1211.0 - Data Communication - Emerging International Trends and Practices of the Australian Bureau of Statistics, 2006  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 10/02/2006  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All  

STRATEGIC DIRECTIONS

A. IMPROVING COMMUNICATION OF STATISTICS

12 The ABS' goal is to increase use of statistics and to improve understanding of the content, caveats, contexts and limitations of the data to assist user assessment of fitness for purpose of the data.


13 To achieve this we are working to provide more relevant, understandable and interesting content. We are also working on effective presentation of information on the web site by utilising opportunities available from web technology.


14 The ABS has recently reviewed literature on cognitive psychology and developed principles and guidelines for content preparation and on-screen presentation (Kraayenbrink, 2004). These address the three fundamental cognitive processes - perception, attention and learning. For a more detailed discussion of the application of cognitive psychology to improve the on-line communication of statistics, see Attachment 1.


15 Using the outcomes of our research, we have developed the following strategies for improving the communication of statistics on the ABS web site:

  • using a layered approach for the presentation of information;
  • developing basic guidelines for presenting/writing for the web;
  • contextual linking of metadata with statistical data; and
  • using the concept of web magazines to ensure that statistical stories are visible to web surfers.

These strategies are consistent with the practice adopted by newspapers in web publishing.


i. Adopting a layered approach

16 Our research on cognitive psychology suggested that using a layered approach is pivotal in reducing cognitive load, improving communication and supporting a diverse range of on-line users with different levels of statistical sophistication. The approach requires setting up information in layers, with the simplest information presented first, and the most complex information last, suitably hyperlinked to allow easy navigation from one layer to the next, and vice versa.


17 Whilst our web site has already been designed with the following data/information layers:

  • Statistical Headline News;
  • Simple story of the data (Main Features);
  • Detailed story (Detailed Publication) and Detailed statistics (Spreadsheets, Data Cubes and Time series); and
  • Related data,

consistency in populating the layers across the diverse subject matter fields within the ABS, and visibly presenting the layers on the web site, were only achieved through the development of a new web page design rolled out in late January 2006. For more information, visit www.abs.gov.au.


ii. Presenting/Writing for the web

18 Our research suggests that many users of web sites consume information on-screen. For effective communication of information, one of the important issues to consider is to ensure that the mind is able to handle the cognitive load. This requires, amongst other things, information to be presented in digestible but self contained chunks - ABS calls these "information nuggets" - with the nuggets suitably hyperlinked to allow interested users to pursue further details; reduction in the "density" of the information, reduction of "propositional complexity" i.e. the number of propositions in a sentence; and presenting information in an easily scannable form etc.


19 In addition, the presentation of the materials on the web site has to support visual perception (i.e. ability to attach meaning to objects presented on the web site) and the content written in a manner to support language comprehension (i.e. ability to connect to, and interpret, the written language). Information presented for on-screen reading should not be too dense, nor verbose, and should also be amenable to scanning for key words by web users. Our directions here are to develop good practice guides and training courses for presenting/writing for the web to achieve effective communication.


iii. Contextual linking of metadata

20 With the exception of electronic publications, the current ABS approach for disseminating ABS metadata on the web site (e.g. Directory of Statistical Sources) is based on a "big bang" approach - we give users a lot of the information, and let the user help themselves to the parts they need to know. Our cognitive research suggests that this is not an effective approach.


21 A more effective approach for communicating metadata is to adopt an approach similar to the one used for communicating statistical data - follow a layered approach and tailor to the needs of the statistical user - whilst ensuring contextual linking between statistical data and metadata. In doing so we aim to minimise cognitive load and improve communication of statistical caveats. Contextual linking is important to ensuring that statistical data will be used in the right contexts and are fit for purpose.


22 We are developing facilities on the ABS web site to provide linked web pages in our publications and in all of our time series spreadsheets and data cubes to contextually link metadata to statistical data . At the collection level, the web pages will provide information on the "Quality Declaration". At the statistical product level, these web pages will appear whenever the information icon next to statistical terms is clicked to provide information on definitions, statistical classifications, questions used in the collection etc.; or the information icon next to statistical data to provide information about statistical errors ie quality statements . The web pages will, as required, contain links to more detailed data or metadata. An example of this is shown in Attachment 2. A prototype for this has been built and is available at https://www.abs.gov.au/about/ePublication.


iv. Publishing web magazines

23 With the ability to hyperlink information on the web site, there are new and innovative ways for effective communication of ABS information. By merely reproducing our paper summary statistical releases electronically, there is a risk that users may omit reading important information, and experience a high cognitive load. This is because "context shrinks when reading on screen". Refer to Attachment 1 for further details.


24 An example to illustrate this point is provided by the Main Features of the publication 8501.0 Retail Trade, Australia (see https://www.abs.gov.au/ausstats/abs@.nsf/mf/8501.0). The Main Features step through "Key Figures", "Key Points", "Notes" and then provide some further key points under the headings "Industry Trends" and "State Trends". With no upfront indication that industry and state trends analyses follow the "Notes" section, it provides a "presentational cue" to the readers that the statistical stories are coming to an end, and could result in the industry and state trend analyses being missed.


25 A better way to present the information involves providing clearly visible and interesting information first. This provides the needed stimuli to draw readers' attention to these "stories" and entice them to read them by clicking the links. An example of how this may be presented is provided in Attachment 3. Note that we have, similar to the treatment by newspapers, "bundled" the three stories together to provide an organisational cue that they come from the same statistical collection, and are related in some way, in this case, different dimension of the analysis.


26 This method of presentation can be extended, of course, to not only summary publications, but detailed publications and compendium publications as well.



Previous PageNext Page