There is data for Gould's hypothesis. The human evolutionary record is one of them. The demonstrated effect of impacts are another. Evolution relies on the paleontological record and is testable in the past only by uncovering new fossils which fit some hypothesized trend whether it verifies some supposed evolutionary force or just change itself. If gouls predicts long periods of stagnation followed by rapid changes and that is what the record shows then he is verified. The more prevalent this trend is the more dominant that method of evolution is. It's true importance can only be established by tracing a large number of linneages over time. The idea that this is untestable (which I have heard) is pure bullshit. You might as well say evolution itself is untestable unless you can replicate it the lab for each and every purported case of evolution.