Granting certification means that the applicant has been found to have reached a standard (minimum level) of competence. But how do you know an applicant is worthy of certification? On what basis do you make the judgment? The question goes to the matter of assessment, and any technical communicator interested in—or concerned about!—certification is interested in the answer.
There are several ways to assess applicants. Which is appropriate? Having studied quality, I know that any measurement has strengths and weaknesses, and any measurement can be subverted. But measuring along multiple dimensions gives a much more accurate result and is much harder to “trick.” It may be a good idea to establish multiple requirements for certification.

Examination
The first thing that comes to mind is examination. There are many variants of the venerable sit-down exam, including multiple choice, fill in the blank, and my favorite, the essay. The strength of this approach is that precise questions can be defined, graded objectively, and administered to many applicants economically, and a passing score can be established. The weakness is that it rewards people who are good at taking tests, not necessarily people who can apply knowledge, and that determining what to ask can be very difficult. Professions that have a defined body of knowledge can use it to define their exams, but we don’t have one yet.
But there are also tests that can establish that the applicant can accomplish some set task. For instance, to become certified, a pump repairman could be given a broken pump and told to fix it. It’s hard to argue with that approach. Nevertheless, many people I’ve spoken to get nervous about the very idea of testing. What else is there?
Education
One possible requirement is that the applicant complete one or more prerequisite courses. For instance, we could declare that anyone who earned a bachelor’s degree from an approved list of schools, attends a list of courses from a approved list of training companies, or simply earns a certain number of continuing-education credits in approved subjects can be certified. The strength of this approach is that we can piggyback on existing standards. The weakness is that it rewards taking courses, not doing work. (And who approves the lists of courses, degrees, and approved schools? That is accreditation, and it’s an equally vexing issue.)
Long-time working professionals hate this approach, arguing that you can become a technical writer without any formal education in the field. Holders of technical-communication degrees, wanting credit for their academic work, naturally feel it merits consideration.
Experience
Another possible requirement is work experience. For instance, we could declare that anyone with X years of experience in the field can be certified. The strength of this approach is that we directly tie competence to holding down a job, for if you’re done the work, you can do the work. The weakness is that the actual skills of people paid as “technical writers” can vary widely.
Unsurprisingly, the opinions expressed on education are reversed here: working professionals love it, new graduates hate it.
Portfolio
Another possibility is to require a portfolio of documents, work products, or both. Now, this is familiar ground for STC and its members: it resembles the publication competitions conducted by a number of chapters and by the Society itself. The strengths of this approach are that it directly measures the ability of the applicant, and that we know how to do it and have an infrastructure in place. The weaknesses are the subjectivity of judging (which is a known issue with the competitions), and that, like the competitions, it doesn’t scale well: the more applicants, the more judges, and the greater the cost.
CV
Finally, we could assess the curriculum vitae of the applicant, in the manner of a university tenure application. Again, STC has experience with this process; it resembles the Associate Fellow and Fellow nomination process. The strength of this approach is that we can ask for both requirements and endorsements (references). The weaknesses are possible breaches of anonymity and that, like portfolio assessment, it doesn’t scale well.
Which One?
Which one method is best for certifying technical communicators? I think none of the above. But a combination of at least two or three is probably a good idea. The answer that gives the strongest, most prestigious (or, for the applicant, most difficult) certification would be all of the above.
What do you think?