LINKING BETWEEN-FILES

The second stage in the methodology links two child level files. This occurs when linking a state and territory file to the CCMS file. This process was performed on a state-by-state basis, i.e. only the NSW children in the CCMS file will be compared with the NSW state file. To determine if two child records on different files were the same child, the following logic was applied to between-file linking.
 
Multiple child records (between files) identified as being the same child based on their Statistical Linkage Key (SLK):
Link on SLK and on SA1
If a child record from each file has the same SLK and both are located in the same SA1, then both records are considered as belonging to the same child. If they don’t link, they are treated as separate episodes.

Link on SLK and on CD
Figure 3.10 shows that if a child’s SA1 code is not supplied, the linking completed by comparing SLK and child’s CD. If a child record from each file has the same SLK and the same CD, then the records are considered as belonging to the same child. If they don’t link, they belong to different children.

Link on SLK only
If a child’s SA1 or CD is not supplied by the data provider or is of a poor quality, then linking is conducted using SLK alone. In this scenario, child records from two different files who share the same SLK are considered as belonging to the same child. If the SLKs do not link, they are considered to belong to different children.

Between-file episode linking was completed for each individual state and territory and conducted between a state and territory file and the CCMS.

Duplicate episode records for a unique child at a service provider:
A child who is enrolled within a service provider can be reported within both a state and territory collection and the CCMS. In some scenarios, each data source may be reporting about different preschool experiences within the same service provider. The different data sources could also be reporting the same preschool experience. Due to the nature of state and territory collections and the CCMS, there is no way to definitively know what these linked episode records at a service provider represent. Operationally, for the 2012 National ECEC Collection, linked episode records at a service provider were considered duplicates. To determine if two or more episodes linked to a child were unique or duplicates of each other, the following business rules were applied. This process applied for each state and territory after both the ‘Within File’ and ‘Between File’ linking processes had been completed.

Duplicate episode record checking within a service provider:
The ABS as part of the Preschool Provider Frame (PPF) (for more information on the PPF refer to Chapter 1: Collection Scope and Coverage) allocates a unique identifier to each unique service. The ABS provider ID is used to identify service providers within and across data files, and across collection years.

Figure 3.11 below, outlines how a child record with two or more episodes that share the same ABS provider ID code were considered as being a duplicate of each other. In this scenario, the state/territory supplied episode is selected over the CCMS source episode and the duplicate episode is removed from the child record. If a child enrolled in one service provider has multiple episodes originating from the same state and territory file, these episode duplicates are not removed from the child record. For a child record with corresponding episodes that have different ABS provider ID codes, each episode is considered to be unique and remain a component of the child record.

Previous Page

Next Page