So long as market investors are human beings rather than machines, market participants will be governed by emotion. The efficient market theory, as Warren Buffett states, works most of the time. But when unusual or exceptional news comes into play, a stock (and/or markets) nearly always overreacts.
The best book I have found on investing is "The Intelligent Investor". There is a clear picture of what works and does not work in investing, and why. There is a fair amount of analysis of the behavior of market participants.
Warren Buffett asserts that he doesn't have much use for what is taught in a typical college business class. As he points out, if professors understand stocks and markets so well, why are so few of them wealthy? People like Ben Graham, Buffett and Peter Lynch are not 'lucky'. They read a great deal, they have keen insight into what makes a stock go up and they are unafraid to buy when prices are low if prospects look good. I would prefer to emulate those who are truly successful rather than those who postulate about what may work.

I am a behavioral economist with a deep belief in the notion that human decision-makers deviate in important ways from the scientific principles laid down in modern rational choice theory. There is no doubt but that very many investors hold erroneous notions of the dynamics of price movements, and having a correct understanding will, on average lead to better returns on one's portfolio. Sheffrin presents the evidence for this position in an interesting and accessible manner.

Shefrin's main advice for investors is absolutely correct, and would improve the asset positions of many poor souls with idiotic notions of stock dynamics. His advice is that if you are not a gifted and dedicated stock expert, you should invest in a low-maintenance cost array of mutual funds, and above all, do not churn your stocks. It doesn't help to be smart, lucky, a stud with the girls, or blessed by God. Moreover, if you think you have one of the "gifted analysts" for a broker, you are to be counted as among the suckers who are never given an even break.

Shefrin has another thesis which he presents with great verve, but which is on very shakey grounds. This is that "gifted stock analysts" can on average, significantly out-perform the market. He believes this MUST be the case if a significant fraction of investors are behaving irrationality. However, there is another possibility, which is that stock brokers as a group gain from the excessive churning that irrational investors permit or ask them to do, but that it is impossible to "beat the market" except by pure luck or by personally studying firm fundamentals and future prospects.

Shefrin's data in favor of the "gifted analyst" is episodic and anecdotal, and there is plenty of data on the other side. For instance, in Malkiel's classic "Random Walk Down Wall Street", he relates the evidence that chimps throwing darts do as well as major brokerage houses. Sheffrin presents contrary evidence for a more recent period in which "gifted experts" outperform the random darts. New evidence, collected by Money magazine, shows that a group of experts did far worse than the darts in 2003. All of this evidence is spotty and anecdotal. The plural of anecdote is not data.

I am not convinced by this book that the efficient markets hypothesis, applied to final returns to investors (after payments to stock brokers and other transactions costs), is not correct. I think the author makes a mistake taking so strong a position when the evidence is so weak on this account. I am certainly not convinced that Malkiel's analysis is in any way overturned by new evidence.

However, if Shefrin convinces a few investors to act more sanely, he will have fulfilled an important social function.