Redesigning the output.
The latter solution is likely to avoid disclosure issues without losing too much detail and therefore is often used. Going back to the income-education tabulation, this could involve the merging of the undergraduate and postgraduate categories, as in the table below, thus reducing the risk of disclosure somewhat.
However, the zero values remain a spain rcs data potential problem (depending on the context) and would have to be tackled in some manner.
Given how important it is that outputs are not disclosive, the UK Data Service has a number of measures to ensure that secure data remains protected, the most significant of which is to limit access to secure data to within the Secure Lab.
The Secure Lab is a secure remote working environment (no points for guessing that!) which contains a range of software, from Word and Excel to Stata and R, that can be used by researchers to conduct their analysis and write up their results.
When outputs are at a “publishable quality” – that is, written up for publication or presentation – they have to go through a process called Statistical Disclosure Control (SDC), or ‘output clearance’, before they are released to the researcher. Requested tables, graphs, maps, or whatever else, are thoroughly checked to ensure that they are not disclosive.