Another problem quickly emerged as a result of the detailed scrutiny of the PTA question. One of the first CDs to be scrutinised consisted of small communities that were close to a regional centre. Many of the PTAs on these forms were said to be ‘shopping’ or ‘visiting family’ in the regional centre. There was no way of knowing, a priori, whether these people had gone only for the day or whether they had gone to stay in the regional centre for a while, and might have been counted there. In this case, the CFO who had been in charge of this CD happened to be in the CMU for the day. His local knowledge of the communities was tapped. His judgment was that people from these communities normally just went into town for the day, and that, moreover, the enumeration in the regional centre happened at a different time from the enumeration in these communities, so on both counts it was unlikely that these PTAs would have been ‘caught’ in town. Accordingly, they were all moved to Question 12 on their forms. In this case, it was not thought necessary to go through all the forms for the regional centre to try to find people counted there.
It is notable that the CFO had not himself made these judgments about moving PTAs before the forms were sent back to Darwin. Had he done so—and had he asked the CIs to go back and collect full details for the PTAs who were moved inside the form—many more people in the population of this CD would have had their full details recorded. Unfortunately, I did not think to ask him why he had not taken this step. There are several possible reasons: he had not understood the circumstances under which PTAs should be moved; he had not checked the forms closely (as per the CFO checklist) before sending them to Darwin; he simply ran out of time and had to make a judgment that it was better to get the forms in by the deadline he had been set than to spend more time chasing the details of PTAs. Given that the original CIs who had completed the forms might no longer be available, this could have been a lengthy process.
Similar problems kept coming up. It became clear that many CIs had not followed the instruction—or had not been instructed clearly enough—to move people from the PTA table in clear cases where they would not have been counted elsewhere, such as being away at ‘sorry business’. It was also clear that most CFOs had not questioned the CIs’ original allocations of people to the PTA category. In most cases, the reasons for moving people—or not moving them—were undocumented.
It also became clear that the prolonged nature of the count, combined with the levels and range of mobility, posed real problems for judgments about whether PTAs were likely to have been counted elsewhere. Theoretically, it should have been possible to check the advance schedules that stipulated when each community in each CD was to be counted, so that, for example, a person listed as a PTA from a community counted in July, who was said to be at a community that was counted towards the end of August, stood a good chance of having returned home before the latter count took place. It would be reasonable then to move such a person back into the form. These judgments depended on schedules being adhered to and on the CFO (at the very least) being aware of the schedules for their own region and also for surrounding regions and major population centres such as Darwin and Alice Springs. The CIs who were assigning people to the PTA category certainly did not have this information and so were not in a position to make such judgments (see my comments in Chapter 4 on my own dilemma over this very point), and there is very little evidence that the CFOs attempted to make such judgments in the field. In any case, most of the advance schedules underwent substantial alterations in the field, so it was almost impossible, in practice, for any CFO to be aware precisely when communities outside their sphere of responsibility had been enumerated.
In future censuses some of these problems could be ameliorated—but not eliminated—by better documentation, on several counts. Firstly, it should be documented whether people who are said to be ‘visitors’ at a dwelling have been asked how long they have been visiting for and whether or not they have been counted elsewhere.[3] Secondly, more details should be asked about where a PTA is, how long they have been away and when they are likely to return, and this should be documented. In particular, if a PTA is said to be in a large town—for example, Darwin or Alice Springs—it becomes well nigh impossible to crosscheck whether they have been counted unless some specific location within the town is recorded. Thirdly, the ‘short form’ used to enumerate the ‘homeless’ should allow for a person’s usual place of residence to be recorded. Many people who go temporarily to Darwin and Alice Springs camp out. It is not possible to crosscheck whether such PTAs have been counted if their usual place of residence is not recorded (see Chapter 5).
The situation could also be ameliorated by better use of technology. If the schedule were in electronic form, and if it were kept updated by the CFOs, and if every CFO had access in the field to this database, it would be a straightforward matter to make judgments—in the field—about the likelihood of PTAs being counted elsewhere. The CFO would simply have to look at the database to see what progress had been made in the CD to which the PTA was said to have gone. If that CD was being counted at more or less the same time, the chances would be high that the PTA would be counted as a visitor there. In the case where a large number of PTAs were said to be in a particular community, it would be possible to check—in the field—whether they were being picked up there.
Thinking further along these lines, it might also be profitable to institute a ‘funerals and festivals’ database, which would also be updated constantly by the CFOs on the basis of local intelligence, detailing where and when ceremonies and festivals were taking place, and which communities were affected, in terms of being the site of such an event or the source of many attendees. Such a database would make planning around these events much more manageable.
Schedules set in stone at the beginning of the census exercise will never be an effective tool in the context of the remote Indigenous census. There are just too many contingent factors at play. I am suggesting here the use of modern technology as a tool to constantly update the schedule as contingencies come into play, and to keep all CFOs—and possibly the CCs as well, because they will be the source of much of the intelligence—updated on the situation, not only in their own areas of responsibility but in neighbouring areas where ‘their’ PTAs are likely to be found.
In a snapshot such as the census, it is difficult to disentangle short-term mobility from longer-term migration. The feeling expressed by some at the CMU was that the count was down in remote areas, and that this most likely reflected migration since 2001 from remote settlements and homelands into regional centres and towns. There are, however, at least three other possibilities:
that the 2001 count was inflated by double-counting and the 2006 count is a more accurate reflection of the real population in remote areas
that the 2006 count in remote areas was an under-count in comparison with the 2001 count
that short-term circular mobility between homelands, hub settlements and regional centres has increased in the intervening period, so that at any particular time fewer people are at their place of usual residence than was the case in 2001.
To these we can add the following possibility: that all these variables are at play to different degrees in different regions of the Northern Territory. Given this situation, it seems to me very unlikely that census data can be used to find definitive answers to questions about mobility or migration. There are, however, certainly indicative patterns that are worth noting for further investigation at a micro-demographic level.
In some areas, particularly in the arid and semi-arid zones, there were many small isolated homelands where all—or nearly all—the dwellings were empty. There are at least two possible reasons. One is migration into larger settlements and towns, which could well be a result of the closure of small Community Development Employment Projects (CDEP) programs and which is likely to increase if the Commonwealth government implements the policies it has foreshadowed for what it considers to be ‘non-viable’ small homelands, such as ceasing to provide funding for housing. If migration is involved, one would expect a corresponding swelling in the ‘non-visitor’ population of towns and regional centres—subject to the proviso that considerable numbers of people might nevertheless call themselves ‘visitors’ because they are not in their own ‘country’. In some areas, however, it could simply reflect the seasonal occupation of homelands, with people moving regularly between their homelands and bigger hub settlements.
In particular instances (such as those described in Chapter 4), the emptiness of a settlement could reflect a very short-term movement, for example, to the site of a funeral ceremony or a festival. Such examples seem to occur commonly everywhere.
Another definite pattern emerged in the areas around service centres such as Tennant Creek or Mataranka. Here it was very common to find large numbers of PTA who were temporarily away in the service centre, shopping or accessing other services. This pattern was repeated on a smaller scale in all areas where a set of satellite communities looked to a particular hub community for basic amenities such as shopping, banking, health services and so on. In such cases, short-term mobility is a more likely explanation than migration, and in nearly all such cases the PTAs should have been put inside the form.
A pattern that was particularly discernable in the relatively densely populated Top End was intense levels of movement between adjacent settlements in culturally defined regions—such as, for example, the Yolngu-speaking region of Arnhem Land. In such regions, many PTA were said to be visiting relatives in nearby communities. Such patterns were also discernable in less densely populated regions, but more intermittently and over much larger areas (see Chapter 5). These patterns too are indicative of short-term mobility rather than migration.
In the next few years, government policy settings could well result in increased semi-permanent migration flows from remote settlements into towns. It will be a complex matter to disentangle the evidence for this migration from the ‘noise’ of mobility more generally, and given the complexity of the patterns described above it would be inadvisable to use census data as any kind of baseline measure. There is an urgent need for regional micro-demographic studies of the nature and causes of mobility and migration in remote Australia.
When the extent of the PTA problem became apparent, the CMU manager attempted to introduce some consistency into the decision-making about who should be moved into the forms, and instructed the checkers to document their decisions on the CMU checklists. She also attempted to get a fix on the scale of the problem, while acknowledging that the CMU did not have the systems to do a proper evaluation. Since the IHFs were not as yet in electronic form on a database, it was extremely difficult and time-consuming to crosscheck for the presence of individuals on more than one form. For example, some large communities had 10 or more boxes of forms, so that if PTA from another community were said to be there, all 10 boxes had to be checked. The DPC manager nevertheless instituted a check on PTAs where this was practicable. In some areas it appeared, from my own observations of this process, that a majority of PTAs had not in fact been counted at the places where they were said to be visiting.
As long as people were listed as PTA at their usual residence, these discrepancies would not have serious consequences for the final estimated resident population (ERP)—if PTAs were counted back into their communities for the purposes of the ERP. In such a case, however, the total Indigenous ERP of the Northern Territory would be significantly different from the total population counted in situ, since many of these PTAs did not appear inside any form unless they were moved back in at the CMU. Since PTAs are not counted back for the purposes of the ERP—and only those listed as visitors in other communities are counted back to their community—there will be a considerable under-count, because of the many PTAs unlikely to have been counted as visitors elsewhere.[4] There are also significant consequences for the quality of the data on all questions except the basic demographic variables of age and sex. There are no other data available for those originally listed as PTA on a form, and if they were not counted elsewhere as visitors then those data were not collected elsewhere either.
In a few cases, the manager of the CMU asked the CFO to go back and collect details of PTA who had been moved back inside the form, or to attempt to do so by phone, however, it was not possible to undertake this exercise across the board.
I took the opportunity at the CMU to scrutinise thoroughly all the forms from the homelands where I had undertaken my observation of the enumeration, and also those from nearby hub communities, in order to get some idea of the scale of the PTA problem in that area, and to assist the CMU in making judgments about which PTAs to put back into the forms. I also attempted to ascertain how many people had been missed altogether and how many had been double-counted, based on my personal knowledge of the local population. The former was a harder task than the latter: noticing an absence is a very different task from finding two instances of a presence.
In Chapter 4, I detailed particular instances of double-counting that I picked up in the course of this exercise, and I will not repeat those here. There were several other instances, particularly of the kind where a person was listed as PTA in one community or homeland, and as a resident at another. I found that one small homeland (population 30) had been overlooked (it was then subsequently visited). I also found one ‘vacant’ dwelling at one homeland, where everyone had been absent at a funeral at the time of the count. It had been intended for a follow-up visit that never happened (the CFO phoned the household from Darwin and the IHF was duly filled in).
For the study-area homelands as a whole—after eliminating all known instances of double-counting—I found that the IHFs listed a total of 598 residents at their own dwellings, 105 PTAs and 51 visitors. On the raw figures therefore, 105 out of 703 usual residents (15 per cent) of the usual population were PTAs, and 51 out of 649 (8 per cent) of the people present at the count were visitors.[5] It must also be remembered that these counts were already something of a fiction (see Chapter 4)—for example, the inhabitants of A3 and A4 had been counted ‘as if’ they were at home, but the forms were filled in at community A, where they were attending a funeral.
I then went through the process of trying to crosscheck whether all PTAs had in fact been counted elsewhere. In some cases, PTAs were said to be at another of the homelands within the group, so crosschecking was relatively straightforward. In other cases, I had to trawl through all the boxes of IHFs from the hub communities in the region. In still other cases, people were said to be further afield and it was not possible to check most of these because the relevant boxes had already been sealed, awaiting transport to Melbourne.
As a result of this exercise, 70 of the 105 PTAs were put back into the forms, either because they were definitely not counted at the places where they were said to have been, or, in the cases where crosschecking was not possible, because they were unlikely to have been counted—for example, because they were at a funeral. That is, in the case of this set of communities, two-thirds of those who were originally listed as PTA—or 10 per cent of the usual resident population—would not have been counted anywhere unless this exercise had been undertaken. For this 10 per cent of the population, only very basic information is available: their sex, age, Indigenous status and usual place of residence.