In academics, GPA and tests generally fail as good statistical indicators of ability. Statistically, sample sizes are too small, variances are not satisfactorily accounted, and metrics like IQ tests and the SAT measure a very specific dimension of intelligence among potentially thousands of dimensions that affect performance on completely different tasks. (I recently wrote a piece that explains at length how academic metrics fail to accurately represent and assess the worth of students as experts and professionals, at https://www.facebook.com/notes/dan-strano/dis-asses-sing-merit-six-arguments-why-our-educational-metrics-are-screwing-us/10151576358443590).
The same problems befall pre-employment screening and interviewing. We rely on too few indicators that do not necessarily correlate with technical and creative abilities necessary for particular jobs.
Some employers look at GPA and standardized tests, which fail as predictors by my arguments above. Some set a hurdle for numbers of years of work experience in the same field, which tells an employer that a candidate was in fact able to hack out a living in the field for years previously, but turns away potentially even more qualified candidates that have not yet broken into their target industry.
Pre-employment testing suffers from a lack of breadth in its ability to measure candidate ability and aptitude. Where years of data collected on GPA fails as a good statistical indicator, two to twelve hours of testing cannot hope to make a more definitive assessment, functioning at its best as a litmus test and at its worst missing large areas of expertise that would help fit a candidate to the job being tested for.
Writing, programming, and other game development skills like visual artistry share a fundamental similarity that make them difficult to teach and hire for: they are media, means, for any other projects, not ends in themselves. Understanding this, it becomes perfectly apparent that employers are using the wrong criteria for selecting writers and programmers, for example.
You do not want a “Java programmer;” you want a UI design specialist who can program Java. You do not want a “C programmer;” you want a fluids programming expert who writes excellent C. You do not want a creative writing graduate; you want someone with stories to tell and the writing ability to tell them artfully.
As an employer of game programmers, you should realize that familiarity with syntax and libraries is actually probably the least of your project's actual demands on the skill of a programmer. Physics simulation requires extensive 3D math, calculus, differential equations, general physics, or advanced topics like fluid and electrodynamic modeling. For mobile apps utilizing accelerometers and magnetometers, programmers need to understand the physics and linear algebra that turn raw sensor instrument data into useful physical information.
For game engine optimization, one should look to those with simulation experience in the physical sciences, which require high-throughput modeling software; most CS majors have actually been taught to stay away from optimization at all costs! Syntax varies trivially. Logic remains logic without regard to syntax, and your physics engine programmers must be physicists after a fashion.
Accreditation programs don't target this combination of skills, and lack of familiarity with a single library can poison the performance of a programmer on one four hour test. Perhaps the best way to ascertain whether a candidate has the necessary characteristic combination of skills is to ask for a portfolio.
For writers, the very periods of apparent professional stagnation in one's work history can actually improve the quality of one's writing. People in creative professions often go for significant periods with relatively little professional advancement to show in their work history—an odd job, a retail day-time occupation to cover expenses—while they produce new creative works, hone their craft, and accrue that intangible stuff of “life experience” that truly becomes fodder for future works and integrally shapes the styles of writers, visual artists, and musicians, while these periods present themselves as stagnant lulls on a one-page résumé.
Of course, one way to asses whether a candidate has been developing creatively or simply oversleeping is to pay close attention to portfolios. A lack of publication with a strong portfolio for writers might signal a lack of professional initiative, but it might signal something different entirely—that their work doesn't appeal to the market for the “ADD generation,” that it isn't the stuff of cookie-cutter best sellers.
The gamer community seeks out works like Evangelion and FLCL. As an artistic community, we can still distinguish the genuine article from formula pieces. Simply because an artist's work hasn't found its audience yet doesn't signal that the work cannot sell. Thankfully, gamers seem to want the next Braid or Limbo, not necessarily best-selling novel material.
Perhaps part of the reason why “indie” games are so successful as to have become a seeming genre in their own right is because big studios are still not accurately assessing the skills of aspiring developers and the saleability of their work, so new developers are forced to go it alone. Cave Story was a labor of love and art. Before its release, one could imagine Daisuke Amaya as a struggling college student and then graduate with a C.V. that didn't belie the project. How many times is a developer like this passed over simply because a large studio couldn't be bothered to try her or his game?