I stand corrected. They have not declined nationally, they've stagnated. If you search for it you'll get the data, but here's a supporting quote:
...from 1973 to 2005... real hourly wages of those in the 90th percentile—where most people have college or advanced degrees—rose by 30 percent or more... among this top 10 percent, the growth was heavily concentrated at the very tip of the top, that is, the top 1 percent. This includes the people who earn the very highest salaries in the U.S. economy, like sports and entertainment stars, investment bankers and venture capitalists, corporate attorneys, and CEOs. In contrast, at the 50th percentile and below—where many people have at most a high school diploma—real wages rose by only 5 to 10 percent
Janet L. Yellen, President and CEO, Federal Reserve Bank of San Francisco, November 6, 2006
As for CPI-adjusted wages - even though CPI isn't perfect, it's still useful. Certainly better than the anecdotal evidence you're asking for. Also, if anything CPI is under-reported, so CPI-adjusted wages should be considered a conservative estimate, which is all we're after here.
...from 1973 to 2005... real hourly wages of those in the 90th percentile—where most people have college or advanced degrees—rose by 30 percent or more... among this top 10 percent, the growth was heavily concentrated at the very tip of the top, that is, the top 1 percent. This includes the people who earn the very highest salaries in the U.S. economy, like sports and entertainment stars, investment bankers and venture capitalists, corporate attorneys, and CEOs. In contrast, at the 50th percentile and below—where many people have at most a high school diploma—real wages rose by only 5 to 10 percent
Janet L. Yellen, President and CEO, Federal Reserve Bank of San Francisco, November 6, 2006
As for CPI-adjusted wages - even though CPI isn't perfect, it's still useful. Certainly better than the anecdotal evidence you're asking for. Also, if anything CPI is under-reported, so CPI-adjusted wages should be considered a conservative estimate, which is all we're after here.