There's a new issue of the Journal of Applied Psychology out. The very last article is a study by Lievens and Sackett, investigating video-based vs. paper-based situational judgment tests.
Here's the low-down: 1159 students completed video version, 1750 completed paper version. Video version scores correlated less with cognitive ability (consistent with previous research, presumably resulting from the test having lower cognitive load) but had higher predictive and incremental validity in predicting interpersonal-related criteria than did scores on the written version. No significant difference in face validity between the two versions.
Don't know which scores correlated better with other criteria, and I won't know until I take a trip over to my local university to get a copy. (The articles are $12 on-line).
Bottom line? Another example of video-based tests successfully predicting job-related criteria, and another reminder that when we talk about predicting job performance we should ask, "what TYPE of job performance?" (BTW, this is often called the "criterion problem")