Linking local labour market statistics to Understanding Society
Dear Understanding Society.
I'm interested in accessing local labour market statistics, and linking them to Understanding Society (waves 1-5) in roughly the same manner as this recent paper: https://www.understandingsociety.ac.uk/research/publications/working-paper/iser/2015-20
I was just checking that applying for the 'Special License' from the UK Data Archive is the way to access this data?
#1 Updated by Gundi Knies almost 4 years ago
- Category set to Data releases
- Assignee set to Gundi Knies
- % Done changed from 0 to 50
from your question it is not clear whether you want to use the data that Alexander Plum and I used for the analysis? Or do you want to link your own local labour market statistics, if so, at what level of geography?
#3 Updated by Tomos Robinson almost 4 years ago
Thank you for your quick reply, I realise now I wasn't very clear in my previous message!
For context, my research question concerns the relationship between maternal labour market supply and child adolescent outcomes. As this relationship is almost certainly endogenous, several papers in the literature have used local area labour market conditions as the first stage in an instrumental variable strategy (as usual, several other papers don't like this IV strategy because of the potential of self-selection into areas, but ah well!)
Like Alexander and yourself do in your paper, ideally I would like to use a few levels of geography for robustness (LSOA, MSOA, LAD) but I realise this potentially isn't possible!
In terms of using the data used in your paper, if you were willing to share I assume I would still have to complete the special license?
#6 Updated by Gundi Knies almost 4 years ago
- Target version set to X M
- % Done changed from 50 to 90
there are restrictions on accessing special licence data in combination, see the data access strategy (follow the link included on https://www.understandingsociety.ac.uk/documentation/getting-started) so it is good that you think carefully about what your requirements are before applying.
Alexander and I cannot share the data we used with others outside our project: onward sharing of special licence and secure data is (for very good reasons!) not permitted.
You could go down different routes. The most flexible approach is to apply for Special Licence access to the LSOA geographical IDs in Understanding Society and use them to link your own local employment data at LSOA level. Then aggregate from LSOA to MSOA and LAD using appropriate, publically available look-up files. Effectively, you apply to use only one SL data file and use it in conjunction with the EUL Understanding Society data and create all the extra data yourself.Some considerations:
- You could also apply for UKHLS households' MSOA codes to be able to be more accurate about where households are located (although LSOA are nested within MSOA boundary changes can occur).
- Applying in the same way for LAD codes may push the sensitivity of the requested data beyond the special license threshold into secure data access.
- Depending on which external LSOA-level data you want to use you need to apply for either the 2001 or the 2011 Census IDs. Applying for both is not possible under Special Licence (secure access may be granted).
The above procedure is very time consuming and cumbersome (in particular when you are not familiar with data linking). An alternative would be to apply to use the Geographical Accessibility data, which offer some readily available local labour market indicators. The data are available under Special License. You would need to make some amends however, as the data do not include the LSOA identifiers itself. The data can also not be used to aggregate to a higher scale.
- You could additionally apply for LAD geographical identifiers and link your own labour market statistics at LAD level. This could push the data into secure access, however (you'll have to try and see how the Data Access Committee decides).
Links to the geographical data products are here: https://www.understandingsociety.ac.uk/about/data-linkage
Hope this helps,
#8 Updated by Tomos Robinson almost 4 years ago
Thank you so much for your detailed reply!
I think the most appropriate way forward for me is to apply for the geographical accessibility data and ask to include the LSOA indicators.
As you imply, I have absolutely no experience linking complex data such as this, and I would like to avoid secure access if possible (because it is probably a bit of a hassle!).
#9 Updated by Gundi Knies almost 4 years ago
if you want both LSOA codes and Geogrphical Accessibility data, you will need to put both of them into the basket and write into the SL application why you need the LSOA code and why you need the Geographical Accessibility data. From what you wrote earlier it is not clear what you hope to achieve by adding LSOA codes: you cannot use it to aggregate the labour market stats to the higher units. If you just want to know who lives in the same LSOA at any point in time and who moves LSOAs, there are indicators for that in the Geographical Accessibility data. Please consult the Geographical Accessibility user guide to see what variables are included.
#10 Updated by Tomos Robinson almost 4 years ago
I've just consulted the Geographical Accessibility user guide, and found two of the 'local labour market indicators' you previously mentioned that I could potentially use:
- Travel time to Employment Centres
- Number (or: proportion) of people in jobseekers allowance in each LSOA
However, I'm still a bit puzzled whether I would need the LSOA indicators too?
In email #6 you said: "You would need to make some amends however, as the data do not include the LSOA identifiers itself"
However, in email #9 you say "It is not clear what you hope to achieve by adding LSOA codes"
Am I correct in saying that if I just want the two variables mentioned above (Travel Time and Jobseekers) I wouldn't need the LSOA codes?
If so, I think this is what I will put on the special access form.
Thank you again for your help!
#11 Updated by Gundi Knies almost 4 years ago
Sorry if I have been unclear. I said you'd need to make amends in response to your proposal to also use MSOA and LAD level data: when you want to rely on the Geographical Accessibility data alone that is not possible. The implicit amendment to your plan is therefore to drop the much more ambitious aggregating and linking of data at other scales. The only reason why you might want the LSOA codes is that you need to link additional data at that scale (and this may affect SL eligibility).
Hope this clarifies,
#12 Updated by Tomos Robinson almost 4 years ago
Yes, I think since talking to you I've realised my plan of having a similar dataset to yours was a bit ambitious!
I think I will stick to the Geographical Accessibility data available.
Thank you very much for all your help.