Editor’s note: Patrick Johnston is president of Vista Research Services Inc., Chicago.
Transferring survey research data between competing software packages has, until recently, been among the more time-consuming and costly tasks of the data processing professional. While software developers have ensured that their own products work well together, little attention has been paid to integrating these products with other software developed both inside and outside of the survey research industry. Given a number of recent trends and developments in information technology, this now poses less of a problem than it once did. Open standards, the expanding role of the Internet, and a desire to integrate data stores across the organization have all hastened the call for interoperability.
The most obvious and immediate need for integration is within the survey research industry itself. If a developer’s Web or CATI data collection package can export directly to the software they developed to create crosstabulations, why would a prospective or existing client want to look beyond this solution? Setting aside the developer’s interest in keeping users under its own software umbrella, several good reasons can be offered: the merit of competing products under consideration, the ability of these products to be integrated with software outside the research industry and their cost vis-à-vis other alternatives. While export and import programs may have been written in certain situations, a more comprehensive solution was needed to allow these disparate products to interface with one another.
Fortunately, a solution now exists which allows competitive software products to exchange data with one another: the Triple-S standard. The U.K.-based Triple-S Group (www.triple-s.org) is a vendor-neutral organization which espouses open standards and is backed by the Association of Survey Computing (ASC). Defined as a “means of transferring the key elements of entire surveys between different survey software packages across various hardware and software platforms,” the Triple-S standard has been adopted by over 50 companies worldwide that develop market research software. Implementers include well-known companies such as CfMC, Voxco, Global Market Insite (GMI) and Pulse Train. These implementers recognize the current trend towards allowing software to work together collaboratively. By design, it is relatively easy for software houses to write export and import programs to support the Triple-S standard. (The Triple-S specification can be downloaded directly from the Triple-S Web site, www.triple-s.org/sssdown.htm ).
The most recent specification, Triple-S XML 2.0, makes use of the Extensible Markup Language (XML). Developed by members of the World Wide Web Consortium (www.w3.org), XML is now in wide use as a means to deliver structured content over the Internet. Many industries, including our own, are also adopting XML as their markup language of choice for structured documents. As a result, outputs from the research industry’s Triple-S XML standard can more easily be integrated with other corporate data and systems. A number of MR software developers are going beyond the Triple-S standard by building other unique features of their software around XML, thus allowing for even greater integration with other products.
Two different types of data are output with the standard Triple-S specification: the metadata and the case data. The metadata is often described as “data about the data” and it is this data which is “marked up” using XML. It includes elements such as a survey’s title, the names assigned to questions/variables, question text, response text and variable types. Along with the Triple-S metadata, case data is provided and, simply put, this data contains the actual answers to the survey. Triple-S case data is usually in fixed, ASCII format with one record per respondent. A recent enhancement to the standard allows for case data to be in comma-separated file format (CSV).
Future seat
By adopting software standards such as Triple-S, the MR industry ensures itself a future seat at the table of corporate decision makers. Beyond the interoperability of packages that handle survey research data, we are increasingly being asked to integrate our data stores with other business functions such as operations and finance. With XML complementing and, in some cases replacing, traditional databases, the most recent XML specification of Triple-S looks especially promising. Our ability to integrate MR data with other departmental data should enable us to move lock step with other business units well into the 21st century.