STATISTICS NEW ZEALAND
POLICY AND PROCEDURES FOR USING THE CLASSIFICATIONS AND RELATED STANDARDS SYSTEM (CARS)
Summary
1 This document sets out the principles, policies and procedures for the use of the Classifications and Related Standards system (CARS).
The Classifications and Related Standards system has been developed as the Statistics New Zealand (SNZ) central repository for all classifications, concordances and codefiles - whether they are standard or survey specific (i.e. non-standard or independent).
CARS will contribute to the critical success factors stated in the corporate mission by:-
·improving efficiency in survey operations
·increasing reliability in survey outputs
·using innovative database technology to meet client needs, and
·contributing to the long run integrity of the statistical system.
CARS will play an important role in helping SNZ fulfil one of its stated responsibilities in the Statistics Act 1975 (Section 14) - to define, lay down and promote standard concepts, procedures, definitions, and classifications for use in official statistics.
The document gives an overview of CARS and its place within the Statistics New Zealand infrastructure. This is not an implementation document nor is it a user manual.
The effective date of this document and the stated policy is November 1996 - the date when the following Policy And Procedures were approved by the Statistics New Zealand Executive Management Committee.
Objectives
2 The purpose of CARS is to:
a Use database technology to provide centralised classification storage, maintenance and access facilities for classification data used both in the development and processing of surveys, and in the subsequent analysis and evaluation of the data.
b Help reduce the time and resource taken to develop surveys and contribute to improved data quality by use of computerised data documentation (metadata) to support the application of standard data item definitions and classifications.
c Facilitate the comparison and analysis of data by storing concordances
Scope
3. CARS is SNZs official central classifications repository which will contain all the current economic, social and geographic classifications, concordances and codefiles used by data collections within SNZ. It will be the repository for:
i historical classifications, codefiles and concordances, and information relating to them, which have previously been used within SNZ.
ii standard classifications
iii survey specific classifications (either non-standard or Independent), concordances and codefiles.
iv non-standard classifications. Although the Department is moving to progressively phase out the usage of non-standard classifications in current surveys, non-standard classifications will remain in CARS as these classifications will always be required for analyses of historical data (including time series).
v all classification categories used for coding survey data at the input stage, and their descriptors and labels to be used for presenting output data in tabular form.
Responsibility
4 The CARS policy supports the Statistics New Zealand Mission. It is the responsibility of all staff to comply with classifications policy and procedures.
5 With the advent of CARS, the role of the Standard Classifications Committee (SCC) has not changed. It is responsible for the definition, implementation and promotion of standard statistical concepts, procedures and classifications. C&S Section is responsible for the management of CARS. See section 4.2 of the procedures document for further information.
Statistics New Zealand Policy
6 It is Statistics New Zealand policy that CARS is the official central repository for all economic, social and geographic classifications, concordances and codefiles used in SNZ. It will be available for use by all departmental data collections.
7 It is Statistics New Zealand policy that existing surveys and data processing systems, as they are redeveloped, use CARS for all their classifications requirements.
8 It is Statistics New Zealand policy that all new surveys and data processing systems plan to use CARS and standard classifications unless permitted by the Standard Classifications Committee to do otherwise.
9 It is Statistics New Zealand policy that systems will source all their classifications, concordances and codefiles from CARS. In most cases, systems will be using standard classifications and concordances, which will be in CARS prior to the date the systems need to use them. (If necessary, Classification and Standards Section, given prior warning, will get them into CARS by the required date.) Some systems will also need to get their own independent or non-standard survey specific classifications and concordances into CARS before using them.
10 It is Statistics New Zealand policy that most systems will not need to keep their own internal copies of classification data, and they will not need their own separately-developed maintenance and access systems for classification data. All the required classifications, concordances and codefiles will be in CARS, and CARS has standard procedures for accessing and maintaining these.
11 It is Statistics New Zealand policy that when a system does needs its own copies of classification data (e.g. proprietary software products which use classifications - such as SUPERCROSS) these copies should still be sourced from CARS - i.e. CARS should have all the classification data being used in SNZ.
12 It is Statistics New Zealand policy that CARS will contain all the valid labels for output data and include access routines for using those labels. Therefore, all outputs for publication, other corporate products, and customer defined outputs should have their classifications derived from CARS to ensure a consistent publication standard.
Procedures for Working with CARS
Contents
1.0 Procedures For Working With Classifications In CARS
1.1 CARS Browsers, Users and User Groups
1.2 Roles and Responsibilities
1.2.1 Standard Classifications Committee
1.2.2 Classifications and Standards Section
1.2.3 Role of CARS Implementation Team
1.3 CARS Help Desk
1.4 Training
4.4.1 Who Will Need To Be Trained To Use CARS
4.4.2 Types Of Training Courses That Will Be Available
4.4.3 Who Will Deliver The Training
4.4.4 How The Training Across SNZ Has Been Planned
1.5 Storing And Maintaining Classification Data In CARS For Commercial Clients
2.0 Benefits
3.0 Glossary
1.0 Procedures For Working With Classifications In CARS
1.1 CARS Browsers, Users and User Groups
All SNZ personnel can access CARS and an access structure has been set up according to the level of use desired.
A CARS Browser is any SNZ staff member who is able to log onto the CARS on-line system and/or is able to access the CARS database directly, and who is not a member of any CARS User Group.
The staff member additionally becomes a CARS User when he/she becomes a member of one of the following CARS User Groups:-
Owner User Group
The classifications and associated metadata in CARS are owned by owner user groups, which are an individual or a group of individuals such as a section, a grouping of sections or other organisational unit within SNZ. An owner user group is responsible for maintaining the classifications, concordances and codefiles it owns.
Reader User Group
Virtually all released data in CARS is able to be read by anyone with access to CARS regardless of whether or not they belong to any user group. However, there may be a small number of confidential classifications in CARS which have a reader user group assigned. A classification which has a reader user group assigned can only be seen by members of that group and the owner user group.
Supervisor User Group
Its role is to manage all the classification data being put into CARS and to facilitate the use of the system.
Survey system applications will be able to directly access CARS through standardised access routines.
1.2 Roles and Responsibilities
1.2.1 Standard Classifications Committee
With the advent of CARS, the role of the Standard Classifications Committee (SCC) has not changed. It is responsible for the definition, implementation and promotion of standard statistical concepts, procedures and classifications.
The Standard Classifications Committee will approve strategies for the review and development of classifications ensuring:
- consistency with SNZ social and economic statistics strategies
- appropriate user consultation (both internal and external)
- the selection of appropriate classification criteria
- that the resulting classification meets user needs
- comparability with international, Australian and earlier New Zealand standards
- that suitable questions can be formulated for the application of the classification
- integrated statistics can be collected and published
- a suitable implementation strategy is agreed to.
The SCC will:
- approve the adoption of standards and standard classifications. (So, this will occur prior to the loading and release of the standard classifications in CARS)
- approve and monitor the implementation strategies for new or revised standards and standard classifications
- monitor the departure from approved standards or the use of non-standard classifications and standards. (Within CARS there are administration reports to help Classifications and Standards Section provide the necessary information to SCC on departures from approved standard classifications)
- monitor the promotion of standards and standard classifications by the Classifications and Standards Section including the provision of training, advice and assistance within SNZ. This training is in the use of classifications - not to be confused with training in the use of CARS (which is discussed later on in this document).
1.2.2 Classifications and Standards Section
The role of the Classifications and Standards Section has changed with CARS coming into production.
The Current Role Of The Classifications and Standards Section (C&S)
The Classifications and Standards Section has a key role in developing and maintaining all statistical standards, standard classifications, concordances, codefiles and other related material. Maintenance of these will now be done in CARS.
An Additional New Role For The Classifications and Standards Section
The C&S Section is responsible for the management of CARS.
Within the C&S Section there will be some staff assigned to membership of the CARS Supervisor User Group which is responsible for:
- defining all classifications to CARS. This includes defining the approval status for any classification, the classification title, and the classification acronym
- defining and maintaining subjects and topics in the CARS Navigator
- deleting obsolete classifications and concordances
- maintaining users and user groups through a consultative process with users
- generating codefiles
- co-ordinating the update of codefiles.
One member of the Supervisor User Group will have the additional responsibility of CARS Administrator. The CARS Administrator ensures that the CARS Help Desk works successfully in SNZ and ensures that Help Desk Queries are resolved in a timely way (please refer section below on the CARS Help Desk). The CARS Administrator will also co-ordinate most of the implementation work for CARS by fielding it to the appropriate people.
Within the C&S Section there will be several Owner User Groups responsible for maintaining all the standard classifications.
1.2.3 Role of CARS Implementation Team
A CARS implementation team will be responsible for the smooth introduction of CARS into SNZ.
During the first year that CARS is in production, the team will comprise:
- the Manager Classification and Standards Section
- a senior systems analyst from Survey Applications Development
- the CARS Administrator
- a senior analyst from Survey Applications Development
- an analyst/programmer from Survey Applications Development
The implementation team will specifically be responsible for:
- planning and carrying out the CARS training programme across SNZ (please refer section below on Training)
- promoting the use of CARS in SNZ
- assisting systems to interface successfully to CARS
- deciding on new reports for CARS
The training and promotional responsibility will include reviewing the forward work plan for the Department and contacting potential new users of CARS with a view to providing assistance to systems needing to interface to CARS.
The AD members of the implementation team will be able to write any special interfaces required, but those systems needing special interfaces should include the cost of developing those interfaces in their budget. Because CARS has standardised virtually all of the ways for dealing with classifications and concordances in SNZ, the cost of any additional special interfaces to CARS should be minimal compared with what, prior to CARS, each system would have spent in developing their own mini classification and concordance system.
At the end of the first year for the CARS production system, the roles and composition of the CARS implementation team will be reviewed.
1.3 CARS Help Desk
Within the SNZ IT Help Desk there is a CARS Service Group (called CARS-All). This means that anyone in SNZ can submit a Help request about CARS via the main SNZ Help Desk. It will assist in the efficiency of resolving CARS Help calls if the requester notes that the call is to do with the CARS system.
Within the SNZ IT Help Desk the CARS Administrator (refer section 4.2.2) will function as Service Group Leader for the CARS-All Service Group.
When a Help Request is submitted to the SNZ Help Desk for a CARS-related problem, the TSS person staffing the Help Desk will assign that call to the Service Group Leader for the CARS-All Service Group. The Service Group Leader will then assign the request to the appropriate Service Agent.
Service Agents for the CARS-All Service Group will include:
- members of the CARS Supervisor Group
- the Senior Systems Analyst for the AD staff providing CARS support
- the CARS senior analyst, and
- the CARS analyst/programmer.
If Help requests are clearly requiring AD support (e.g. a bug in the CARS system), then the usual procedure will be for the Service Group Leader to create an AD Service Request and within that to make a doclink to the SNZ IT Help Desk request (so the detailed request is only held in one place). The AD Service Request will be forwarded to one of the AD staff nominated as a member of the CARS support team (i.e. a Service Agent for CARS). It will then be the CARS Administrator (or Service Group Leader) or another member of the CARS support team who eventually flags the SNZ IT Help Desk request as closed.
The initial Service Categories for CARS calls are as follows:-
CARS - Access
CARS - Bugs
CARS - Database
CARS - General Queries (AD)
CARS - General Queries (C&S)
CARS - Interfacing Applications
CARS - Loading
CARS - Minor Enhancements
CARS - Training
CARS - WinSCM
1.4 Training
1.4.1 Who Will Need To Be Trained To Use CARS
Nearly everyone in SNZ working with data will need to be trained to use CARS.
1.4.2 Types Of Training Courses That Will Be Available
There will be three types of training courses provided:-
A CARS browsers course
This course is designed for those who will want to view classification data in CARS but who are not responsible for getting any data into or out of CARS.
A CARS classification owners course
This course is designed for those who will want to both view classification data in CARS and get classification data into CARS. Typically attendees for this course will have non-standard or independent classifications and concordances that they need to get into CARS. They will also learn to maintain classifications and concordances in CARS.
A CARS programmers course
This course is designed for those who will want to view classification data in CARS, get classification data into CARS, and write programs to access data in CARS. Typically attendees for this course will be programmers and para-programmers using SAS, EXCEL, ACCESS or SQL WINDOWS.
1.4.3 Who Will Deliver The Training
The training will be delivered by members of the CARS implementation team. Each course will be run by two trainers - one a Classifications and Standards Section representative, and the other a Survey Applications Development representative from the CARS implementation team.
1.4.4 How The Training Across SNZ Has Been Planned
It will take some time to train everyone in the Department who will eventually need to use CARS. However, not everyone will need to use CARS straight away.
People will need to be trained to use CARS when they want to write programs to access CARS data or when they are working with systems that access CARS data.
For the first year of the CARS production system (i.e. the year to 30 June 1997) it is planned to target programmers and para-programmers in SNZ for CARS training. There are approximately 100 such staff, and it is estimated that at least ten CARS Programmers Courses will need to be run to cater for them.
A list of those programmers and para-programmers needing training has been developed, together with approximate dates for when the training would be most appropriate.
The concept for training is that it is not just running a course. In order to ensure that the training is successful, the trainers, prior to each course, will investigate what the trainees need to use CARS for. This will usually be established by actually visiting the trainees in their workplace, seeing what their classification work involves and noting what types of PC and software they are running. Then, after the course, the trainers will again endeavour to visit the trainees in their workplace to ensure that they have already made (or know how to make) the shift to using CARS instead of other ways they had of working with classification data.
During the second year, the remaining needs for bulk training will be met by the CARS implementation team, and a plan for the ongoing training of new staff members will be developed. It is envisaged that there will be a timetable for regularly running each of the three CARS training courses.
1.5 Storing And Maintaining Classification Data In CARS For Commercial Clients
The CARS system can easily and efficiently store classifications and concordances for commercial clients. If a classification or concordance is required to be kept confidential, then the CARS security system can cater for this by way of the reader group concept mentioned earlier in section 4.1.
Members of a reader group for a particular confidential classification or concordance can access that classification data in exactly the same way as they could access any other classification data in CARS. So, if that confidential classification data needs to be used with other data in SNZ to produce some specific output for a client, then someone from the reader group for that classification data could write a program to access that confidential data in CARS.
A charging policy for storing and maintaining classification data in CARS for commercial clients will be developed once an assessment has been made of the potential market. It will take account of the capital costs of developing the system in CARS for the protection of confidential information. This policy will be based on the general principle that when confidential classification data stored in CARS is accessed to produce output from SNZ, then the charge of producing that output should include a component for the storage and maintenance of the classification data in CARS.
It is also envisaged that some specific commercial products will be developed from CARS. Likely products include classification query tools based on CARS and WinSCM (WinSCM being the ABS coding tool that interfaced to CARS). These products are likely to be available on CD ROM, diskette or via the Internet.
2.0 Benefits
CARS is SNZs official central repository for all economic, social and geographic classifications, concordances and codefiles, for use by all departmental data collections. The benefits arising include:
- Common storage facilities will make it easier to find and access classifications, concordances and codefiles, and this will benefit the survey development process.
- Common procedures for browsing, updating and accessing classifications and concordances in CARS will help both reduce the time and resource required to operate surveys, and eliminate the need to provide these facilities separately in departmental survey systems.
- Having all the standard classifications in one place and in one format facilitates the use of standard classifications in surveys. The lack of a centralised classifications system can result in surveys utilising their own variations of standard classifications, i.e. non-standard classifications. Increased use of standard classifications will result in closer integration of data from various surveys. This will ultimately result in the department being better able to satisfy customer requirements.
- Classification and Standards Section will be better able to expand their service to SNZ in terms of classification advice and timely promulgation of new and revised standard classifications, concordances and codefiles.
- Many SNZ divisions report a need for centrally held and universally accessible concordances. Time-consuming work in producing standard and non-standard outputs sometimes involves either finding whether a concordance exists or manually constructing a concordance because one cant be found (even though it may exist somewhere else in SNZ). In addition to concording different classifications (e.g. NZSIC and ANZSIC), concordances in CARS will precisely document the changes between versions of one classification. This will facilitate the provision of time series analyses.
- The provision of a centralised computer-assisted-coding facility within CARS will considerably reduce the time required to develop survey processing systems, and will improve coding accuracy and consistency. In time, the same computer-assisted-coding facility will be available in a packaged form to external users. Another benefit of using the computer-assisted-coding facility within CARS is the coding consistency gained when several surveys coding to the same version of a classification all use the same source codefiles.
- CARS will also result in considerable saving in staff training and familiarisation, especially in areas of SNZ with relatively high staff turnover, because survey processing, research and programming staff will only need to learn one way of handling classifications data rather than many different ways.
- The storage in CARS of standard labels for output will result in increased standardisation and improved presentation of tables in the Departments publications.
- Because CARS verifies the consistency of classifications data before it is released for general access by all CARS users, there will be a reduction in the the risk of error from inadvertently using corrupt classifications.
- SNZs ability to market classification maintenance facilities will be enhanced, especially as particular classifications can be kept confidential to the owner or to specified readers in CARS.
3.0 Glossary
Classification:
A classification is an ordered set of related categories used to group data according to their similarities.
Concordance:
A concordance specifies the translation of categories from a source classification version to a target classification version using a specific compilation method or criterion.A concordance may be between different versions of the same classification and between versions of different classifications.
Codefile:
A codefile is a comprehensive listing of probable survey responses and the categories to which they are coded. It may include classification category descriptors and synonyms for these.
Standard Classification;
A standard classification is a classification which is approved by the Standard Classifications Committee (SCC) or by an officer acting under SCCs delegated authority.It can be a: -
New Zealand Standard, or
Interim Standard
Examples of these are:
New Zealand Standard Classification of Occupations 1995 (NZSCO95)
New Zealand Standard Classification of Countries 1995 (NZSCC95)
Australian and New Zealand Standard Industrial Classification - NZ Version 1996 (ANZSIC96)
Meshblock 1996 (MB96)
Territorial Local Authority 1995 (TLA95)
Non-Standard Classification:
A non-standard classification is a variation of a standard classification, where a variation is category changes:-
not approved by the Standard Classifications Committee, or
approved by the Standard Classifications Committee for limited use.
Any classification which has codes and/or descriptors different from those in the corresponding standard classification is non-standard.
Independent Classification:
An independent classification is a classification which is not a standard or a non-standard classification. An example is Job Search Methods in the Household Labour Force Survey.
Statistical Standard:
A Related Standard (also known as a standard or statistical standard) is a set of components which, when used together, produce consistent information for a particular output variable across collections and over time.
A standard defines key definitions and concepts, questionnaire modules, codefiles, coding rules, and input and output categories.
Standards are adopted by general consent after a formal development process.
Label:
A label is an abbreviated category descriptor. A standard label has a length not exceeding forty characters and is used particularly in standard classifications to provide standardised wording for survey output.