I took the DLAB (Defense Language Aptitude Battery) along with all of the other military students who went to DLI (Defense Language Institute). DLI is without a doubt the most intense language school in the world, where you have to go from learning the alphabet to reading, speaking, and writing at a college level, and have at most 1 and a half years to do it. In the class that came after us, 2 people passed the arabic course out of 30.
As far as the predictability of someone's aptitude, the DLAB was a good baseline to show that a person had better language acquisition than the average student. When students barely met the baseline of 100 points on the test, they had an incredibly tough time in the beginning with a language as different from english as arabic. What the test didn't measure, however, is a person's will to persevere through the slow progress, constant mistakes, and utter frustration. Although the low scoring students had a tough time in the beginning, most of them passed in the end. The student's that ended up failing out of our class were, ironically, the students that did exceptionally well in the beginning. I mostly attribute this to them relying on the ease of understanding in the beginning, rather than studying at the level that the lower scoring students had had to learn how to do when they had started out.
Language learning is not a linear path where you can acquire x amount of words per day. Your language learning is affected by so many factors such as sleep, healthy living, happiness, motivation, among others. There also seems to be a point for almost everyone I was with where constant study results in a moment of clarity, where you start to believe in your skill in the language and lose the fear of failing in conversation.
All in all, I would say that learning a language is almost exactly like learning programming. The material to learn either is readily available on the internet, and each purely depend on your desire to learn them. There will always be those that just can't quite jump high enough to get past the initial hurdles, and that's what these tests are there to provide -- a way for them to get filtered out before money is invested in their ability. In that, it's similar in a way to giving someone a fizzbang test. It won't tell you that you have a great programmer, it will just give you a baseline assessment that you don't have someone that shouldn't be a programmer.
Fellow DLI alum here (Persian-Farsi, '98) and I agree with everything you've said. The DLAB was a decent stab at predicting my success of completing DLI, but it was grit and determination that allowed me to pass the DLPTS with 2s.
Funny that you associate it with programming-- I used that same hardheadedness to learn Python and now I write software for a living.
As far as the predictability of someone's aptitude, the DLAB was a good baseline to show that a person had better language acquisition than the average student. When students barely met the baseline of 100 points on the test, they had an incredibly tough time in the beginning with a language as different from english as arabic. What the test didn't measure, however, is a person's will to persevere through the slow progress, constant mistakes, and utter frustration. Although the low scoring students had a tough time in the beginning, most of them passed in the end. The student's that ended up failing out of our class were, ironically, the students that did exceptionally well in the beginning. I mostly attribute this to them relying on the ease of understanding in the beginning, rather than studying at the level that the lower scoring students had had to learn how to do when they had started out.
Language learning is not a linear path where you can acquire x amount of words per day. Your language learning is affected by so many factors such as sleep, healthy living, happiness, motivation, among others. There also seems to be a point for almost everyone I was with where constant study results in a moment of clarity, where you start to believe in your skill in the language and lose the fear of failing in conversation.
All in all, I would say that learning a language is almost exactly like learning programming. The material to learn either is readily available on the internet, and each purely depend on your desire to learn them. There will always be those that just can't quite jump high enough to get past the initial hurdles, and that's what these tests are there to provide -- a way for them to get filtered out before money is invested in their ability. In that, it's similar in a way to giving someone a fizzbang test. It won't tell you that you have a great programmer, it will just give you a baseline assessment that you don't have someone that shouldn't be a programmer.