The lists of localities for the coding of workplace addresses were provided by the State Transit Authorities (STAs), and were much less comprehensive than the ABS’s own locality lists used in other questions. This was the case particularly for remote areas of the country—the majority of discrete Indigenous communities seemed to be missing.^[5] This threatened to force the coders into a lengthy and tortuous coding process. They were essentially instructed to bypass the process and code the State, followed by ‘Community further investigation’. It is hoped that these STA lists will be more comprehensive for 2011.

In family-coding mode, the coder had access to all the forms relating to a household—in instances where there was more than one form because there were more than 12 people in the household. This was not the case, however, when coding some other questions, such as the language question, and this caused problems if the answer was ‘Same as Person 1’. If this answer appeared on the second or a subsequent form for the household, the coder no longer had easy access to the answer given for Person 1, on the first form. There seems no good reason why the system should not allow for the first form to be readily accessible for all coding.

In Chapter 7, I noted that at the Darwin CMU the answer to Question 15 (‘Where does this person live most of the time?’) was checked carefully for each individual. If this question had been left blank, and it was obvious from the answer to Question 12 that the individual was not a visitor, the ‘This community’ box was marked. It transpired that this was a very important exercise, since according to the coding conventions at the DPC the answer to Question 15 overrode the answer to Question 12. If the ‘This community’ box was left unchecked at Question 15, the individual could not be included in the family coding for the household even if they had replied, at Question 12, that they were not a visitor. This was one of several examples that I noted at various points in the process from data collection to data processing where a very small detail could have very large consequences. Thanks to the Darwin CMU, this will not have been a problem for the Northern Territory, but I do not know what happened in other CMUs.

There were other glitches in the system that required ad hoc coding solutions, but it seemed to me that these were dealt with efficiently through the impromptu tutorial sessions and the more formal training sessions.