Editor's note: Tim Macer is a U.K.-based independent specialist and adviser in the use of technology for survey research.

Coding verbatim responses is a bit like doing the dishes after hosting a dinner party: a somewhat tedious and time-consuming experience, but ultimately satisfying when you see the results stacked neatly away, ready for use later. At least, that was the case before dishwashers became commonplace. Spectacularly successful though CATI has been down the years, the technology for handling all the open-ends that pile up after the party is over is still stuck in the suds-and-scrub era.

After interviewing, coding open-ends is often one of the costliest parts of the overall process. One agency that runs a weekly omnibus estimates the total cost of asking an open-ended question is 20 times higher than asking a similar closed question. This is not always a cost that agencies can pass on in full.

Fast-forward to Web-based interviewing, and the cost gulf widens. It costs virtually nothing to collect 10,000 completes, but a small fortune to code all the "other specifies" and "Tell me what you..." questions that go with them. Not surprisingly, it is from the Web that a superb idea comes that could rid coding of its kitchen sink associations for ever.

Ascribe is a Web-based verbatim management system from Language Logic that allows coders, coding supervisors, researchers, and even end-clients to work with open-ended responses from research projects, and with each other, to achieve perfectly coded data in less time and without the effort. It works equally well across all interviewing modes, and has several users that swear by it for paper-based studies.

To get your surveys into Ascribe you need to go through a "load" process to import the verbatim responses. You can also import some other "closed" data. Ascribe is not an analysis package as such, though it does contain some surprisingly advanced analytical capabilities (more on that later). So, at the other end, the data must be exported in such a way that they can be tied into the rest of the data for the survey. The ease or difficulty of these stages depends largely on the different packages being used upstream or downstream. As Language Logic has entered into a sales partnership relationship with SPSS MR, there is particularly good integration with the SPSS family of products.

One clever feature of the load process is something Language Logic calls "incremental loading" that lets you upload data from your CATI or Web survey package on a daily basis. Throw in the whole lot, and Ascribe will automatically discard any duplicates, and only add in what has changed - though you can fool it if you are inconsistent with your interview numbers from one batch to the next.

Tame your open-ends

Once your data are ready for coding, you can use seemingly dozens of different ways to tame your open-ends. All start with the same assumption that the best coding tool is the human mind. This is no auto-coding system: The tools exist to support and empower the coder to take better coding decisions.

This flexibility means you can replicate traditional coding methods or explore new ways to work by giving the coder more discretion and allow him or her to evolve the list of coded answers during the coding process. For example, you can text-search then group items together using drag-and-drop, nominating one answer to give the group its identity and its text. You can then create more groups or use drag-and-drop to assign more individual answers to this group, and so on, until all of the answers are classified. It is as simple as that.

Although you are using no more than a Web browser, it is remarkable the extent to which drag-and-drop and right mouse clicks have been made to work, thanks to the Java application behind it all. The context-sensitive right-click menus are especially handy while learning Ascribe, when you reach one of those "What do I do now?" moments. The user interface is particularly well designed to be highly productive in use: There is almost no typing involved and the mouse button menus mean that most operations are achieved in one, two, or three mouse clicks. All of these precious seconds saved add up over the life of each project to massive time savings. The few places in the interface where the process or the terms used are counterintuitive are therefore more noticeable. But they are likely to be short-lived, as Language Logic has a reputation among its users for fixing things quickly.

The service is offered on an ASP (application service provider) basis, so once you have opened an account with Language Logic, you can start coding from your Web browser. Pricing is calculated on the volume of work you do, not on any software license or monthly rental costs.

Any location

Web deployment also means that you can work from any location. Ascribe was originally devised as a means to allow for coders to work from home, recognizing that the Internet can be a liberator for the many workers that must juggle work and home responsibilities daily. Equally, location-independent coding means that projects can be monitored and supervised from afar - and the unreformed workaholics, by visiting an Internet café, can even check on their projects while supposedly vacationing.

The system recognizes these different types of user. The administrator password allows you to set up usernames and passwords for other workers on your account and grant different privileges to coders, supervisors, research directors, and even clients. Allowing clients to observe or even participate in the coding process may be controversial for some agencies, but is likely to be very appealing for clients.

Making the information available, so that coding decisions can be monitored, improved and even reversed, is a strong point of Ascribe. It even provides a tutoring mode, so that a new coder can code the same data as someone more experienced; the system will then highlight the differences as an aid to learning.

Entire spectrum

Ascribe seems to work equally well across the entire spectrum of survey research, from large agencies like ACNielsen BASES, which is using Ascribe to code 1.5 million verbatims a year, to in-house research units with just a handful of projects to administer.

Mark Thatcher, vice president at ACNielsen BASES, considers Ascribe's greatest benefit to be the low level of technical expertise required to work with the data. "Coding is a non-technical, manual process," he observes. "In the past, before we could realize any efficiencies we had to increase the skill set of the end user. This need for training usually brought with it some initial costs and downtime that we would just as soon have avoided. Ascribe enables anyone with a basic familiarity with an Internet browser to use this tool and have power over the data."

Used at ACNielsen BASES to code Web survey responses, the tool is increasingly being utilized in a collaborative way to develop code lists for international studies. Ascribe is truly multi-lingual and, with double-byte support, supports all the non-Roman alphabets like Arabic, Hindi, and Japanese.

The move to using Ascribe has, in Thatcher's opinion, added status and value to the role of the coder within the organization. "The coders feel more empowered and more respected by both internal and external clients. And the coders love it. In some cases, they have claimed productivity increases of up to 200 percent."

Such large-scale use puts any system to the test, and Thatcher has encountered some problems relating pre-coded and verbatim responses on the same question. It has proven a nasty problem to solve, and the solution offered is, in Thatcher's opinion, only "satisfactory."

Another problem users report is in communication between users when several people are working on the same project. It is not easy to see what codes other people have defined or applied; some functionality to highlight these, or allow users to communicate would help.

Amy Hatton uses Ascribe virtually single-handed in her role as marketing research manager at information services provider and publishing firm LexisNexis in Dayton, Ohio. Among other information sources, LexisNexis provides online legal research services to legal firms and registered law school students.

Colleagues often had to give up on coding the open-ended responses from surveys and other internal and external projects, due to pressures of time. Rather than ignore this valuable information, Hatton selected Ascribe to help her complete the task in less time. "It is a great tool, because it is more efficient than the alternative, which is to do it by hand," she says. "In research, coding has got to be one of the most unfulfilling and tedious jobs, and this tool actually makes it interesting, or as interesting as it can be! You can go back to change or recode things, so it is very flexible."

Typically, Hatton can take care of a project with around 5,000 open-ended responses in a day, compared to around a week to do the task manually.

Ascribe also contains a versatile tab tool which allows you to crosstab any of the closed data you have imported as well as any coded questions. For each cell, it will give you a count of the verbatim responses from where you can drill down and see the original responses. It is another powerful facility, and something few tab packages can offer.

Hatton has found that, by importing all of her pre-coded questions and demographics into Ascribe, this crosstab tool is adequate for all her analysis on simple studies. When more is required, she exports the coded data to SPSS. Her only grumble is with the data importation process, which can be time-consuming, taking up to two hours on some projects. For more typical uses, imports do not take anywhere near that long, as they do not import all of the data, only the relevant fields.

Better results, less time

Ascribe is a shining example of technology and the Internet working at their best to get better results in less time and at less cost, while freeing us from the old constraints of location. In the office, at home or even in the office at home, it is all the same. And unlike a dishwasher, you can actually see exactly what is going on inside.