Document Start
Paper Title: Disclosure Review Board - Terms of Reference
Purpose of the paper
This paper describes proposed terms of reference for the Disclosure Review Board.
Background/Context/Known Facts
At the 2011 May Management Meeting, confidentiality was identified as an enterprise risk.
As part of a strategy to mitigate disclosure risks for aggregate statistics, SMG agreed to the
establishment of a Disclosure Review Board. The Disclosure Review Board replaced the
Microdata Review Panel with the extended role to include aggregated statistics.
Consultation undertaken in preparing this paper
The proposed membership and modus operandi have been discussed with members of the
Disclosure Review Board.
1
01
Terms of Reference of the Disclosure Review Board
Background
1.
At the Management Meeting in May 2011, confidentiality was identified as an
enterprise risk. SMG discussed a strategy for improving confidentiality capability for
aggregate statistics in July 2011. Methodology was asked to develop a proposal to establish
a Disclosure Review Board (DRB), which extended the role of the Microdata Review Panel
(MDRP). The DRB replaced the Microdata Review Panel in February 2012.
2.
The DRB retained the MDRP's role of providing advice on the disclosure risks
associated with microdata dissemination. The consideration by DRB on the risks associated
with releasing aggregate statistics is predominantly on an exceptional basis, i.e. individual
delegates responsible for approving the publication may seek DRB's advice where there are
issues not adequately covered by existing policies. As well, the DRB may initiate
investigations where the board has identified specific systemic or complex issues of concern.
3.
In April 2015 the General Manager TableBuilder approval criteria was implemented.
TableBuilder releases which met the criteria are now approved by the respective General
Manager and are no longer required to be considered by DRB.
4.
As a result of organisational restructure, DRB membership was reviewed in July
2015.
5.
This paper describes the role, membership and procedures for the DRB.
Role
6.
The DRB chair is responsible for advising the Australian Statistician and delegates on
the risks of an individual person or organisation being identified in statistical information
disseminated by the ABS, and the approach to mitigating the risks, so that ABS products are
released in a manner that complies with legislation and policies relating to confidentiality and
disclosure. It is the role of the DRB members to provide advice to the chair to assist in final
decision making.
Membership
7.
The Board will comprise the following members:
General Manager, Methodology Division (chair)
Program Manager, Communications and Dissemination Branch
Program Manager, Macroeconomic Division
Program Manager, Methodology Transformation Branch
Program Manager, Industry Division
Program Manager, Population and Social Statistics Division
Director, Strategic Projects and Partnerships Division
Director, Policy, Legislation and Assurance Section.
8.
The Director of the Data Integration, Access and Confidentiality Methodology Unit
(DIACMU) will be an advisor to the DRB.
9.
Secretariat support will be provided by the Customised and Microdata Delivery
(CAMD) section.
2
02
Modus Operandi
10.
Scope for Microdata: ABS Policy requires all proposed microdata releases to be
considered by the DRB. The DRB may initiate investigations into particular outputs where
particular interests or concerns arise.
11.
Scope for Aggregate Data: For aggregate statistics, the delegate and related staff
releasing statistical output may approach the DRB for advice. This normally occurs when
they consider that the release involves complex issues not adequately covered by
established policies. The line management for each business area will continue to be
responsible for ensuring that outputs from that area are released appropriately, and should
provide guidance to their areas about which outputs should be considered for discussion by
the DRB. The DRB may initiate investigations into particular outputs where particular
interests or concerns arise.
12.
A subset of TableBuilder products (those which do not meet the General Manager
approval criteria), must be considered by the DRB. DRB provides advice to the Australian
Statistician and delegates for the respective output of the TableBuilder product.
13.
Delegations: For aggregate statistics, the Statistician has delegated the power to
publish results or abstracts from compilations or analyses (including TableBuilder) to
respective officers. The Statistician is himself responsible for approving the release of
microdata. DRB advises the Australian Statistician and delegates for the respective output.
All delegates have the responsibility of
ensuring that the confidentiality requirements of the Census and Statistics Act are met in all
releases of data.
14.
Submission: Guidelines for submissions to the DRB will be developed and
announced jointly by the DRB Secretariat and Data Integration, Access and Confidentiality
Methodology Unit (DIACMU). The DRB Secretariat will provide information and advice about
the role of the DRB, and will also provide administrative support to business areas who are
lodging submissions to the DRB. Coordination of assessment work required for obtaining
advice from DRB will be undertaken by the DRB Secretariat for microdata and TableBuilder,
and by DIACMU for all other aggregate statistics. Proposing areas will be required to work
with the DRB Secretariat for microdata and TableBuilder and DIACMU for aggregate data to
undertake the assessment work.
15.
Submission of proposals for review of release of aggregate statistics should be made
only at the Director level with SES endorsement. Proposals must go via the DRB Secretariat
to the DRB. The DRB steers the assessment of disclosure risks associated with the
proposed output and the manner in which they will be released, and advises the delegates
on acceptability of the risks and appropriate measures to mitigate the risks.
16.
Continuous Improvement: The DRB will regularly review the procedures for
assessing disclosure risks and tools for risk mitigation, and will commission work to develop
or improve the procedures. DRB will support Microdata Futures and Policy Legislation and
Assurance with their work on the future directions of microdata. Where appropriate DRB will
actively seek to influence ABS policy regarding micro and aggregate data. DRB will
champion the extension of ABS staff's skills and knowledge of confidentiality policies and
practice, through appropriate training and provision of self help materials.
17.
Frequency of meetings: The DRB will meet monthly, although some issues may be
discussed out of session.
3
03
4
04
5
05
Document Start
SUBMISSION PROPOSAL FOR CONFIDENTIALISED UNIT RECORD
FILES
IMPORTANT INFORMATION
Overview diagram of the CURF process:
Read the following documents before completing this submission proposal:
SECTION A - SUMMARY AND BACKGROUND
If you have any questions about this proposal contact Microdata Access Strategies (MAS)
Survey title
(including reference year )
Proposed CURF Title
Branch
Section
Contact Person Name /Phone
Survey area Director
Survey area FAS
A1. CURF Format : (tick all that apply)
Discuss with MAS if unsure
Basic CURF
Expanded CURF
Clause 7A release (release of a microdata file to an individual organisation) - link to policy and
legislation section 10-12
Other, specify:
A2. What are the other methods of release planned for data from this survey ? (tick all that
apply)
:
TableBuilder
ABS.Stat
Information
Other, specify in the box below
consultancies
06
A3. Is this a repeat CURF ? (Repeat CURFs are CURFs released from the same collection series
for a new reference period.)
No
Yes
A4. Is this a new edition / reissue (tick all that apply)
No (this is a new CURF release, the data has not been released previously )
Reissue due to data errors on previously approved file
New edition including previously unreleased data
The date the original file was released on the web: dd/mm/yyyy
A5. Administrative data used in CURF product .
If administrative data is being used on your CURF product, have you sought permission from the
data custodians?
No
Yes
N/A - no administrative data on the CURF
CURF BACKGROUND
A6. Topic description :
Core issues to be included in topic description:
sample size, sampling method and scope (inclusion and exclusion)
type of selection and reporting unit
planned structure of file including levels (include a diagram if possible)
For CURFs with previously released editions or earlier reference periods , the following issues
must be included in topic description:
populations changes from previous CURF
new and changed data items in proposed CURF in comparison to data items in previous
CURF
any other changes from previous CURF that may contribute to confidentiality risk
07
SECTION B - DATA ITEMS
B1. DATA ITEM LISTING
Detailed data item list which needs to include:
full details of all proposed data item categorisation, including ranges and top coding. This
must be the CURF (not the MURF) data items
concordance between Basic and Expanded CURFs (if applicable)
concordance between current and previous CURF releases (if applicable)
any data item categories that depart from CURF standards
The information needs to be provided in accordance with specifications outlined in the Data Items
Listing Checklist.
Include doclink to detailed data item list
B2. ONE WAY FREQUENCIES
Provide one-way frequencies and list failed frequencies (only). This information should be
provided using the 'One way frequencies report - template',
Include doclink to One way frequencies report
B3. ADHERENCE TO CURF STANDARDS
.
Departures from Standards should provide adequate justification for the DRB as
specified in
B3.1. Which sets of standards do you propose to apply to this CURF ?
Basic CURF income, labour, education surveys
Basic CURF other social surveys
Expanded CURF general standards
Expanded CURF indigenous standards
Other; details in the box below:
B3.2. Will there be any departure from the CURF standards for any proposed data items ?
No
Yes; details provided in the box below:
Data item
CURF standard
Justifications
08
SECTION C - SAMPLING INFORMATION
Note: This section is business collection specific , proceed to Section D for all other collections.
For business collections
To assist the assessment
of the CURF, you must provide information on whether there is a completely enumerated sector
of the sample to be included on the CURF and what are the criteria to completely enumerate a
sector. Units in the completely enumerated sector are more likely to require to be
confidentialised.
C1.
Is any part of the population completely enumerated ?
No
Yes; details provided in the box below:
C2.
What are the selection and reporting units for this collection ?
No
Yes; data item listed in the box below:
09
SECTION D - GEOGRAPHIC INFORMATION
D1 Please list all geographic data items to be released on the CURF . (If no geographic
variables, go to Section E)
Data item name
Data item categories
CURF standard
In addition to explicit geographic identifiers on the CURF, the data items, record identifiers, file
structure, or survey weights may provide additional geographic information by inference .
Therefore, steps must be taken to avoid inadvertently identifying geographic areas different from
the explicit geography on the CURF.
D2 Will the CURF records be listed in a particular sequence
No
Yes; details in the table below:
Sequence CURF Steps taken to reduce geographic inference
records are listed
D3 Are there other data items or combinations of data items on the CURF that may imply
geography of residence or business operation ?
Data items other than specific geographic data items may imply geography of residence or
business operation. For example, usual mode of travel to work shown as "tram" or "ferry" would
suggest a geographic location of Melbourne or Sydney respectively; Mining would implies
geography of business.
No
Yes; details in the table below:
Data item name
Steps taken to reduce geographic inference
D4 Is the level at which the sample weights are calculated different from the level of geography
on the CURF ?
No
Yes; describe steps taken to make sample weight less identifying in box below :
D5 Please list all data items and levels of detail used in the weight calibration process
:
Data item name
Level of detail used in weight calibration process
D6 Is there any other sampling information that you think might have geographic significance ?
Sampling information may also provide some geographic indicators. For example, certain
weights may distinguish between geographic areas.
No
Yes; describe steps taken to avoid disclosure of geographic indicator in box below :
10
11
SECTION E - ITEMS WITH UNUSUAL RISK OF DISCLOSURE
SPONTANEOUS RECOGNITION
E1. List all survey specific data items (i.e. not specified in the CURF standards)
either singularly
or in combination , considered as a risk for spontaneous recognition , and describe steps
taken to minimise the risk of disclosure
Data item name
Steps taken to minimise the risk of disclosure
E2. Are there any data items that will be top and /or bottom coded or have their values
expressed as ranges , deciles, quintiles , or percentiles ? (Do not need to include CURF
standard data items)
No
Yes, details in the table below:
Data item name
Basic CURF
Expanded CURF
ranges, deciles, quintiles ,
ranges, deciles, quintiles , percentiles ,
percentiles , top/bottom code * (if
top/bottom code * (if applicable)
applicable)
*NOTE: The data item list should show the collapsed data items as they will appear on the
CURF.
E3. Is it possible for any collapsed data item (e.g top coded ) to be broken by other data items
or records included on file ? For instance, if number of persons in household is included on
CURF, action must be taken to delete any records for household members exceeding the top
code limit.
Examples of breaking col apsed data items
No
Yes; details in table below:
Data item name
Proposed action to address issue
E4. (Indigenous CURFs only ) Are there any records with Country of Birth and Year of Arrival not in
Australia ?
No
Yes;
list records or counts in the box below. These records may need to be masked:
E5. Are there any variables on the file or combinations of variables that could be used as a proxy to
create a new variable which has not been included on the file ?
No
Yes;
list the variables of concern in the box below:
INCOME OR OTHER DOLLAR VALUES /FINANCIAL DATA
12
Particularly high income/sales or significant losses is a visible characteristic of individuals ,
households or businesses. Therefore income data is considered as sensitive. Consideration
should be given to the way in which situations of high (+ or -) income will be shown on the
CURF. List matching of exact income value can also be a risk.
E6. Will the CURF contain income or other dollar value /financial data ?
No
Yes; details in the table below:
Data item name
Data item categories
E7. If the CURF contains income or other dollar value /financial data , please describe ways that
the extreme situations (+ or -) will be treated :
E8. For basic CURFs which are not income surveys , will equivalised household and /or income
unit income be included in addition to personal and /or household income ?
Not applicable - expanded CURF or basic/expanded Income survey
No
Yes;
analysis will be required as per instructions listed in the following document
13
SECTION F - MATCHING CURF DATA AGAINST EXTERNAL FILES
F1. Identify any external administrative list , register or data item (s) that could be matched .
Lists previously identified have been recorded in a master list , and any lists identified for this
CURF assessment should be added to the List Matching Register. Some previously identified
lists may be relevant to your CURF but you should ensure that the information is still current ,
before including in your submission.
F2. Have you completed the list matching checking using the List Matching Register ? (This is
not required if CURF is being released as an Expanded CURF only )
No
Yes
N/A - Expanded CURF only
All lists identified in the List Matching Register need to be included in question F 3 below,
specifying applied treatment or justification as to why treatment is not required . You must also
assess the risk of list matching, taking into consideration the factors listed in F3, such as the
motivation of the organisation.
F3. Measures to reduce potential for matching .
Efforts must be made to reduce the potential for CURF data to be matched to external or internal
data. Internal data includes other CURFs, as there can be overlaps in records sampled. You
should consider:
the number of data items available for matching purposes, e.g. whether data items
available may be deleted or recoded or whether 'noise' (i.e. small amounts of random
variation) has been introduced into the data items
the resources needed to perform the match
the motivation of organisations to perform the match
the age of the data collected
the accessibility, reliability, and completeness of the external and internal file, and
the sensitivity or uniqueness of the data collected.
Some factors that make matching easier are listed below and information is requested on what
are the steps that will be taken before the CURF is released to reduce the matching potential.
(NOTE: This information is essential even if you are not aware of any external files that could be
used in matching.)
Proposed CURF list/dataset/organi Preliminary DRB Final DRB
Treatment applied
includes (select all
sation
meeting
meeting
that apply)
(Include the Data Risk assessment Risk assessment
item(s) /
(risk level before (after treatment)
subpopulation
any assessment
identified )
or treatment)
sample case Not applicable
Low
Low
selected from
Medium
Medium
a list provided
High
High
by a source
outside the
ABS.
14
Describe how
and by whom
sample cases
were selected
from the list:
every member
Low
Low
of a
Medium
Medium
subpopulatio
High
High
n. For
example,
large
employers,
high income
individuals,
doctors,
scientists of a
specified
type, or
inmates of
certain types
of institutions.
data item(s)
Low
Low
that might
Medium
Medium
isolate a
High
High
small, readily
identifiable
population.
For example,
detailed
occupation
data, or
detailed
family type,
particularly if
available in
combination
with
geographic
data.
data item(s)
Low
Low
from external
Medium
Medium
source(s) that
High
High
could serve
as a link to
external file.
For example,
energy
consumption
or cost
records from
an energy
supplier;
welfare or
social
15
security data
from a private
or
government
agency;
mortgage
details from a
lending
institution.
data item(s)
Low
Low
that might
Medium
Medium
identify
High
High
residence in
or attendance
at a particular
type of
institution of
which there
may be only
one in an
identified
area. For
example,
boarding
schools,
nursing
homes,
educational
institutions,
etc.
data item(s)
Low
Low
frequently
Medium
Medium
used for
High
High
matching. For
example,
exact date of
birth, sex,
and race; or
other items
that are likely
to be identical
on both files
such as an
exact income
amount,
exact date of
entry to
Australia.
other issues
Low
Low
that might
Medium
Medium
facilitate list
High
High
matching
none of the
Not Applicable
above
16
SECTION G - MATCHING CURF DATA WITH OTHER CURFS WITH
OVERLAPPING SAMPLE
G1. Does the CURF have overlapping sample with other CURFs released or other CURFs
which are expected to be released in the
future.
No
Yes
Identify where there is overlapping sample with CURFs already released and where there will
be overlapping sample with CURFS being released in the future. The DRB timetable includes
a list of CURFs released and upcoming. If you have any questions talk to MAS.
G2. Identify all CURFs that have been released or are expected to be released where there is
overlapping sample with this CURF .
CURFs already released :
CURFs expected to be released :
G3. Measures to reduce the risk of disclosure should a match occur .
You must reduce the risk associated with CURFs having overlapping sample with other CURFs .
Consider:
what are the match keys between the files, how many data items are available for matching
purposes?
consider the types of files eg Expanded accessed via the RADL or Basic CURFs on CD
ROM
the amount of sample overlap (how many records)
the extra information a user would gain from matching the CURF files
DRB have an agreed framework for mitigating the disclosure risks. Read through the framework
and follow the procedures outlined in part 2 of the document
17
SECTION H - OTHER ISSUES
IDENTIFICATION OF INDIVIDUALS
Well known or high profile individuals and businesses in the survey sample are more likely to
be identifiable, for example, large corporations, politicians or media personalities.
Additional measures need to be taken to preserve the confidentiality of these
individuals or businesses.
H1. Is it known whether any well known or high profile individuals are in the survey sample
?
No
Yes;
describe measures to preserve the confidentiality of these individuals
in the box below:
SPECIAL TABULATIONS FROM AN INTERNAL VERSION OF A FILE
Confidentiality problems may arise if special tabulations are made from an internal version of a
file such as the MURF, which includes detail omitted from the CURF and are then released to
external clients. For example, the tabulations might provide specific geography not included on
the CURF cross-tabulated by multiple data items on the file.
There is also a risk that special tabulations could be used in comparison with CURF data ,
(where different levels of detail or data item categorisation is included on the CURF ) to the
MURF. Special data request made to information consultancy must be closely monitored to
ensure tables cannot be used in conjunction with the CURF to breach confidentiality of
respondents through differencing the results from the CURF with the tables.
H2. Do you have any concerns about confidentiality problems which may arise if special
tabulations are made from the internal main unit record file , which includes detail omitted
from the CURF ?
No
Yes;
describe your concern in the box below:
H3. Are any restrictions to be placed on future special tabulations as a result of this CURF being
released?
No
Yes;
describe the restriction (s) in the box below:
18
Document Start
Policy and Legislation : Section 12-09
Header
Final
Manual Category
Ü
B. Policy and Legislation
v2010/02
Manual ID - No & Title:
Ü
Policy - Policy and Legislation
Last Updated:
17 Jan 2014
Chapter No. & Title:
Ü 12. Confidentiality and Disclosure
Section No. & Title:
09. Disclosure of Confidentialised
Microdata (Clause 7)
Subsection No. & Title:
Document Version:
Ü 2010/02
Status
Status:
*** Final ***
Clearance Authorised:
Australian Statistician
Contact Info
NOTE: This policy is currently under review. Some or all of the information contained within this policy
may be outdated. For further advice, please contact the data custodian.
DISCLOSURE OF CONFIDENTIALISED MICRODATA - CLAUSE 7 POLICY
KEY POINTS
1
The purpose of Clause 7 of the
Statistics Determination 1983 is to assist and encourage
informed decision making, by enabling wider access to ABS data in the form of unidentified
individual statistical records (microdata), for research and analysis purposes.
2
Clause 7 allows the Statistician to disclose, for statistical purposes only, a confidentialised
microdata or unit record file (CURF), in a manner not likely to enable the identification of the person
or organisation to which the record relates.
19
5
Clause 7 of the Statistics Determination is the main means by which microdata can be
disclosed. For information about the other limited provisions for the disclosure of microdata, see
the Frequently Asked Questions section below.
6
The Statistician must provide his/her approval in writing to release a new CURF under
Clause 7.
7
While the release of CURFs has become an expected output for most household surveys,
the Statistics Determination is an enabling provision only. Release of information under Clause 7 of
the Statistics Determination is at the discretion of the Statistician (and his/her delegate).
POLICY (INCLUDING DELEGATIONS)
9
It is ABS policy that two broad types of CURF can be made available to external users under
Clause 7:
a
Those that are confidentialised to a degree well beyond the requirement that their
release is not likely to enable the identification of a person or organisation. These are
known as basic CURFs. Basic CURFs are available on CD ROM, and users can therefore
access them in their own environment; and
b
Those which contain finer levels of detail and are confidentialised to a lesser degree,
such that additional protections are required to ensure their release is in a manner not
likely to enable the identification of a person or organisation (for example, to prevent
them being linked with other datasets)
. These are known as expanded CURFs. Expanded
CURFs are accessible only within an ABS controlled environment, such as the secure
Remote Access Data Laboratory (RADL), the Remote Execution Environment for
Microdata (REEM), and ABS Data Laboratory (ABSDL) environments.
10
It is ABS policy that CURFs will be released from most social surveys, unless there is an
expectation of no demand from major clients, or it is not possible to produce a useful file because of
confidentiality restrictions.
11
It is ABS policy that CURFs shall not be released before the first publication from the
relevant collection has been released.
2
09/11/2016
20
12
Consistent with our legislation, it is ABS policy that all directly identifying information must
be removed from microdata to be released under Clause 7. Such information includes name,
address and any other information (such as register numbers issued by other bodies) that identify
particular persons or organisations.
13
Once any identifying information is removed, it is ABS policy that microdata to be released
under Clause 7 is confidentialised (or modified) to reduce the risk of a person or organisation being
identifiable within the resultant CURF. The Disclosure Review Board (DRB) has been established for
the purpose of closely examining and assessing this level of risk within CURFs.
14
It is ABS policy that DRB advises and assures the Statistician that the risk of identification is
sufficiently low to meet the requirement that disclosure of microdata under Clause 7 is 'in a manner
not likely to enable the identification of a a person or organisation'. In doing so, DRB may request
that certain actions be undertaken prior to the finalisation of the CURF. It is ABS policy that the
relevant subject matter First Assistant Statistician assures the Statistician that the actions requested
by the DRB are undertaken.
Approval to create a new CURF suitable for release under Clause 7
15
It is ABS policy that the Statistician must provide written approval for the creation of a new
CURF suitable for release under Clause 7. This function has not been delegated. The Statistician
must be satisfied that the disclosure is in a manner not likely to enable the identification of a person
or organisation. As part of this, the approval must specify the type of release (for example, whether
the CURF will be available on CD ROM, RADL/REEM or just the ABSDL).
16
It is ABS policy that all requests for a written approval to create a new CURF suitable for
release be submitted to the Statistician through the Audit, Policy and Parliamentary Liaison Section.
For a repeat cycle of a survey that has a previously released CURF, approval in principle should be
sought no less than 10 working days prior to the planned release of that CURF. However, for new
surveys, or those with substantial changes to the nature or content of the material proposed for
release, the approval should be submitted no less than 20 working days prior to the planned release
of that CURF.
Delegations for releasing approved CURFs to users
18
Once the Statistician has approved a new CURF for release, the actual release of that CURF
to users has been delegated (as described below), with the delegate being accountable for ensuring
the conditions of the Statistician's written approval for the creation of the CURF, and all other
administrative requirements of Clause 7 and this policy are met.
3
09/11/2016
21
19
It is ABS policy that the release of an approved
basic CURF to an organisation or individual
within Australia must be approved by the Director, Microdata Access Strategies Section (MASS),
conditional on the signing of an appropriate undertaking (see section on undertakings below).
20
It is ABS policy that the release of an approved
basic CURF to an organisation or individual
outside Australia must be approved by the Statistician, conditional on the signing of an appropriate
undertaking. A written approval in principle from the Statistician is required prior to seeking the
undertaking (see section on undertakings below). For additional considerations about the release of
CURFs outside Australia, see the Frequently Asked Questions section below.
21
It is ABS policy that the release of an approved
expanded CURF to an organisation or
individual
within Australia must be approved by the relevant subject matter First Assistant
Statistician, conditional on the signing of an appropriate undertaking (see section on undertakings
below).
22
It is ABS policy that the release of an approved
expanded CURF to an organisation or
individual
outside Australia must be approved by the Statistician, conditional on the signing of an
appropriate undertaking. A written approval in principle from the Statistician is required prior to
seeking the undertaking (see section on undertakings below).
23 It is ABS policy that only the Statistician may approve the release of a basic CURF for inclusion in
an international database. In addition, under 3(d)(i) Clause 7 of the 1983 Statistics
Determination, the statistician may provide approval in writing for the CURF to be disclosed to
their users through the international database without obtaining undertakings from these
users. Before seeking the Statistician’s approval, the SMA must ensure that three additional
conditions have been met:
a. The SMA has assessed the protocols for accessing the database;
b. The DRB has reviewed the CURF
and provided their advice accordingly; and
c. Undertakings from the owner of the international database have been obtained.
24 For additional considerations about the release of CURFs outside Australia, see the Frequently
Asked Questions section below.
Undertakings
4
09/11/2016
22
25
Consistent with the requirements of Clause 7, prior to the release of a CURF to a user, an
undertaking must be given by the individual (in the case of an individual) or by a responsible officer
(in the case of an organisation) to whom it will be released. Clause 7 specifies that each individual in
an organisation signs a relevant undertaking only 'if the Statistician considers it necessary'. This has
been the approach in the past, however, it is now ABS policy that an undertaking, given on behalf of
an official body or any other organisation, covers all members of that organisation.
26
To be considered a member of a particular organisation, that organisation must take legal
responsibility for the member (e.g. students at a particular university are considered members of
that university, as the university is legally responsible for the students' use of an approved CURF).
27
It is ABS policy that approval for an organisation to have access to a CURF does not extend
to any of their consultants. In these circumstances, the consultant must gain separate approval for
access to the CURF.
28
Consistent with the requirements of Clause 7, it is ABS policy that such an undertaking must
state that:
a
No attempt will be made to identify particular persons or organisations to which the
information relates;
b
The information will be used only for statistical purposes; and
c
The information will not be disclosed to anyone else (or outside the official
body/organisation being given access) without the approval in writing of the
Statistician.
These restrictions on disclosure apply to the individual statistical records themselves and
not to transformed aggregates derived from those individual statistical records. The
production of single cell aggregates for basic CURFs are not in breach of the undertaking as
the level of confidentialisation of basic CURFs is such that individual records are not likely to
enable identification.
29
It is ABS policy that for requests involving supply of basic CURFs to international databases,
the subject matter area should undertake an assessment of the protocols for storage, and access by
individual researchers, in addition to each organisation who has access to the database having to
sign an undertaking.
Period of access to a CURF
30
It is ABS policy that there is no renewal process and no fixed period of access for basic
CURFs.
31
It is ABS policy that access to an expanded CURF is for a fixed period only, with access
renewal required at least annually. Access to the CURF is to be removed if the annual renewal
process is not completed.
Charging for access to CURFs
5
09/11/2016
23
33
It is ABS policy that secondary distribution arrangements, where disclosure to a third party
end-user is arranged at the discretion of an intermediary, are not permitted
LEGISLATION
35
The release of information in the form of unidentified individual statistical records is
authorised by Clause 7 of the Statistics Determination:
STATISTICS DETERMINATION 1983 -- Clause 7
Disclosure of unidentified information
(1)
Information in the form of individual statistical records
may, with the approval in writing of the Statistician, be
disclosed if:
(a)
all identifying information such as name and address
has been removed; and
(b)
the information is disclosed in a manner that is not
likely to enable the identification of the particular
person or organisation to which it relates; and
(c)
the Statistician has been given a relevant undertaking
by each person required by subclause (2) to give a
relevant undertaking in relation to the information.
(2)
The persons required to give a relevant undertaking are:
(a)
for information to be disclosed to an individual — the
individual; and
(b)
for information to be disclosed to an official body:
(i) the responsible Minister in relation to, or a
responsible officer of, the official body; and
(ii) if the Statistician considers it necessary in a
particular case — each individual in the official body
who will have access to the information; and
(c)
for information to be disclosed to an organisation
other than an official body:
(i) a responsible officer of the organisation; and
(ii) if the Statistician considers it necessary in a
particular case — each individual in the organisation
who will have access to the information.
(3)
In this clause:
relevant undertaking means an undertaking
in writing that use of the information in relation to which
the undertaking is given is subject to the following
conditions:
(a)
no attempt will be made to identify particular
persons or organisations to which the information
6
09/11/2016
24
relates;
(b)
the information will be used only for statistical
purposes;
(c)
for information to be disclosed to an individual — the
information will not be disclosed to anyone without
the approval in writing of the Statistician;
(d)
for information to be disclosed to an official body or
other organisation:
(i) the information will not be disclosed to anyone
outside the body or organisation without the
approval in writing of the Statistician; and
(ii) if the Statistician considers it necessary in a
particular case — the information will not be
disclosed to an individual in the body or organisation
who has not given a relevant undertaking;
(e)
if the Statistician considers it necessary in a particular
case — either or both of the following:
(i) the information, and all copies (if any) of the
information, will be returned to the Statistician as
soon as the statistical purposes for which it was
disclosed have been achieved;
(ii) access by officers to information, documents or
premises will be given as may be necessary for the
purpose of conducting a compliance audit concerning
observance of the conditions under which the
information is disclosed;
(f)
any other condition that, in the opinion of the
Statistician, is reasonably necessary in a particular
case.
BACKGROUND
36
The Statistics Determination provides for the disclosure, with the approval in writing of the
Statistician, of specified classes of information under very restricted conditions, which would not
otherwise be legal following the proclamation of the
Census and Statistics Amendment Act (no 2)
1981 on 1 March 1983. The Statistics Determination thus enabled the continuation of a number of
long standing practices followed by the ABS in releasing statistics.
37
More information on the background to the Statistics Determination can be found in the
Explanatory Statement accompanying the legislation when it was introduced.
38
The ABS includes in its Annual Report a table of releases of new unidentified information
approved during the previous year under Clause 7 of the Statistics Determination.
7
09/11/2016
25
8
09/11/2016
26
9
09/11/2016
27
REFERENCES AND RELATED INFORMATION
52
Census and Statistics Act 1905
53
Statistics Determination 1983
54
In the first instance, any queries relating to the application of this policy should be discussed
with your own line management. Any further queries should then be directed to the Audit, Policy
and Parliamentary Liaison Section.
55
If you are having any technical difficulties eg. broken links, please contact the Audit, Policy
and Parliamentary Liaison Section
.
For further information, contact:
Head, Audit, Policy and Parliamentary Liaison
10
09/11/2016
28
Document Start
Policy and Legislation : Section 12-10
Header
Final
Manual Category
Ü
B. Policy and Legislation
v2010/01
Manual ID - No & Title:
Ü
Policy - Policy and Legislation
Last Updated:
24 Jun 2016
Chapter No. & Title:
Ü 12. Confidentiality and Disclosure
Section No. & Title:
10. Disclosure of Information to
enable the Statistician to perform
Relevant Functions (Clause 7A)
Subsection No. & Title:
Document Version:
Ü 2010/01
Status
Status:
*** Final ***
Clearance Authorised:
Director, Audit, Policy and
Parliamentary Liaison
Contact Info
DISCLOSURE OF INFORM ATION TO ENABLE THE STATISTICIAN TO PERFORM
RELEVANT FUNCTIONS (CLAUSE 7A)
KEY POINTS
1
Clause 7A of the
Statistics Determination 1983 allows the ABS to contract out work
which requires access to limited confidential information (as defined under the
Census and
Statistics Act 1905 ). The use of external resources must only occur for work that is not
possible or practicable to undertake internally.
2
Under Clause 7A, identifiable information relating to an organisation or business may
be disclosed. However, ABS policy limits the disclosure of this identifiable information
predominantly to business names and addresses that are essential for the ABS to
undertake routine processing functions (see paragraph 11 below).
3
Information about a person, family, household etc of a personal or domestic nature
may be disclosed, but it must be confidentialised.
29
POLICY (INCLUDING DELEGATIONS)
5
It is ABS policy that any disclosure of information under Clause 7A must enable the
ABS to make use of external expertise or resources to undertake work
within its legislated
functions that is either not possible or not practicable to undertake within the ABS's own
resources. This includes the involvement of contractors in routine processing functions such
as despatch of questionnaires and reminders. Such work may also include research
requiring the use of highly specific analytical skills which do not exist, and are not required
on an ongoing basis, within the ABS. Some recent examples of disclosure under Clause 7A
are included in Frequently Asked Questions below.
6
It is ABS policy that work approved under Clause 7A must directly assist the conduct
of work in the forward work program.
7
It is ABS policy that Clause 7A cannot be used for peer review of ABS data or to
facilitate pre-embargo access to ABS data by a paying client.
8
It is ABS policy that disclosure of information under Clause 7A will only be approved
when it is assessed as being the only viable means of assisting the Statistician to perform
specific statistical functions. Alternatives can be assessed as being not viable on grounds
such as cost effectiveness, timeliness, lack of ABS resources or lack of ABS expertise
(where this expertise is not required on an ongoing basis within the ABS).
9
It is ABS policy that all approvals for disclosure under Clause 7A must have an
expiry date, with renewal taking place at least annually. Note that a single approval may be
given for a series of disclosures of information under Clause 7A for a nominated person or
organisation (including contractors) to undertake related specified activities over a defined
period of time. In these cases, the wording of the approval and the undertaking must satisfy
all legislative and policy requirements, and each disclosure must be clearly specified (e.g.
regular mail outs of ABS forms for a particular quarterly survey, for up to a year).
11
Under Clause 7A, identifiable information relating to an organisation or business may
be disclosed. However, it is ABS policy that disclosure of identifiable information is limited
predominantly to business names and addresses that are essential for the ABS to
undertake routine processing functions (such as the production of despatch labels for
business surveys). If you have any doubts as to whether a proposed disclosure would be
consistent with this policy, please contact Audit, Policy and Parliamentary Liaison Section.
12
Information about a person, family, household etc of a personal or domestic nature
may be disclosed, but it must be confidentialised. It is ABS policy that an abundance of
caution must be applied when considering whether to release information of confidentialised
personal or domestic nature (that is, in a manner that is not likely to enable the identification
2
09/11/2016
30
of a person) under Clause 7A. All releases which are approved will need to go through a full
assessment by the Microdata Review Panel, similar to releases under Clause 7.
13
The
Statistics Determination is an enabling provision only. Therefore release of
information under Clause 7A of the determination is at the discretion of the Statistician. It is
ABS policy that Clause 7A will only be used when the benefits of disclosure are clear and
would not have an adverse affect on: the trust of the providers who supplied the information;
those about whom the disclosed information relates; and the broader community.
Seeking approval for disclosure of business information
14
It is ABS policy that in seeking the approval for the disclosure of business information
under Clause 7A of the
Statistics Determination , the following three-step process must be
undertaken:
a. approval-in-principle by the Statistician;
b. arranging the contract, MOU or other agreement and obtaining signed
undertakings; and
c. formal approval by the Statistician.
15
Details relating to each of the above steps can be found in the Clause 7A procedures
document.
Seeking approval for disclosure of unidentifiable personal or domestic information
16
It is ABS policy that in seeking the approval of the Statistician for the disclosure of
personal or domestic information under Clause 7A of the
Statistics Determination , the
following four-step process must be undertaken:
a. agreement to proceed;
b. approval-in-principle (which includes the Microdata Review Panel assessment of
the information to be release, similar to Clause 7 releases);
c. arranging the contract, MOU or other agreement and obtaining signed
undertakings; and
d. formal approval.
17
Details relating to each of the above steps can be found in the Clause 7A procedures
document.
LEGISLATION
18
The disclosure of information to enable the Statistician to perform relevant functions
is authorised by Clause 7A of the
Statistics Determination :
STATISTICS DETERMINATION 1983 - REG 7A -
Disclosure of information to enable the Statistician to perform
relevant functions
(1) Information relating to a person, other than information of a personal or domestic
nature which is likely to enable the identification of that person , and information relating
to an organization may, with the approval in writing of the Statistician, be disclosed to a
person or organization where:
(a)
the disclosure is to assist the Statistician
3
09/11/2016
31
to perform statistical functions; and
(b)
the Statistician has been given a relevant
undertaking for the purposes of this
clause.
(2) A reference in this clause to a relevant undertaking shall be read as a reference to an
undertaking in writing by:
(a)
in the case of information to be
disclosed to a person, being an
individual — that person; or
(b)
in the case of information to be
disclosed to an organization — each
proprietor, partner, officer or employee
of that organization who has access to
the information, that the information
will be used only to assist the
Statistician to perform the statistical
functions as specified in the undertaking
and that the information will not be
disclosed to another person or
organization other than a proprietor,
partner, officer or employee of the same
organization who has also given the
Statistician an undertaking pursuant to
this clause.
REFERENCES AND RELATED INFORMATION
21
Census and Statistics Act 1905
22
Statistics Determination 1983
4
09/11/2016
32
23
In the first instance, any queries relating to the application of this policy should be
discussed with your own line management. Any further queries should then be directed to
Audit, Policy and Parliamentary Liaison Section.
24
If you are having any technical difficulties e.g. broken links, please contact the Audit,
Policy and Parliamentary Liaison Section
.
For further information, contact:
Head, Audit, Policy and Parlimentary Liaison
5
09/11/2016
33
Document Start
Corporate File
Methodology Solution Cycle
Document
34
The Methodology Solution cycle
The methodology solution cycle has 5 broad phases, although in reality they overlap, and
iterate. They are :
1 Initiation and Research : Problem recognition and formulation of methodological
framework
2 Experimentation : Pilot test potential solutions
3 Development : Develop infrastructure, and commence implementation
4 Delivery and Adoption : Develop expertise and achieve benefits
5 Support and Review : Continuous improvement, and adapting to new situations
1 Initiation and Research
Unlike the IT development cycle which is often driven by new opportunities, methodological
research is usually demand driven. Either a recurrent pattern begins to emerge from
problems encountered in existing operations, or the workprograms of client areas signal a
need for solutions to new problems.
Initial steps are to recognise the problem, and identify its most salient methodological
features. This usually involves identifying the general issues underlying specific existing
operational problems.
Then begins a search for potential solutions. Frequently MDMD staff develop solutions from
scratch, by modifying existing methods, or inventing new ones. Increasingly we are seeking
to look outwards, by maintaining a good knowledge of the relevant literature, attending
conferences, visiting and being visited by colleagues from other NSOs, and especially by
utilising the
We also have fruitful links with the
in the price index area.
During this stage there is little demand on our internal clients. MDMD staff may need access
to existing data to investigate the nature of the problem, but typically once permission has
been granted, MDMD staff access and manipulate the data without much assistance from
clients. Timelines are difficult to specify precisely, and our major internal need is to ensure
that the problem merits the degree of effort being put into it, and for feedback about the
relevance and suitability of potential solutions. It is a stage where external scanning and
consultation can prove particularly helpful.
2 Experimentation
Theoretical solutions must be proved in practice. And often, experimentation and trial is
needed to guide the development of the theoretical solutions in the first place.
During this stage the demands on the clients we are seeking to serve begin to grow. If the
potential solutions involve new algorithms, MDMD staff may well develop code to implement
them as part of their research work. But frequently new data will be needed, or some form of
field trial of the new method. MDMD does not have an independent budget for trials, nor any
operational capacity. Essentially we need keen guineau pigs, willing to explore the new
method as part of their pilot test work, or as an add-on or modification of live operations
(carefully treated to avoid risk to statistical outputs of course).
As trials proceed MDMD also needs active involvement in their evaluation, with feasibility,
and cost concerns typically needing to be offset against the methodological merits of a new
method.
35
Timelines begin to firm up, though flexibility is still needed, as the very purpose of conducting
a trial is to uncover aspects which have been overlooked, or new issues which emerge from
practical experience.
At the end of this stage MDMD staff know what needs to be done.
3 Development
To proceed to this stage, MDMD looks for and needs commitment from client areas.
Information about the new method must be disseminated beyond the initial pilot sites, and be
considered by potential adopters who, till this stage, may not have been closely involved with
the project. New problems may emerge, or more likely, additional complications to existing
ones.
Adoption of new or improved methods means changes to systems, both IT (which implies an
investment of TA resources, usually) and people. In ABS, this usually requires both
acceptance by the senior executive of our client areas (to authorise the changes and supply
funds for any necessary developments), and by individual operational areas, who will have to
adopt the new methods and systems once developed.
If the project proceeds, timelines and budgets tend to be fairly well specified (though
inevitably we move through the various stages of the IT hype cycle). When done well,
planning will also begin for the deployment of the new methods, with MDMD looking to
explain and educate about the new techniques, and operational areas making provision in
forward workplans to adopt the changes.
The challenges here are to produce systems and methods which fit into existing processes
and systems. Very rarely does a methodological change revolutionise an entire operation.
Usually a component of a process is improved or replaced. End to end issues (both in an IT
and a business process sense) require careful attention.
At the end of this stage MDMD staff have the necessary tools and know how to do what
needs to be one.
4 Delivery and Adoption
The latter stage of development, or initial stage of delivery and adoption, involves
acceptance testing and signoff on the new systems. In an IT sense, this may mean testing
new tools within the context of existing systems. While MDMD staff can check that the
algorithms in new tools operate correctly, client areas must confirm that the new and old
components link together correctly, receive the correct inputs, and send the outputs on
successfully to the next stages in the processing cycle.
During the deployment period, there is a need to fine tune the application of the new method
to each specific instance of its use, and to explore and fine tune assorted parameters and
options. During this stage, within MDMD, knowledge about the new method must spread
from the team involved in the research and development, to the wider group of people who
support all the clients who will be adopting the new method. And our MDMD expertise has to
grow from the theoretical, and that garnered from experimentation and pilot tests, to a broad
and deep familiarity with practical applications.
Each adopting client area needs to review and quality assure the outputs, to make sure that
all aspects (statistical and operational) of the new method are understood and performing as
designed.
36
At the end of this stage MDMD staff have experience in applying the solution, and know why
various options are chosen and in what circumstances.
In fact, in an ideal world, this understanding and appreciation would be spread to the client
areas. The new methods would be understood, and the consequences of various choices
well appreciated by client areas. MDMD staff would move from their role of supporting a new
method, to one of consulting and advising about the more complex aspects, or particularly
tricky new problems which emerge during the operational use of a well understood existing
method.
5 Support and Review
Once a solution has been adopted, there is an ongoing role for MDMD in supporting and
maintaining the solution. One very effective means of support is ensuring that the solution is
widely understood and effective documentation exists that clearly communicate how the
solution works. It can take time to build up to this! So part of this step is continually
improving the quality of documentation, the user-friendliness of the solution, and the
knowledge and understanding of the solution throughout ABS.
This step also encompasses on-going support for the solution. As documentation improves,
and competency in / knowledge of the solution outside MDMD improves, there should be a
gradual but consistent decline in the amount of routine support and maintenance required
from MDMD. However, it is also quite possible that new and unexpected problems may
continue to crop up for some time, requiring MDMD assistance.
It is also important to regularly review and evaluate the solutions we implement, to check
they are performing as required, and to continually improve them in light of what we learn
about their operation in practice.
37
The methodological solution curve !
38
Document Start
Policy and procedure for dealing with the risk of matching CURF data with
other CURF data with overlapping samples
This policy and procedure covers the risks associated with releasing CURFs that have overlapping sample with other CURFs released, or
CURFs expected to be released in the future. The risk occurs when the same records may appear on multiple CURFs due to an overlap in
sample. Because most CURFs contains a number of demographic variables these common variables can be used to match records on one
CURF with records on another (or possibly multiple other CURFs). If a users was to match the records from the multiple files this increases the
risk of disclosure. Efforts must be made to reduce this risk.
This document is split in two sections:
Part 1: General framework for assessing and mitigating the identification risk
This section outlines the issue and concerns when there is overlapping sample for CURFs. It also includes DRB's general comments for
assessing and mitigating the risk.
Part 2: Procedure for mitigating the risk of disclosure - Application of the framework
This section includes a step by step guide for MAS and survey areas.
PART 1: General framework for assessing and mitigating the identification risk
1) As an overall mitigation action, any masking applied to the records of one file should be applied consistently to the same records in
subsequent files released. Additionally, where subsequent masking is recommended for an overlapping unit, this masking should be applied to
the survey specific items (ie additional masking cannot be applied to the overlapping data items that have already been released on an earlier
file) of the file proposed to be released, and must be carried out in such a way that any assessed disclosure risk would be adequately mitigated
2) The level of risk of disclosure is probably increased from combining the data items on the overlapping CURFS. Identifying the level of risk can
be done by a conceptual inspection of the data items on the multiple CURFs.
3) After you have identified the level of risk from the combination of data items, determine how likely it is for a user to link the multiple files
what are the match keys between the two files? what variables are common on the multiple CURFs
consider types of files eg Expanded CURF accessed via RADL or Basic CURFs accessed on CD ROM or other ABS releases including
clause 7as
may need to actually match the CURFS to assess how likely
4) Given there is a reasonable chance of a user linking the files an assessment and further action needs to be undertaken to reduce the risk of
39
disclosure.
identify any records at increased risk of identification from the combination of data items. For example, if the Basic CURF is already
released and looking to release overlapping expanded CURF then look at:
(i) high risk records on the Basic CURF and determine if any would have been masked if the additional information from the
overlapping CURF was also available and
(ii) high risk records on the Expanded file and determine if any would have been masked if the additional information from the
overlapping CURF was also available
5) Mitigate the risk
masking should be applied consistently across all files(ie if something is masked on one file it should be masked consistently on all other
files released)
consider whether administrative arrangements can/should be put in place to stop matching taking place
Example: implement an administrative access arrangement where users are provided access to only one file at the same time(where there
is overlapping sample). All saved files in RADL would be deleted when removing access to one file and being provided with access to the
next. If Basic/Expanded combination can request return of Basic file before access to Expanded (but can't rely on return of file).
in RADL ensure daily auditing to identify users attempting to extract a large number of records and/or large number of data items for one
record.
undertake additional masking identified in step 4.
40
41
42
Status:
BETA
Secure data linking environment
Functional separation
Functional separation is the part of the process for ensuring data linking activities are done in a safe and secure manner. It is an integral part of
the linking process and is a requirement for the ABS to maintain its accreditation as an Integrating authority. This involves placing project
members into separate roles during a linking process. Access to data will vary depending on the role that each project member performs. Doing
this enables the protection of individuals and organisations in the datasets being used for linking, as it prevents any one project member from
having access to both linking and analytical information from both datasets at the one point in time.
There are four roles as follows:
Librarian: A staff member in this role performs processes such as the acquisition of data to be used for linking purposes, standardisation of the
data and creation of the files for linking.
Linker: A staff member in this role performs the linkage of the two datasets. Their access is limited to fields they require for linking and clerical
review (when applicable), which can include identifying information such as name and address information
Assembler: A staff member in this role creates the Analysis file. Their access is slightly more expansive than the other roles, as they have
access to concordance files, which allow the assembler to identify the records on the analytical files that have been linked, as well
as having access to the analytical fields themselves. However, they do not have access to the fields used for linking.
Analyst: A staff member in this role performs analysis on the linked dataset. Their access is limited to data needed for analytical purposes,
which typically does not include identifying information.
Note that it is possible for a staff member of a linking project to perform more than one role, so long as they are not performing multiple roles at
the same point in time. Also note that there may sometimes be overlap between 'linking' and 'analytical' variables. For example, Age and Sex
are two variables that are useful for both linking and analysis.
The diagram below demonstrates how the different roles in functional separation can be applied to a linking project so that no staff member in
any role has access to all data on both files at any one point in time:
43
Using the diagram above, the following procedures would be conducted:
44
1. The Librarian provides linking fields and a linking ID they create for each record to the Linker. They also provide a concordance file to the
assembler, which contains the original and linking IDs for their respective file.
2. The Librarians for each analysis file provide the original ID and analytical fields to the Assembler(note that the analysis file may or may not
be separate from the linking source file, depending on the way in which data is stored by the custodian).
3. Once the files have been linked, the Linker provides a file containing linked record IDs to the Assembler
4. The Assembler uses the concordance and linked record files to attach analytical variables from both datasets to the linked records.
5. The Assembler replaces the linking IDs on the file with new'analysis' IDs. Once this is done, the file is passed on to the Analyst
Functional separation is maintained using this model, as no one role has access to both the linking and analytical fields from both datasets. The
Assembler here is a minor exception as they do have access to both the original and linking IDs from both datasets, however they do not have
access to any other linking field aside from the IDs themselves.
The way that datasets obtained for and produced by the linkage process are stored should reflect functional separation arrangements. A
number of storage locations may be required, with different access settings applied to each location. Some of these locations may be restricted
to only one type of role, whereas others may be a location that acts as a 'transfer' point for data from one role to another, for example when the
Linker needs to provide the linked records file to the Assembler.
The table below provides an example of range of storage locations that could be created for different datasets generated in the linkage process,
and the roles that should have access to these datasets.
File type
Roles allowed access
Librarian
Linker
Assembler
Analyst
Linkage source file*
Y
Analysis source file*
Y
Subset files containing linking IDs and linking variables
Y
Y
Subset files containing original and linking IDs
Y
Y
Subset files containing original IDs and analysis variables
Y
Y
Linked records file
Y
Y
Analysis file
Y
Y
Outputs produced from analysis file
Y
* - A separate folder for each of the datasets being linked should be created.
Audit reports for data linking environment
The data linking environment consists of secure linking servers that are accessible using remote connection. Access to these servers is
restricted to those who have been granted access by the DLC and TSD staff who maintain the servers.
In addition to the restricted access, audit reports are automatically generated on a daily basis. DLC routinely check these reports in order to
investigate any suspicious activity that may be occurring within the data linking environment.
45
