ABS Labour Force Survey not biased
The Treasurer is right in saying that the NT employment and unemployment numbers released by the ABS are volatile. The ABS has continually cautioned that this is the case when looking at increases or decreases in the numbers of employed and unemployed from month to month.
Trend estimates for NT employment and unemployment should be looked at over a number of months to get a good indicator of underlying changes in these numbers. Used in this way, they will provide a good basis for the informed decision making you seek in your editorial (17/01/05).
Minister Stirling has also stated that the sample for ABS' NT Labour Force Survey is not representative of the total population because it includes a disproportionately high number of unemployed households. This is not correct. The survey sample is representative.
Due to the relatively small sample size in the Territory, households with unusual characteristics, which are occasionally included in the sample rotation, can have a greater impact on the estimates than would be the case if a much larger sample were used. Increasing the sample to this much larger size would impose an unacceptably high survey response burden on the NT population. However, the survey's current sample design ensures that, on average, it is not biased towards any particular section of the population.
Robyn Elliott
Regional Director
Australian Bureau of Statistics, NT
ABS letter to the NT News editor 'ABS labour force figures not biased' on 20 January 2005