I don't believe that the data supports dilution of the viewing audience due to a change in viewing habits.
Gunther Heinrich's "All Star Trek Movies and Episodes in Two Charts" is no longer at
http://www.madmind.de/2009/05/02/all-star-trek-movies-and-episodes-in-two-charts/ where it used to be, at least since July. But his chart plotting the ratings for TNG, DS9, VOY, and ENT is visible at
https://scifi.stackexchange.com/questions/126469/what-was-the-highest-rated-episode-of-star-trek.
The key feature that I'd like to draw attention to is that, for each of the first three post-TNG series, there was a marked drop in viewership after each premier, a rapid decay over the first few episodes that occurred before the decay rate for each series settled into a flatter, steadier decline. To me, this indicates interest that people had in
Star Trek. People tuned in, hoping to find something they would like. But they failed to find what they wanted, and they turned away. Those who remained gradually lost interest over the run of each show.
In other words, many more people would have watched, had they found shows interesting. They actually sampled the shows, but they didn't find them interesting, so they didn't keep watching. The fact that they tuned in to sample the shows runs counter to the premise that the viewing audience was diluted by a change in viewing habits. If viewing habits had been the determining factor of the overall ratings level for each show, this sudden drop from a much higher level of initial viewership would not have occurred.