthetic data will lower the barrier to entry for researchers considering using LEO data, increasing and speeding up the production of additional policy and practice relevant research for the benefit of the public.</p>
<p>In creating a low fidelity synthetic version of LEO, we have created a new dataset which retains the overall structure of the data (e.g. numbers of data tables and variables, and variable names), and the structure of each of the variables it contains (e.g. variable types, missingness, and counts), but which does not preserve any relationships across variables (e.g. correlations). Importantly, the resulting synthetic data does not contain any personal data, i.e. real data relating to an identifiable person.</p>