I think that one of the main problems with these stories is that they are too general and too broad. The US is way too diverse and much too different from these kind of broad statistics to be meaningful.

For actually meaningful statistics, you have to look at the local level. If the modal share of bike commuters in your city has doubled in the past 5 years, that's going to be significant...at least for your community. If your city has added 100 miles of bike lanes and far more people are bicycling, that's going to be significant, too. Again, for your community.

But trying to divine whether statistics taken from the entire country are meaningful or not is much more difficult - an increase in biking in NY, SF, and Boston might be enough to nudge the car ownership metric down, but it won't really be significant for people who don't live in those communities...people who don't live in those communities may see no change at all.