By Terence Hogarth, ETF Skills Lab Scientific Director & University of Warwick
Introduction
Advances in skills anticipation have accelerated significantly over the relatively recent past. Natural language processing (NLP), more than anything else, has transformed research undertaken to support skills anticipation, providing access to data previously out of the skills analyst’s reach.
Computer scientists and software engineers have applied their capabilities to skills anticipation, so that now we can automatically analyse online job advertisements and so much before besides - such as bibliographical databases and vocational curricula. Text data from such sources can now be collated and classified in near real-time.
When combined with more traditional forms of skills analysis, such as skills surveys and forecasts, the use of NLP techniques can provide unique insights into the demand for, and supply of, skills. This is readily evident from the European Training Foundation report on the Future of Skills in ETF Partner Countries.
With so much skills anticipation data now within the analyst’s easy grasp, this inevitably begs questions about how it can be better marshalled to support policymaking. Another important consideration is how to avoid situations in which the availability of so much detailed information about the specific skills required in a particular job creates situations of data overload - a situation, if you will, of not being able to see the wood for the trees.
How to better exploit the improved data
Making the most of the increasingly available data and analysis to inform skills anticipation-related policymaking is therefore an important challenge. Getting it right can hope to significantly improve outcomes for workers, jobseekers, employers and governments.
Recent reviews of how to connect research to policy have drawn attention to, amongst other things, the importance of engagement between policymakers and researchers - from both the private and public sectors - to develop shared agendas. This requires the building of effective professional partnerships and data sharing, such as granting access to administrative databases and delivering easy access to research findings.
At a more fundamental level, the provision of sufficient funding to support policy-based research is an important prerequisite. The need for a shared language is also critically important. Any tendency for researchers to lapse into academic jargon will in all likelihood obfuscate the communication of findings. Meanwhile, a failure of policymakers to understand the theoretical perspectives that inform research designs is equally likely to damage the chances of successful knowledge exchange.
Much of the above prescription is about structures and governance. It is about regulation which provides the basis for creating forums in which engagement can take place and providing the resources to pursue shared interests. Where it works it creates a community of practice of researchers and policymakers each cognisant if not sympathetic to the other’s perspectives and requirements. The ETF’s Multilevel VET Governance Toolkit (2020) provides guidance here.
An important consideration here is that partnerships and ‘communities of practice’ take time to develop. There may also be an attendant danger if governance structures become set in stone with the result that there are insiders and outsiders. Insiders come to represent established interests already active in various committees, working groups and so on, whereas outsiders may be the ‘new kids on the block’ with little access to established structures.
Looking further ahead
Future technological change may yet throw up new, transformational techniques with implications for the analysis of the demand for, and supply of skills, and the degree of match or mismatch between them. The governance of skills anticipation needs to be continually open to ideas from the ‘new kids on the block’ mentioned above.
While there is every likelihood that skills anticipation will need a combination of new techniques allied to more traditional approaches to effectively inform policymaking, you can never be sure. Policymakers need to be open to the application of new research methods to the challenges with which they are faced.
Where there is, perhaps, more certainty with respect to the problems and puzzles policymakers and researchers want to solve. Many of these seem universal over time and geography revolving around, as they do, the timely anticipation of emerging skill needs: what are they, how far in advance can they be predicted, how can the relative importance of future skills be weighted (some skills are likely to be more important than others), and how can the supply-side respond in sufficient time to minimise economically damaging skill mismatches.
If this provides the core of a shared agenda for both policy and research, then there is the basis for shared, mutually beneficial engagement that is open to newcomers.
Please log in or sign up to comment.