My response to Mike Hedges AM Bevan Foundation blog on the use of statistics in programmes that tackle poverty
Mike Hedges, Assembly Member for Swansea East/Dwyrain Abertawe blogged at the Bevan Foundation on how a greater flexibility in interpretation and design of data should be used to enhance the identification of people in need. My response to it is available to view on the above link. It is expanded upon here in order to articulate the value that community development brings to this process.
I’m not sure using units smaller than LSOAs necessarily leads to more accurate identification. LSOAs are already small and have been used in the most recent editions of the WIMD precisely because they allow for a more nuanced sub-ward analysis of deprivation at community level. They are constrained by the arrangement of electoral wards above them, which would presumably be costly to rearrange. With such rapid growth in housing in some wards (Butetown springs to mind) the LSOAs need to evolve to reflect the new communities that spring up; this is not necessarily about the size of LSOAs, but the cohesion and sensitivity with which a ward is carved up into them. At a macro level it is less about the size of the LSOAs and, rightly as Mike Hedges points out, about how flexibly they are interpreted by the programmes that use such data and whether other data is eligible to complement WIMD and census data. The experience of Communities First (CF) is salient here.
CF, through its use of Results Based Accountability, requires a story behind the baseline. In essence ‘what else does one know about a community beyond what the statistics suggest’. This is welcome. I recall CF community development workers (CDWs) in the Dulais Valley citing broadband connectivity data that suggested it was among the most 2-3% ‘dis-connected’ communities in the whole UK. Data related to digital connectivity, whether it is use or availability thereof, is not an indicator that WIMD draws upon; though arguably with the increased shift towards online access to job searches and availability of financial products and transactions it is a key indicator that shapes deprivation. CF allowed for additional data and research to shape the argument for resources towards particular tackling poverty activities. CDWs do not merely raise awareness of such a statistic but are well-placed to interrogate the assumptions that it informs, such as the extent to which it affects accessibility to employment advice and job adverts, and the effect on morale, confidence and preparation for the recruitment process. In such an instance the story behind the baseline does not narrate itself, and certainly not on a collective basis.
The emphasis on the size of LSOAs potentially draws attention away from the underlying indicators that the WIMD draw on. Mike Hedges focuses on two housing related indicators: tenure and council tax band. This is particularly interesting for two reasons. Firstly, that the last WIMD in 2011 deliberately reduced the weighting in the calculation of the overall WIMD of the housing domain from 10% to 5% because it drew on census data from 2001 and this was felt to be less robust than it might have been. Thus, irrespective of which indicators are used, the key issue is that the weighting of the different domains reflects the proportionality to which different indicators cause, aggregate or reflect poverty. Secondly, tenure and council tax seem reasonable indicators to accompany the current housing domain indicators of overcrowding and presence of central heating. One might argue however that data related to affordability of housing might be more pertinent again; or even availability of housing. In respect of tenure, is the status of tenure or security of tenure that is a more pertinent indicator to levels of deprivation within a community? This reveals how politically-laden the identification of indicators actually is. Why is there no business start-up related indicator? Or self-employment related indicator? Whatever the indicator, the data has to be available consistently at whatever unit level is employed because the more gaps there are the harder it is to be flexible in the interpretation of data for which Mike Hedges calls. Again CF’s experience is helpful.
The gaps in ‘NEET’ data at LSOA level made for a very patchy understanding of even the statistical extent of the problem and provided for a muddled picture among CF clusters. If the extent of a problem is not accurately known, how can progress be accurately measured? Perhaps this is why WIMD does not use ‘NEETs’ as an underlying indicator.
Community development is crucial in advocating on behalf of less vocal and/or visible interests. In this way it is able to draw attention to other indicators that can inform the analysis and measurement of disadvantage. Issues about statistical rigor remain, such that there may be legitimate technical reasons why something cannot be used. But the advocacy role is two-way and CDWs can help explain to communities why indicators are not adopted. I recall working in a Gwent valleys community where there were concerns about the mortality rate from breast cancer in that sub-ward community. The data, the Local Health Board told us, was available at sub-ward level (this was in the pre-LSOA days) but to circulate it would risk revealing the identities of the individuals who comprised the statistics, which might be insensitive and distressing, as well as breaching data protection legislation and confidentiality protocols. My and others’ roles were to facilitate that dialogue. Did the unavailability of that data affect project planning? Or our understanding of the experience and psychology of, and services for, breast cancer? Possibly. But it was a reminder that there is always a human face behind statistics; human faces that can articulate the experience and knowledge that shapes the stories behind the baseline…if they are given a suitable, safe opportunity to do so. Community development helps create such opportunities and allows them to enter the political nexus that exists around debates related to disadvantage in a way that limits the extent to which that experience can be exploited for political gain.
Returning to Mike Hedges blog, it is extremely helpful that he puts statistics and policies’ use of them under the spotlight. The opportunity to participate in the construction and design of WIMD is one which should be more prominent than it traditionally has been. A more public profile would allow the debate about what is relevant in defining ‘in need’ to be pluralised, and this is crucial and goes beyond not just finding out where people in a pre-determined and possibly remotely-determined need are.