Best practice for linking up indices of deprivation for earlier cohorts

[cyskirrow]

We're planning to control for indices of deprivation in our analysis of early cohorts from NPD in ECHILD. I can see that there are IDACI indices from 2016 onwards in the NPD absence tables, but this is late for the defined cohort we would like to be looking at - many of whom will have left schooling by then.

What is the best practice approach for linking up deprivation indices (e.g. IDACI, IMD) to a broader cohort from NPD? Is there a best practice way of, for example linking up IMD to regional identifiers (e.g. MSOA provided in NPD) data?

Thanks for your help!

[tony-stone]

Hi @cyskirrow

IMD measures are calculated for LSOAs, a geographic unit smaller than MSOAs. Due to our data providers' / host's policies on the deidentification of data, the smallest geographic unit within ECHILD is MSOA.

However, the Department for Education, appreciating the need for IMD information in analyses, attached a number of decile of IMD measures (ie maintaining the deidentified nature of the data) to a variety of NPD tables - not just the absence data.

I don't recall precisely which tables contain these IMD measures, and the best table(s) to use would depend on how your cohort is defined, but Spring census tables are often used to define populations and do contain IMD measures.