James Scott, one of our SecureLab experts, offers practical advice on how to ensure your outputs are released quickly
It should go without saying that User Support and Training staff want to facilitate the release of high quality outputs from SecureLab. As you will remember from your training, for legal and practical reasons, there are certain procedures to be adhered to, and criteria to be met when using controlled data in a secure environment such as SecureLab, and this is why we must perform statistical disclosure control (SDC) checks.
1. Folder and file-naming conventions are important
This might seem relatively unimportant, perhaps even trivial, but I can assure that it’s not. The standard sub-folder naming convention looks like this (project number_year_month_day, separated with underscores): 998701_2021_03_15. This is the subfolder you need to create and then add your output content to. This subfolder should be in the ‘SDC’ folder. Sticking to the convention means that the newest output will always be at the bottom of the list.
When it comes to naming your output file(s), you should use a meaningful name that corresponds with the one you used on the Output Request form, not just something generic like ‘output’ or ‘release material’. By adhering to the folder and file-naming conventions you ensure clarity, we can be sure that we are checking the correct file(s), and avoid wasting resources which will have a knock-on effect for all SecureLab users.
2. Supplying underlying Ns is a must
We cover this pretty extensively on the training course, so you might be surprised to learn that we receive a number of outputs each week that do not include (all of) the underlying unweighted Ns. We must ensure that the threshold of 10 is met; not just in tables but for percentages, regressions, scatterplots, minima, maxima, median…everything! If these are not supplied, we cannot release the output and will have to contact you, delaying the release of your output (and requiring double the resource for us, as we will have to re-check the output twice more!).
3. Provide accurate explanations and clarity
We ask that you submit publishable outputs. The list of acceptable types is on our website, and in the SecureLab Handbook. Tables must be properly formatted and sufficient explanation on data used, methodology, and what each aspect of the output shows must be included. We can’t be experts on every dataset, field of inquiry, or statistical technique that researchers may employ, and if we’re not 100% sure what we’re looking at then we’ll contact you for more information, delaying the release of your output.
4. Include accurate data citation
However ‘safe’ we consider the content of your output to be, it won’t be released if you haven’t included the full data citation, including the DOI (digital object identifier) number. Citing data helps to ensure studies continue to be funded and the resultant data made available to researchers, so is incredibly important for all of us!
5. All tables, figures, graphs, and outputs must have meaningful names and be numbered appropriately
We hold a huge number of datasets, so we cannot be sure what every variable name used in these datasets means. Therefore, researchers must give all variables in every table, graph etc. meaningful names that we can understand at a glance, alongside sufficient textual explanation. Moreover, all tables, figures etc. must be numbered. Imagine that we have to contact you to make changes to a figure, and because it isn’t numbered we have to refer to it as “the figure on page 32”. If you add explanation, remove parts, change parts etc., that figure may no longer be on page 32, causing potential confusion (in the interim we will have a seen a great many outputs from other users so yours will not be fresh in our mind!)
6. Carefully triple check you have fully addressed all the tick box list before submitting – we cannot release any research if there are any discrepancies or gaps
Finally, when you are in the process of making your output request via the form on our website, please please please do read the tick box list carefully. These tick boxes are intended to serve as a final-stage reminder of some of the minimum requirements. It makes the team’s collective heart sink when we open an output and it is clear from the outset that some of these criteria haven’t been met, despite the relevant boxes being ticked. If you work your way down the list and think “I’m not sure if I’ve done that” or “I’ve submitted something inappropriate” then you should halt the request and go back to your output to check and/or amend as necessary, not continue with your request! This tick box list is there to help you get your output released quickly!
If the criteria are not met, then we will ask you to make changes and your output request will go to the back of the queue. Factor in that during busy times it may take three or four days for us to open your output in the first instance, and further – unnecessary – delays will be frustrating for everyone.
The tips above are for your benefit and ours. The ideal scenario is that you request your output, we check it, and release it to you immediately. The tips will make this scenario much more likely, and that’s better for everyone.
More tips on using SecureLab efficiently