Practically all observations in The Press or provided in comments are "anecdotal." That is, they aren't part of a controlled study (practically never done for transportation, due to the difficulty of defining an adequate control) or one that looks at a large enough amount of data to be credibly analyzed for statistical significance. That doesn't make them any less valid as random observations, and those with agendas will happily use or dismiss them as they desire. They're just that, though: random observations, which don't indicate a trend or have much context other than as a snapshot.

As for induced demand, in fact, it exists. It's a component of a realistic modeling process for projecting the effect and useful life of a project and the performance of a transportation network. In practice, in urban areas, it's more like latent demand - trips that aren't done under current conditions but that become reasonable with reduced congestion. It's not necessarily a bad thing - if you build something, you *do* hope somebody will come and use it, but you also need to account for how many latent trips will come out of the woodwork to know how much actual useful growth would be accommodated. Again, doing that accounting is standard modeling practice.