WEIGHTS AND ESTIMATION
NOT APPLICABLE CATEGORIES
WEIGHTS AND ESTIMATION
As the survey was conducted on a sample of households in Australia, it is important to take account of the method of sample selection when deriving estimates from the CURF. This is particularly important as a person's chance of selection in the survey varied depending on the state or territory in which they lived.
Each person record contains a weight which for the Pregnancy and Employment Transitions Survey is called FINPRSWT. The weight indicates the number of people in the civilian population represented by that person.
The Pregnancy and Employment Transitions Survey weight, FINPRSWT, appears on all 1,351 records. The estimates in the Pregnancy and Employment Transitions Survey publication are based on these records, that is women with a child under the age of two years during November 2011. Therefore when using FINPRSWT, to match the published Pregnancy and Employment Transitions Survey total estimate, the filter POP1=1 must be used.
Where estimates are derived from the CURF, it is essential that they are calculated by adding the weights of persons in each category and not just by counting the number in each category. If each person's 'weight' is ignored, then no account would be taken of a person's chance of selection or of different response rates across population groups, and the resulting estimates could be significantly biased and would only represent distributions within the actual selected sample and not the population of interest. The application of weights will ensure that the subsequent estimates conform to an independently estimated distribution of the population by age and sex, rather than to the age and sex distribution within the sample itself.
For further information see the Explanatory Notes in the publication Pregnancy and Employment Transitions, Australia, November 2011 (cat. no. 4913.0).
Standard errors for each estimate produced from this CURF can be calculated using the replicate weights provided on the file.
Each person record contains a set of 30 replicate weights. Replicate weights applicable to the Pregnancy and Employment Transitions Survey data items, contain the prefix 'WPX02'. By using these weights, it is possible to calculate standard errors for weighted estimates produced from the microdata. This method is known as the 30 group Jackknife variance estimator. For data items that are only applicable to the Pregnancy and Employment Transitions Survey, refer to Data Item List.
Under the Jackknife method of replicate weighting, weights were derived as follows:
- 30 replicate groups were formed with each group formed to mirror the overall sample (where units from a collection district all belong to the same replicate group and a unit can belong to only one replicate group)
- one replicate group was dropped from the file and then the remaining records were weighted in the same manner as for the full sample
- records in that group that were dropped received a weight of zero.
This process was repeated for each replicate group (i.e. a total of 30 times). Ultimately each record had 30 replicate weights attached to it with one of these being the zero weight.
Replicate weights enable variances of estimates to be calculated relatively simply. They also enable unit records analyses such as chi-square and logistic regression to be conducted which take into account the sample design. Replicate weights for any variable of interest can be calculated from the 30 replicate groups, giving 30 replicate estimates. The distribution of this set of replicate estimates, in conjunction with the full sample estimate (based on the general weight) is then used to approximate the variance of the full sample.
To obtain the standard error of a weighted estimate y, the same estimate is calculated using each of the 30 replicate weights. The variability between these replicate estimates (denoting y(g) for group number g) is used to measure the standard error of the original weighted estimate y using the formula:
g = the replicate groups
y(g) = the weighted estimate, having applied the weights for replicate group 'g'
y = the weighted estimate from the full sample.
The 30 group Jackknife method can be applied not just to estimates of population total, but also where the estimate y is a function of estimates of population total, such as a proportion, difference or ratio. For more information on the 30 group Jackknife method of SE estimation, see Research Paper: Weighting and Standard Error Estimation for ABS Household Surveys (Methodology Advisory Committee), July 1999
(cat. no. 1352.0.55.029).
Use of the 30 group Jackknife method for complex estimates, such as regression parameters from a statistical model, is not straightforward and may not be appropriate. The method as described does not apply to investigations where survey weights are not used, such as in unweighted statistical modelling.
The following table has been provided to enable CURF users to check some of the relative standard errors they have produced.
NOT APPLICABLE CATEGORIES
Women with child under 2 years—By selected characteristics
Relative Standard Error
|Age of women (years)|
|40 and over |
|Social marital status|
|Country of birth and period of arrival|
|Born in Australia|
|Arrived before 2001|
|Arrived 2001 to the date of interview|
Many data items included in the microdata include a 'Not applicable' category. The classification value of the 'Not applicable' category, where relevant, is shown in the relevant data item lists available on the Downloads