"To do two things at once," said the Roman sage Publilius Syrus, "is to do neither." And this was 2,000 years ago, long before people tried to drive while talking on their cellphones and digging for tollbooth change and yelling at the kids and (ahem) listening to the radio.
It is difficult to appreciate the affect of multitasking on personal effectiveness. The fifteen seconds mentioned in the NPR article doesn't sound like much. But the additional evidence of amount of focus and other studies suggests there are complexities with trying to do two things at once that essentially guarantee we won't do either of them well. It is also clear to me and my colleagues that focusing on one activity at a time leads to a much calmer feeling at the end of the day. (But please don't ask to look at all the windows I have open on my desktop just now.)
In projects, multitasking is easier to define. For example: We have three parallel tasks to do, and each takes five days. If I work on each in sequence, I get the first done on the fifth day, the second on the tenth day, and the third on the fifteenth day. However, if I "multitask," and spend one day on each, the first is not completed until day 13, an extra eight days. That starts to look significant.