Here are two articles that seem to be saying different things about public transit in the big cities of the US. This article is a rundown of the cities where large proportions of the population don't own a car. The percentages they come up with (27.9% of Chicagoans with no car, and 26.3 commuting to work with public transit) seem to apply only to actual city residents, not the suburbanites (who in Chicago's and many other cities' case comprise a much larger population than the city proper). So in some ways it tells an overly rosy picture of how many people are using public transit.
This other article tells a very different story, showing that for most cities in the US, even those with good public transit systems, driving is faster than public transit. There's a study where a team of researchers divided each city up into gridded squares, and then calculated square by square what the fastest way of getting to every other square in the city would be. The results show that in the majority of cases, public transit is slower than biking or driving. Now this study has a similar weakness to the other one, in that it doesn't look at the larger metro area. If they did, the parts of the geographic area where it is easier to get to by car would look even higher. But if they looked just at the city proper, but factoring in the traffic generated by the cars of the entire metro population trying to come in and out, public transit would look a lot more attractive, even in fact for many suburban commuters. Which I guess explains why the public transit ridership described in the first article I cited is in fact higher than you'd expect from the second article's studies.
Another funny thing I noted--Chicago doesn't compare well to DC in terms of the area of the city where it's fastest to get by bike. While most parts of DC you click show that a quarter to half of the city is best accessed by bike, no point in Chicago gives you much more than 8 or 10% of the city fastest reached by bike. I assume this is due to a number of factors. Chicago's grid layout really is an efficient way to get around by car, while DC has a grid that just doesn't quite work that well. Furthermore, Chicago is almost four times as big geographically as DC, so much of the city is just inherently farther away from any given point. Lastly, the proportion of Chicago that the researchers' program comes up with that is best covered by car includes a lot of sparsely-settled areas. Population density in Chicago is overall a bit higher than in DC, but it is very unevenly distributed. So you've got large sections of the North Side, lakefront, and near downtown that have really high densities of people and amenities (the type of stuff you'd be likely to be trying to get to), and then large swathes of the South Side with very low density of people, businesses, and other attractions. So for most of the city's population, trying to do most of the things they're likely to want to do, the effective size of the city shrinks, and the percent of that smaller city that is accessible by bike would go up.
In any case, what these maps have shown me is that my anecdotal impression is right--Washington DC is a great place to get around by bike, and when I have the choice, I'll take bike instead of the Metro.