Statistical technique: roadmaps

12 October 2010

In every survey key questions are always asked on how things have changed, what has improved, whether policies have worked, and what developments are going on. Statisticians from ORC International are now able to robustly quantify these changes, by looking at results of repeated surveys and asking themselves the question "has the level of X changed or have things remained the same?"

This analysis works in conjunction with segmentation, a statistical technique that identifies common characteristics of distinct groups. The new technique called "roadmaps", shows shifts in segments and what has driven those shifts.

For example, in a hypothetical study on how work-based learning is perceived by employers internationally, in 2006 the UK might have been identified as one of the countries in the segment "WBL neutralist", which is statistically correlated with employers saying that learning has to be outside working hours. In 2010, UK employers might have moved to the segment "WBL advocates", where only 20% are reporting that their learning has be outside working hours, and now 80% support learning in working hours.

We could then look at the drivers for these shifts such as policy developments, financial incentives, applications of technology and changes in social attitudes. This is done in close collaboration with experts and by scrutinising survey responses.