Fixed that for you.
NYTimes.com : 393 errors. whitehouse.gov: 26 errors. physicsforums.com: 24 errors.
Haha interesting. Nonetheless, there are definitely certain pages that do hinder navigation (columbia's astronomy departmental webpage in particular). And many of them don't look very professional at all.
If you are deciding on your school based on the department webpage you probably aren't in the right discipline.
People obviously don't decide schools based on their department webpages. But the department webpage *does* matter when they initially stumble upon a webpage. Then they might leave it, or they might investigate it further (of course, the highly interested students will always investigate more, but there are many other talented students who may be applying to departments in other fields as well, and who have a limited time to investigate the webpage of each department). A very high number of people do leave the webpage on their first try. Plus, good webpages will also attract undergrads who know little (but who may be good undergrad researchers).
That's all well and good, but people don't go to graduate school for physics to design nice webpages.
Of course they don't. They don't go to graduate school to grade papers either.
Most professors don't really care when all is said and done. Professors don't advertise through their webpages either.
The fact is, people are spending more and more of their time online, and so webpages are becoming more and more important (well, young people are using Facebook more, but they will have to use webpages to reach professors). Good webpages do mean that more people will look at their research. Many of the most famous physicists (among undergrads today) aren't necessarily the ones with the most highly-cited research, but rather, the ones with the most web presence (Sean Carroll, in particular).
Plus, search engine optimization does matter. People are more likely to find your research if they google in "HII spectral lines" and you're at the top of the search results. And then they may be more likely to cite your paper.