The Second Meeting of
the Working Party on the Application of New Technology
to Population Data
Singapore, 1-3 April
1998
STAT/WPA(2)/5(Australia)
March 1998
ECONOMIC AND SOCIAL COMMISSION FOR ASIA AND THE
PACIFIC
Working Party on the Application of New Technology
to Population Data
Second Meeting
1-3 April 1998
Singapore
Internet application
at various stages of population data collection
and dissemination: experiences of the Australian
Bureau of Statistics
Dr Rob Edmondson
Director, Technology Services Division, Client
Relations Manager
for Population Statistics Group and Methodology
Division
Australian Bureau of Statistics rob.edmondson@abs.gov.au
March 1998
Apart from an extensive Web
site, actual internet usage by the ABS is quite
limited due to strong policies on security and
privacy. There are, however, a number of non-internet
electronic gateways, and Lotus Notes has provided
an "intranet" within the ABS for many years. These
services could be hosted on the internet if different
policies were adopted. In addition, internet-related
technologies are being increasingly used in a
variety of dissemination products, particularly
CD-ROM products.
"Real"
Internet Use.
The oldest internet service
provided by the ABS is a Gopher site for subscription
clients containing thousands of timeseries updated
daily. More recently, the ABS has developed and
enhanced an extensive Web site (http://www.abs.gov.au)
based on Lotus Domino technology, which enables
the site to be maintained as a replica of an internal
Notes database. This site provides key national
indicators, together with more detailed data that
provides a broad statistical picture of Australia's
economic and social environment, some quite detailed
Census tabulations at local area level (including
optional map-based area selection), the main feature
pages of some publications, various information
and discussion papers, media releases etc. It
currently has around 10,000 'hits' per day.
The ABS provides internet connection
to all authorised employees via a
firewall. This provides email and limited Web
browsing services. We also
mirror a number of non-sensitive Notes databases
externally, including our Web site. Apart from
the main Web site, these connections primarily
provide general research and external communication
facilities such as email relating to this meeting,
and data dissemination capabilities. From a statistical
point of view, these services permit closer contact
with key client communities during early survey
proposal and design phases, better liaison with
related organisation (such as Stats NZ) on various
surveys and statistical issues, general release
of information papers; general public enquiries;
help-line product support facilities; release
of public-good statistics; product demonstration
and marketing; and another distribution channel
for the results of ad-hoc consultancies.
The ABS is moving to release
a wide range of information about its various
collections - concepts, sources, methods, public
good data, full descriptions of data potentially
available, etc. Similar information is being collected
about Australian Government collections generally
as part of the recent "Statistical Clearing House"
project, and this information will also be published
on the Web. The result is expected to be increased
consistency and comparability between collections,
but also increased interest in some of the less
visible or less frequently collected information.
Population statistics in particular provide a
wide range of information on a wide range of topics,
but at infrequent intervals, and should benefit
from the increased visibility, and the ease of
searching much larger volumes of information using
the web.
With electronic access to publications
increasingly available, we expect both increasing
use of data and reducing demand for pre-printed
publications. The result should be a better service
to the community, together with reduced cost of
dissemination. Provided that electronic commerce
develops in the manner expected, we also expect
to make increasing amounts of more detailed information
available at a fee. This data is currently unpublished
and available only though relatively expensive
consultancy services.
Using the Web for dissemination
also brings some challenges - providing consistent
and more detailed information in timely ways to
a greater variety of channels is certainly not
easy, and the information is then accessed and
presented by an increasing number of third party
tools. There is considerable comfort in publishing
only limited statistics on paper after careful
consideration of content, presentation, and explanatory
material.
There is a great deal of interest
in moving some of our data capture
operations onto the net, but security and confidentiality
concerns must be
adequately addressed first. So far, data capture
over the internet has been limited to non-statistical
data.
Non-internet
external connectivity that could be internet hosted
In addition to the internet
connections, the ABS maintains secure
communications facilities for a variety of administrative
by-product data
sources and field systems. These typically use
encryption and authentication facilities, and
don't provide direct connection to the internal
network. Administrative by-product data capture
for population statistics is rare compared to
the economic field, but opportunities exist, particularly
in demographic statistics, and in some social
fields with a strong institutional component such
as crime and health. In contrast, the use of external
connectivity to support field operations in Population
statistics is far more advanced.
Field managers were provided
with PCs & modems during the last Census for
use in recruitment, to enable manager-manager
information exchange, progress reports, and headoffice-manager
query resolution. The results were very encouraging,
and more extensive use is planned in the next
Census, perhaps extending to pay and allowance
automation etc.
Computer assisted interviewing
has been introduced into the household survey
program in the last few years with modem-equiped
notebooks being used as the capture vehicle. This
has provided the ability to field larger and more
ambitious surveys, with complex sequencing, in-field
editing and during-interview query resolution.
Instruments and workloads are downloaded to the
machines and completed workloads returned via
the modem links. (note: data is kept encrypted
at all times, both on the PC and during transmission.)
As mentioned above, the possibility
exists to permit questionnaires to be (optionally)
filled in on-line by respondents, but the ABS
hasn't tried this approach in the population statistics
field yet. There also hasn't been any significant
use of the internet in the population statistics
field for post-field query or non-response resolution,
(we don't even collect email addresses) but this
may emerge in time.
Internal
connectivity that could be internet hosted
Lotus Notes provides our primary
office automation environment, being extensively
used for email and discussion databases, and for
a wide variety of other tailored applications
- particularly administrative applications such
as organisation structure, automated leave processing,
management information systems, planning and tracking
databases, and the like. These have been generally
accepted and have proved very useful. The ease
of use and consistency of the browser interface
is particularly attractive for these occasionally-used
applications. We now have thousands of such databases
containing information ranging from an extensive
collection of executive reports, minutes etc,
to a large number of single-topic or single-workgroup
databases or single process/workflow databases.
There are also a growing number
of databases tailored to mainstream statistical
processes. Some are new ways of doing old things
- submitting batch jobs from a browser interfaces,
setting up dispatch and collection control facilities
in browser-enabled databases, accessing very detailed
management information via browser interfaces,
etc. Some are new ways of doing things - routing
edit failures and queries by email, collecting
and developing survey specifications in a way
that can be automatically converted into an optical
mark recognition questionnaire (or perhaps a computer
assisted interview instrument, or a SAS edit program
etc). Developing classification standards in a
manner that supports the publication of the standard
but which also feeds into computer based coding
systems and into the preparation of publications
by collections using the standards.
The ABS has also been trialing
the use of Lotus Notes as a computer assisted
data entry tool. The early results have been encouraging,
and further development is planned. These systems
could be fielded for direct respondent completion
over the internet once security and privacy concerns
have been addressed. Developments such as these
enable the full integration of previously disparate
facilities - the same database, indeed often the
same documents, provide for dispatch and collection
control, management information, data entry, query
raising, query routing, query resolution, etc.
Some
other (non-network) use of internet technologies
Web technology is sometimes
used without network connection being required,
primarily electronic publishing formats including
the use of acrobat format files (or similar) for
publications. We are currently trialing the electronic
delivery of some significant publications to key
clients in acrobat format. A key issue is ensuring
that the electronic and the paper versions are
the same. We are also interested in issuing complete
collections of publications to libraries and similar
institutions in a compact but accessible form,
particularly one that overcomes many of the limitations
of microfiche. In this we are not alone - we have
increasing been getting technical manuals and
the like in internet formats, but not from the
internet.
