Nine years ago at the blog -- reposting this partially because RadioLab hasn't gotten any better but mainly because I really like the title.

West Coast Stat Views (on Observational Epidemiology and more) 2024-11-21

Thursday, October 8, 2015

"So long and thanks for all the fish"

I've fallen into this trap before. I'm driving down the road channel surfing the radio and I come across a snippet of something that sounds interesting on one of the NPR stations. I stop and listen long enough to start to get into the topic before I realize that this is RadioLab. Now I am faced with a difficult choice: I can change the station despite having become curious about what's going to happen; Or I can continue to listen until the inevitable annoyance and disappointment kicks in. I have been in this situation often enough to know that these are the only two possibilities. No matter how promising the opening or how intriguing the subject, I will regret it if I listen to the whole thing. RadioLab beautifully illustrates the somewhat counterintuitive principle that if you're going to imitate someone, you are often better off imitating the mediocre than the great. The show is clearly trying to be the next This American Life. All of the cute touches and distinctive mannerisms are aped, but without any sense of taste, proportion, style, or restraint. The result is an overproduced, painfully self-satisfied show narrated by two grown men who can't get enough of each other. All of this might be forgiven if the people behind the show were anywhere near as smart as the TAL crew and had something of interest to say. One of the reasons that This American Life works is because what might otherwise tip over into excessive production on another show is supported by a foundation of extraordinarily solid journalism. No one is better at bringing clarity and insight to a big story like patent abuse or the financial meltdown of 2008. The chatty tone, the sound montage and all of the other potential distractions only serve to enhance the story because the reporting is so good. RadioLab specializes in even bigger topics like the nature of language. Unfortunately, this added ambition only highlights the producers' limitations. Instead of clarity we get oversimplification; instead of insight we get lots of TED talk style geewhiz pseudo-profundities. Which brings me to today's show ("today" being a relative term but anyway...). When I tuned in, a researcher was discussing her work with dolphins in the sixties. I stuck with it through the discussion of giving the animals LSD, but then they got to the weird part... When I turned the radio back on, the story featured a different researcher had moved to the present day. The methodology was more conventional but the annoyance factor was just as high. Putting aside an enthusiasm level that would have been slightly excessive had the reporter been the first astronaut to land on Mars, the approach to the underlying scientific questions was awful.Even Malcolm Gladwell would've thrown up. The big payoff also contained the most unintentionally telling part of the program, but first a little bit of background: according to the program (and I have no reason to doubt this), each individual dolphin has a distinct signature whistle. The reporter (who was also the producer of the segment) and the hosts consistently discussed this in anthropomorphic terms as the dolphins' names. As an experiment, the researchers had come up with something like a voice recognition system for dolphins that categorized certain sounds as "words" and would also "speak" certain new words that represent, among other things, individual divers. The big climactic moment came when one of the divers "spoke" her name, at which point one of the dolphins turned and "spoke" his signature whistle. This is where we hit the unintentionally revealing part. One of the hosts asked if this represents a major linguistic breakthrough, at which point the reporter suddenly switches to conscientious mode and tells us how rigorous the researcher is. We are told this would have to happen 35 times before... Then we literally get the sound cue from Close Encounters of the Third Kind and a discussion of all of the deep philosophical conversations we might have with dolphins in the future. This is what thirty plus years of TED Talks leads to, an entire generation of journalists and writers who think this is what science is about: take some isolated study or statistic, ignore the context and previous body of research, and instead start drawing sweeping inferences and telling elaborate narratives. Preferably with lots of cute banter.