Market Research Industry has changed very substantially in the last 10 years, with a few very large multi-national corporations now commanding a large share of the market. This growth, a combination of acquisition and organic, has resulted in a complex set of problems for these businesses when conducting international and mixed mode research. Typically 5 or more separate and different data collection systems may be required in order to effectively complete the work in a reasonable timescale. Generally, this will mean 5 or more different CATI or CAWI scripts, leading to 5 or more different data sets in different places all leading to management complexity and time consuming and often complex data manipulation in order to report on the project.
What sort of system could be built to deal with these challenges, using contemporary technology and architectural designs?
This paper discusses the design considerations for such a system, as encapsulated in a software development project that has resulted in a new mixed-mode interviewing system called Bellview Fusion, from Pulse Train.
In particular the paper will explore …
The requirement for centrally controlled system.
Although data collection sites themselves may be geographically scattered, and some degree of local autonomy required, it is also a requirement that the system can be centrally monitored and controlled.
The requirement for a single data repository
All of the data should reside, in a common format, in a single, large, resilient data repository. This eliminates the need to reformat data at the consolidation stage, and allow near real-time reporting on results.
The requirement for customisation and automation
The systems should be open, with documented interfaces, and both programmable and extensible. This delivers a high degree of flexibility and the ability to integrate the data collection system into the workflow and financial systems of host organisations.
The requirement for Mixed-Mode data capture
Data collected via the Web (CAWI), phone (CATI), Paper (PAPI) and face-to-face (CAPI) should all reside in a common format in the data repository.
In an effort to offer more choice to respondents and to increase response rates, they must be allowed to chose their preferred mode of interview, and must be allowed to switch between modes as many times as required.
Considerable meta- and para-data must be maintained by the system in order to unravel the collected baseline collected data.
A tightly coupled/loosely coupled system
Where robust WAN/LAN links are in place these should be exploited.
However, if a ‘local’ site loses contact with the centre, it must cache sufficient information to allow it to continue to function in a disconnected state, with a re-synchronisation occurring when the network is re-established.
The employment of a distributed workforce must be catered for
Predictive/Auto-diallers at CATI centres must be available to local and distributed agents
All of these pose interesting technical challenges from a system architecture and component design perspective, and the paper will explore some of these issues as they relate to Bellview Fusion and modern data capture systems in general.
| Back to: Top | Programme | Page last updated on 31 August, 2003 |