Seems like a popular and workable strategy, but like a lot of strategies, how well it works depends a lot on the difficulty level and size of map.
I play Emperor difficulty, standard size map, random civ, random map type, default barbs. I estimate that in about 10% of my game starts would a Science Broker approach have a reasonable chance of success. In the about 50% of the game starts, an early war is a necessity to get enough room to breath. In another 25% of games an expansionist wipes another player out very quickly and becomes dominant very early. There is little chance for a human player to keep up with research in that case with my starting conditions. In some other games, some civs just wants to fight (early unique unit) and nothing can convince them otherwise.
Changing the starting conditions, such as a lower difficulty level, a large or huge map, raging barbs, a scientific civ (obviously), fewer players, would increase the percentage.
Large or huge maps give you more time to get rolling, as the AI does not get a movement bonus, so it takes a long time to deliver an army to your door. A large map also slows the tech trading among the AI players, which is the heart of their overall tech advantage.
Raging barbs helps because it crimps AI expansion as they always send one escort with settlers. More barbs also costs them gold early in the game when barbs are more dangerous.
In reading the board, I have come to the conclusion that Diety difficulty on a Huge map lets you try about as many strategies as Monarch on a Standard map. Not to say that the end game is as easy, but many more starting strategies have a chance of success.