The CFO encouraged the Junjuwa CIs to start the count a week early—a suggestion that made a lot of sense given their enthusiasm and my availability as an observer and as a logistical resource with a car and a licence, and the fact that the week before census week was ‘slack week’.[7] Since people’s finances are getting low, there is generally much less drinking or mobility during slack week; people are easier to find and are less likely to be distracted by all of the activity that occurs during pay week; and there are less likely to be so many visitors in town for shopping. Of course, counting this community early raises questions about the potential for double-counting of people counted in Junjuwa and then counted again somewhere else in census week.
There were four CIs working in the community of Junjuwa and each was allocated 15 houses. In addition, there were a number of much smaller communities that this group of collectors was responsible for: Darlngunaya, Bungardi, Burawa, Biridu, Galamunda and Warangarri.
By and large, the CC took responsibility for these other communities, in part because I had a car and therefore she and I could go off and count those people. One young man was trained up to do the count in Darlngunaya, where he resided. This is the largest of the smaller communities, containing about 10 houses.
In the event, the real count began on 1 August. The CC and I drove around Junjuwa, finding the CIs, giving them their forms and the map on which the CC had allocated them particular areas. We then drove over to Burawa, a small community of about six houses, to begin the interview process. This was the first time the CC had attempted to fill in a form, since that process did not occur during the training. This first interview took 45 minutes, in part because the CC was trying to recall points made in the training video—two weeks earlier—about individual questions on the form.
In total, there were five CIs, one CC and myself engaged in the census for Bunuba Inc. Apart from the young man counting Darlngunaya, the four remaining CIs were in fact two couples, who worked together. This seemed a very effective arrangement for a range of reasons. One partner could act as a backup interviewer if there was a person in the house with whom the other partner had an avoidance relationship. One person might be far more confident in writing, so they could write the answers while the other asked the questions. Finally, it is more enjoyable and socially acceptable not to work alone.
One issue that arose was with people who worked full time. There were a number of houses that I visited with the CC where no one was home during the day. The CC herself was a mother of five children and, not owning a motor car, she was unable to chase up these people in the evening. The current approach seems to be based on the assumption that Indigenous people will be at home and available for interview during working hours, and does not seem to make provision for people in such communities who work full-time—apart from the CC or CI following them up after hours, which is not always possible if the census staff do not live in the same community. Such situations must surely arise regularly in Indigenous communities. In the end, the CC made the decision to leave a form at these households, to be collected later.
The problem with observational reporting is that it is not necessarily so obvious to the observer what is working well, as opposed to what is not working well, in the current system. An opportunity did arise, however, to observe a context in which the census process worked less well. Mindi Rardi is a largely non-Bunuba community in Fitzroy Crossing, which was MWW’s responsibility in this context. The CC, however, offered to help out the MWW staff, so the four CIs from Junjuwa and I spent two mornings enumerating the people there.
In general, the CIs were less comfortable working in this other community, even though they knew many people there. They were much keener to get forms over and done with quickly, and would not push people for answers. For example, the count of children in Mindi Rardi will be very much less than it should be, for the CIs would write down only the immediate response to the question ‘Who stops here?’, which often overlooked children. Others also would have been missed, since there was no follow-up or return visit to record people who happened to be at card games or at the supermarket on the day. While some of these people might have been noted when the forms were checked at MWW, those who were visitors would not have been recorded at all.
One of the CCs stated at the training that she hoped to have the parts of Fitzroy Crossing for which MWW was responsible completed by 3 August. After that date, she was going on leave for four weeks; however, the count for MWW’s Fitzroy Crossing communities did not begin until well into the official census week, after 7 August. Part of the reason for the late start was that MWW still did not have enough forms (they were about 120 short) well into census week.
In any case, a single MWW staff member did most of the count of people in town at Kurnangki and Loanbung, as well as other communities such as Djimbalakudindj, about 100 kilometres to the west. By this time, this particular MWW staffer was acting CEO, so was having to carry out this work after hours.
When people are really visitors—that is, when the are not just extended family members from outstations—for example, if they are in-laws from a different language group, the interviewers sometimes did not feel comfortable asking them the full range of questions. On one occasion, I was asked by the two CIs to enumerate a house of eight young people from as far south as Jigalong who were still in town after a funeral, and who were camping in one of the older houses at Mindi Rardi. In a sense, it was handy for the CIs to have someone like me around, a kind of neutral figure, to ask and explain these ‘rude’ questions to strangers in town.
According to the CFO Field Manual, there are three types of enumeration: standard (the approach taken in the Kimberley), rolling and assisted. Arguably, while the first kind of approach worked sufficiently well in certain areas, the problem was that because there was absolutely no oversight in many communities no one was aware if nothing had been done until it was almost too late. Taking the third approach[8] would have meant that the progress on the count could at least have been monitored. Taking the first approach and applying it across the board does not acknowledge the fact that there can be significant variations in human capacity across communities—a truism that might not be apparent to a CFO, especially one with a relatively short-term relationship with communities and community members. The approach, then, has potential to be very ‘hit and miss’.
The intent with the census was that the majority of the southern Kimberley would be completed in some sense by the end of the week of 7 August. What ensued was that the CFO was still training in that week and that, for a significant number of communities, absolutely nothing happened once the training ended. Some very large communities were not enumerated until well into September when the CFO was joined by an ABS colleague from Perth and by two other CFOs who had finished their areas. The communities of Wankgatjungka, Yiyili, Looma and Jarlmadangah all required unplanned-for help from outside. My understanding was that the absolute deadline for forms to be ‘in the mail’ was 15 September, the date I was to leave Fitzroy Crossing. The count was, however, still under way when my observations ceased.