I don't have a lot to say about the first chapter of Class Dismissed. It lays out the case that economic inequality is real and growing in the US, and that the effects of inequality (relative economic disadvantage) are real, even if not as obvious as the effect of material disadvantage (i.e. disadvantage on an absolute scale rather than a relative scale). John Marsh is not an economist, but he's definitely well-read and he draws on a lot of data sets (primarily government statistics) rather than secondary analyses. Not being an expert on these sorts of macroeconomic statistics myself, I'd be reluctant to cite Marsh's analyses as the final word on any detailed issue, but the overall picture is reasonable enough and well-referenced.
I'm half-way through chapter 2, where he starts arguing that we can't educate our way out of this problem. I have issues with a few specifics, like his annoying habit of saying that when a particular type of labor is abundant the demand is low. I know what he means--he means that if there are lots of people available then employers won't have to search very hard or compete very much, therefore those types of workers are not "in demand" in the colloquial sense of the term. However, in economics there is a very important conceptual distinction between supply and demand. Also, on page 43 (to go back to chapter 1 but stay on the theme of harping on economics) he makes a rather assertion about the effects of immigration on wages, and his only reference is to a NYT article. I wish he had directly cited whichever primary sources are referenced in the NYT.
Still, his big picture analysis is dead-on: He keeps pointing out that if we flooded the market with more graduates in a particular area of expertise the compensation would go down. That reduction in salaries is definitely not an argument against educating people; even if we view education through a purely economic lens there are still economic benefits to having educated people available, and the people who get those jobs are (often) better off than they would be without that education. Still, the formula is not simply "These jobs pay $X, so educate more people and we'll have more people making $X." John Marsh makes that point where so many others fail to get it. Educating more people WILL reduce the college wage premium relative to less-killed jobs. The only question is whether the reduction in the wage premium will be pocketed by consumers (cheaper services), employers (cheaper labor, bigger profits), and/or the less-educated (college grads lifted out of their ranks eschew the jobs done by the less-educated, so the less-educated have less competition and their wages go up). It is by no means obvious that the gains won't be pocketed by employers, and the past few decades of economic history give us no reason for optimism.
He also notes that the "knowledge worker" jobs are often (though of course not always) the ones that can be most easily moved across national borders and done remotely, while many of the jobs most likely to stay "in person", i.e. not be outsourced or off-shored, are not necessarily jobs that pay well or require a lot of education. Again, this is not an argument against getting an education, but it is an argument against treating education as our society's guaranteed (and only) path to economic equality. The tea leaves do not necessarily point that way, despite all the protestations to the contrary.