I’m quite capable of writing seriously and in a proper academic style. But this is a blog. If I can’t have fun and speak my mind, why bother? Despite that, I think there may be some serious points and interesting finds here. I don’t think anyone has done this before: if I’m wrong, please let me know.
This mini-project used three data sets, all available to the public. One (a portion of the 2011 National Census) has questionable data. The next (CRIDE’s 2014 survey) comes with a warning attached that it should only be used for analysis and debate rather than drawing any hard conclusions. The last (NRCPD’s May 2015 registration breakdown) only represents those interpreters who choose to register with NRCPD. Therefore the following should also be taken as an exploration – stuff that is or is not likely – rather than solid fact. I’m the first to point out other people’s terrible use of stats and I hope I’ve been equally critical and honest about my own dalliances.
If you’re already bored you can skip ahead to the pretty maps.
NRCPD recently began releasing a breakdown of qualified registered interpreters by region for the first time. When I saw this, I realised there is the potential to have a pop at some very old questions: how many BSL users per interpreter are there? and where is the most work for interpreters? and where do Deaf people compete most for interpreter availability? The answers would obviously be of interest to those invested in Deaf Industry strategy and marketing, but there are political concerns as well. For example, I regret that it has been suggested to me more than once that there are “too many interpreters” (only by interpreters, obviously).
But counting registered interpreters is the trivial part. The question of how many Deaf people there are and where they are has always been problematic, because we don’t all necessarily agree on what they are. Humans always come unstuck on methods and definitions, because words don’t mean anything in themselves. For areas, grey is the new black. Determinations of identity are always going to be muddy, because people are fabulously muddy.
But if you don’t ever try, you don’t progress. Despite these issues, there is always some kind of data. Data trumps ideology every time. You just have to pull your finger out.
The apparent finding of the 2011 National Census for England and Wales (released in early 2013) that there are only about 22,000 sign language users in those countries triggered a series of robustly disbelieving dismissals from a range of respectable organisations such as the BDA, RAD and SignHealth. Received wisdom had said that we have anything from 50,000 to 125,000 “BSL users”, although it was never quite clear to me how we knew that or what it meant. Does it include anyone whose first language is BSL, regardless of audiological status? Do CODAs, hearing children of Deaf adults, count as “Deaf” if we are linguists concerned with native languages? What about hearing late learners of BSL?
Not long afterwards, when the separate Scottish Census found 13,000 “signers” in Scotland alone despite the lower population density, the BDA decided they could extrapolate that figure out across the rest of the UK, arriving at a new “more accurate than ever before” figure of 156,000. I’ve not been able to locate any justification for how this was reckoned: I suspect it was a back-of-the-envelope calculation on the order of [British/Scottish]*13,000, because using 2015 numbers, I immediately get 157,374.
Regardless, the main problem with the Census, just like many lesser surveys, is that it asked a problematic question (“What is your main language?“) and produced problematic answers. The responses from sign language users were bundled into one of three pigeonholes: “Sign Language (British Sign Language)”, “Sign Language (Any Other)” and “Sign Language (Any Sign Language Communication System)”. I have no idea what that third category means or how it differs from the second. Is “Other” referring to non-British signed languages, to “home sign”? Is “Any System” talking about SSE, Makaton, Paget-Gorman? We don’t know, and this is even before you start to ask questions about the accessibility of the census-taking. There is some irony in a project which asks people in written English whether they can read written English.
You can even ask hard questions about the nature of asking hard questions: does everyone understand your question the same way as you do? People have complex motivations when answering questions so closely linked to their personal identity. How much bias do you introduce by explaining what you’re looking for? Most frustratingly of all for a minority community is the monolingual’s assumption that we all have a “main language”. If 99% of the UK population spoke Volapük 49% of the time, this Census might have completely omitted to notice.
At any rate, the numbers may well be off the mark, but we could ask ourselves whether or not they are at least consistently off the mark. It was the wrong question: was it more wrong in Birmingham than Cardiff? The strengths of a census over a mere survey are that (a) absolutely everyone is supposed to participate, with heel-draggers systematically hunted down, and (b) it is rolled out across the country near-simultaneously using the same collection methods and identical questions. The numbers are probably not an accurate representation of Deaf BSL users, but we are interested here in relative geographical distribution, so we still might be able to usefully compare regional ratios to each other.
The only other systematic and in my view reputable source of national data that concerns deafness and sign language use is the annual survey carried out by CRIDE, a consortium of researchers from NDCS, BATOD and UCL, which has engaged with Local Authorities (LAs) over four increasingly successful iterations to try and map out services provided to d/Deaf children and young adults. Those of us with a specialist interest in Deaf education, and against a background of, shall we say, highly optimistic calls for only fully qualified interpreters to be used in mainstream classrooms, might find some interesting products from a blend with regional interpreter stats. At the very least, a different source might serve as a counterpoint to the Census data, even if it is not directly comparable.
CRIDE also attempt to track different types of specialist teaching and support work staff who engage with the children and young adults, including (in theory) interpreters, Teachers of the Deaf, Communication Support Workers (CSWs) and Teaching Assistants (TAs). It was very much my desire to explore those statistics in the same way: however, a regional breakdown is not provided, only national (England versus Wales). I have contacted NDCS asking if they have these breakdowns and if they are willing to share them: if they respond, a further post will look at the CSW issue.
Signature/NRCPD is a voluntary regulator. It is not compulsory for practising interpreters to register with them and no amount of empty rhetoric about “integrity” and “real interpreters” will change that. There are many unregistered but qualified and/or competent interpreters working in the community, but we have no idea exactly how many or where. We could assume that unregistered interpreters are spread fairly uniformly and do not roam the country seasonally in one giant warband like nomadic hunter-gatherers pursuing mammoths, although that does sound like fun and maybe we should try it.
On the other hand, despite the implications in certain interpreter training curricula, interpreters are not isolated, powerless and entirely unaccountable individuals, unable to make ethical decisions without external sources of authority: they do interact with each other in their local communities. They swap ideas, influence each other. So opinions about exactly what constitutes “professionalism” might indeed vary significantly on a regional basis, in manner similar to the anecdotal ASLI/VLP geographical split. We just don’t know.
This exercise will hold on to a background assumption that unregistered interpreter populations correlate with registered ones, but as it is NRCPD data being used, we will try and confine analysis to implications for NRCPD-registered interpreters only. We will also try and avoid talking about ourselves in the first person plural.
Limitations and assumptions
- Scotland and Northern Ireland are not considered here. They both have their own Census, and the Scottish interpreter population is split between SASLI and NRCPD.
- As discussed above, the Census data about the number of sign language users is questionable, and CRIDE’s is only concerned with children and young adults. However, while individual figures per region might be inaccurate or tentative and not represent correct magnitudes, I am assuming there is still some value in directly comparing regions to each other.
- Again, the Census categories for sign language use are flawed. A decision was taken to use the “Sign Language (Total)” figure rather than “Sign Language (British Sign Language)”. Interpreters do work regularly with a range of communication needs, including users of Sign Supported English and migrants for whom BSL is a fourth, fifth or sixth communication method, as well as British people who can sign but identify as first language English users; furthermore, we do still have older members of the community for whom the term “BSL” was a new and in some cases unwelcome invention. We will refer to the group as “SL users” to make it clear we are not necessarily talking about the Deaf community or BSL.
- The Census data is from 2011, but the earliest regional breakdown of NRCPD interpreters is from February 2015. It is assumed here that Deaf populations have not migrated considerably in the last four years.
- The CRIDE summaries do not give a full regional breakdown of the proportion of audiologically deaf children who are BSL users: a blanket figure for England is given but no figure for Wales. Details of how this was treated are given in the methods below. CRIDE also point out that 20% of the Local Authorities in England could only partially account for the language use profiles of their children: one reading of the summary is that CRIDE assumes that those unaccounted for fit the same distribution of signing/non-signing as those who were explicitly counted.
- A definition of the NRCPD and CRIDE regions could not be located. It is assumed here that like the Census, they mean the former Governmental Office Regions (GORs).
- The GORs vary in geographical size, population, infrastructure, political control and economic conditions including employment rates and provision of public services. A simple estimate of BSL users per interpreter is not the same thing as a measure of “accessibility” and does not take actual demand into account. The apparent size of a region also has no direct relationship with population clustering (urban areas) so the fancy coloured maps are really just for giggles.
The QS204EW table in the 2011 Census database was queried for a breakdown of total “usual respondents” (England and Wales inhabitants over the age of 3, ie. the region’s population) and the number of “Sign Language (Total)” respondents for each GOR. These figures were tabulated against NRCPD’s May 2015 breakdown of registered qualified interpreters by region. Trainee interpreters were not included.
These three columns were sufficient to produce variety of tables which estimate:
- the proportion of sign language users to each region’s general population;
- the ratio of sign language users to registered interpreters;
- population and regional standard distribution scores.
Parts of those tables are listed below along with some lurid and arguably misleading choropleth maps.
The CRIDE 2014 UK-wide summary report provides a number for deaf children and young adults receiving Local Authority services in each region, but this refers to audiologically deaf and hard of hearing children and not specifically to users of signed communication. They did ask about the number of sign language users “belonging” to each LA, but no detail is given below UK member nation levels. For England, it is stated that deaf children who “use sign language as their main language or in some combination with another language” constitute 9% of the “belonging” children: no indication is given that this proportion varies significantly across the regions. A figure for Wales is not given. I contacted CRIDE representatives and asked them if they have more detail, but while waiting for a response and for the purposes of this simple exploration, I have simply estimated the number of sign language using children in each region by taking 9% of the total.
These figures were again tabulated against NRCPD’s May 2015 breakdown of registered qualified interpreters by region. Trainee interpreters were not included.
The ratio of SL-using children to registered qualified interpreters was then calculated per region along with standard deviation from the mean. Some exploratory cross-referencing was also made between CRIDE and Census data.
- To my great surprise, London had the worst ratio of SL users to interpreters by a margin, despite having the second largest overall population to recruit trainees from and being the UK’s prime target for economic migration, but not by a hugely significant margin (1.25 standard deviations). It also has the highest number of SL users. Looking at the ratios of interpreters to the general population (not tabulated here), London’s is lower than the mean but not significantly lower (-0.85 standard deviations).
- London’s percentage of SL users in the general population, however, varied significantly from the mean by 2.1 standard deviations: either the Deaf community significantly clusters in London (ie. even more than everybody already clusters in London) or SL-using Londoners found the Census more accessible for some reason.
- The mean percentage of SL users per region was 0.041% (1 person in every 2,439) and no other region than London varied from this significantly. If you agree with the apparent professional and political consensus that the Census underestimates SL use by at least a factor of 3 and possibly (if the BDA’s Scottish Census extrapolation had merit) a factor of 7, the ratio of BSL users per region would be closer to the order of 1 person per 500. If we assumed that there were really 75,000 Deaf BSL users in the country instead of the Census’s 22,000, then there might be a worst case of 145 users (of all ages) per interpreter in London and and a best case of 63 per interpreter in the East Midlands (rounding to the nearest entire person). The “BDA theory” of 156,000 would boost that to nearly 300 SL users per interpreter in London, which seems extreme to this London-based interpreter.
- The best ratio for potential interpreter service users is found in the East Midlands, where the Census found less than half London’s number of SL users per interpreter, with the South East not far behind. The ratio of interpreters to the general population in the East Midlands is close to being significantly different from the mean (1.98 standard deviations) but as mentioned, the proportion of SL users to the general population does not differ significantly there (0.23 standard deviations). Something in the East Midlands water produces a lot of interpreters.
- With these data, London does not score very well again, but is not the very worst. The results are bottom-heavy this time. If a political will to enforce fully qualified and registered interpreters for education services were somehow carried out, the relatively worst region to be a deaf child would be Wales, followed closely by Yorkshire & the Humber and the North East (Signature’s home region).
- The best prospects for SL-using children whose parents want qualified interpreters are again in the East Midlands and the South East – Wales has double the number of children per interpreter in comparison.
- No region’s ratio of children per interpreter varies significantly from the mean in the typical social science sense of “significant”, with the East Midlands at -1.36 standard deviations and Wales at 1.4. Nor did any region’s ratio of SL-using children per general population vary significantly from the mean.
- Comparing CRIDE’s estimated number of SL-using children in each region to the Census estimate for the whole SL-using population was interesting. If you assumed that both the Census and CRIDE were absolutely accurate, 23% of the SL-using individuals in Wales are children and young people; the lowest case would be 14% for London (bearing in mind a suggestion above that the London Census figure for SL users might be anomalously high in relation to other areas, or not). The Census’s age count of the entire population indicates that overall, 21% of the population were aged 0-17 in 2011 (which explains why I have such a hard time on buses). Either the Census count for SL users is not as wide of the mark as some believe, or the CRIDE tally of SL-using children and young adults is equally underestimated. CRIDE’s summaries do suggest that some LAs’ data leave some children unaccounted for (for a variety of reasons including known unknowns and unknown unknowns), but not however by a factor of 7 or even 3.
Conclusions and recommendations
As with language, interpretation of data is subjective, depending on your identity, history, position. Are you an interpreter entirely concerned with potential work who doesn’t like competition? Then London may be the best place, relatively, to look for it. Are you Deaf and struggling to access services? Then you may already be relatively fortunate if you live in the East Midlands and should consider avoiding London. Do you have deaf children who use some form of signed language? Then the Home Counties (except Essex) may be a relatively privileged place to live (no real surprise) and your prospects for qualified support are substantially more limited in Wales and the north east of England. There is some clear, if not extreme, variation of interpreter accessibility across the regions and we might therefore also ask whether the private training industry is sufficient or competent to meet actual regional needs. Finally we have a vague suggestion that the Census estimate of sign language users is not necessarily as low as some people want it to be, but this needs further investigation.
This analysis has a substantial number of limitations, discussed above. One of many things also seriously lacking from NRCPD is access to the length of experience of interpreters and proper data about turnover. We know from the interpreter population explosion that the average length of experience must have dropped considerably since the 1980-1990s when a double-digit number of hoary professionals had the entire field in their pockets, which might partly explain an anecdotal perception in the Deaf community that interpreters are not as good as they were in past (nostalgia was better back then as well). You may well be stubbing your toe on an interpreter every five minutes in the South East and East Midlands, but for all we know from the data above, they are all inexperienced with the printer toner on their NVQs still warm, and all the ones with 25 years of experience are huddled in Cornwall in some kind of primitivist commune. We need, and NRPCD must provide, numbers and qualitative data about turnover (interpreters who leave and why they leave, not just a running total), demographics of the interpreting workforce, levels of burnout and an analysis of demand before anyone can realistically model the state of play. If NRCPD or UKCOD cannot organise this, then perhaps they will have to make way for organisations which can.
In the education domain and with regard to the ratio of SL-using children to interpreters, the best current case (East Midlands) has one existing interpreter serving 3.4 children on a full time basis. This is already unrealistic and we know that the numbers of children are probably slightly under-reported. Furthermore, we know from CRIDE and from earlier work such as the 2010 BATOD DESF survey that only a miniscule fraction of the staff regularly engaging with young deaf people are qualified interpreters. CRIDE estimate the total number of signing deaf children and young adults for 2014 at about 4,300. This means that if in order to fulfil the current demands of full-time education, all of the current 900-odd registered interpreters were somehow forcibly removed from their current assignments and made to work in education (depriving every Deaf working adult and public service user from interpreter access) there would still be four-fifths of children left wanting – and this is assuming that every existing interpreter is actually available to work full time (it has been suggested to me anecdotally that ‘terps are typically part-time supplementary earners for their households).
It has been my observation that those demanding qualified interpreter provision for deaf children are entirely silent on the issue of how we might fund and administer the development of a workforce of a few thousand more qualified interpreters, and this is not even considering post-compulsory education (all of the FE and university populations, while we still have them). By my back-of-the-envelope reckoning of current industry rates, and making the ludicrous assumption that there are suitable recruits, it would cost approximately £28 million to drag enough holders of Level 2 BSL up to RSLI status for every schoolchild. You could cut that in half if you allowed a flat ratio of two children per interpreter across the country; you might be able to cut it in half again if you ejected the current powerbrokers of the interpreting curriculum and redesigned it from scratch in a fantasy socialist regime that restored the public sector. So that’s just £7 million – pocket change.
Is any of that likely to happen if we’re content to sit back and let market forces magically sort it all out? I’ve seen interpreters blithely tell potential trainees that they can “always” fund themselves. To me that sentiment tastes of privilege.
Perhaps instead we should be asking why we abandoned Deaf schools in the first place.
I hope this is useful. I encourage people to share, use, criticise and improve on this, but if you use it elsewhere, please credit me (and I’d enjoy hearing from you). Be well.
(Edited 21/3/2016: Some of the ideas above are updated and re-checked in a later post.)