meaning: The Macer View/13

	The Macer View
	Comment and analysis from Tim Macer Number 13, November 2005

Resources

Software reviews

Software database

Reports & Thinkpieces

Conference papers

Press articles

The Macer View

Conference analysis: The ASC Conference on 'Maximising Data Value' provided Tim Macer with some real food for thought.

At the Association for Survey Computing’s two-day methodology conference in September on the theme of maximising data value, it was apparent that the benefits of re-using data are often lost on a research industry obsessed with collecting new data. Speakers described strategies to harmonise survey design, open up and share research widely, to mix research modes and employ new data linkage techniques to bring together disparate datasets as a viable and at times superior alternative to fielding new studies all the time. Not that any of this appears particularly easy to achieve, from the observations of various presenters. Common research habits and poor integrated support in research software for cross-survey work seem to act as a considerable brake on innovation and some of the methods proposed require for a head for statistical methodology, certainly stronger than this reviewer possesses.

It was a coup for the event’s organisers that the first keynote speaker was Karen Dunnel, recently appointed National Statistician, and for whom it was her first public address in the role.

For the Office for National Statistics, a current priority is to harmonise a whole range of established and often overlapping longitudinal surveys to create a single continuous population multi-modalsurvey for the UK.

The project’s loftier goals, which brings together the massive Labour Force Survey, the General Household Survey and others, is to produce better estimates, eliminate current conflicts between different measures, and to increase the availability of small area data. But also, as Dunnell stated “the pressure is on to achieve economies each year and the belief is that will improve efficiency.” Ears pricked up as she revealed that ONS has been able to reverse declining response rates by retaining interviewers to handle refusals better, as well as using mixed modes and reissuing sample more effectively.

In a fascinating glimpse into research buyer territory, Philip Cookson, a research director at Microsoft HQ in Redmond, revealed how Microsoft was moving towards a more integrated approach to its research and could potentially be setting a standard by doing so.

“We wanted to generate integrated insights across studies,” said Cookson. “But with over 100 research vendors, often using different technology, it is impractical to enforce a single software vendor approach.”

Instead, Microsoft is in the process of creating a central database-driven repository of questions, questionnaires and survey response data, built around the open QEDml standard which Cookson helped to found when previously working at Philology. The software giant will be encouraging its research vendors to reduce the amount of set-up and reprogramming being done by using QEDml or creating XML translators to import questionnaires automatically.

Three finalists for this year’s ASC/MRS Technology Award finalists were announced at the event: Gamma Associates’ AuditHost, Pulse Train’s Bellview Fusion and IMS’s Research Reporter. The winner will be revealed at October’s Research Excellence and Effectiveness Awards.