Opinions are registered with dials that go high for good and low for bad; a red button signals you’ve had enough. (I Spikelost 60 percent of its audience in its first two minutes.) The graphsare so specific, they tell producers which lines weren’t funny andwhich characters caused a downward spike by merely entering a room.Results are broken down by demographics like age, gender, and evenrace. Networks pay more attention to their particular target. CBSwouldn’t want a room full of young viewers, and Comedy Central wouldn’tcare to have older ones. For a series like Brothers & Sisters, an intense family drama, ABC wanted viewers who would normally like to see talky family dramas. Mainly women.

Men’sscores dropped considerably during the talky relationship scenes ofthat series, says Castler, while women’s perked up. That’s a given.Another focus group truism: Women often lose interest when scantilyclad women are the focus, but men’s scores, no matter how bad theprogram was beforehand, shoot up. (In our mock test, the male criticswere far more, shall we say, conciliatory during the fight scenes in a women’s prison. Hmm.)

Thefrustrating part for network execs is that they still never really knowif a show will work or not. Television testing is hardly an exactscience, says Marti Noxon, who was executive producer of Brothers & Sistersbefore quitting in August. It’s tricky, she adds, since overvaluing orundervaluing the process could get you in trouble. Sometimes changesare good. Other times, they rob a series of its heart.

“It’s aweird process,” Noxon says. “And for creatives, it’s especially hard —to feel that you’re living and dying by dials. But imagine what we’dget if we relied on the tastes of business-minded executives.”

We know: I Spike.