Some interesting quotes from a study of the efficacy of anti-terrorism measures (PDF):
Considerable effort has been made over the years by the Department of Homeland Security to penetrate into the mind of potential terrorists and to imagine which targets they might prefer to attack... Although the list has remained secret, there have been a number of leaks indicating that miniature golf courses are included, as well as Weekee Wachee Springs, a roadside waterpark in Florida... a Mule Day Parade, a casket company, a petting zoo, a flea market, a groundhog zoo, and some, but not all, Wal-Marts.
(Although, to be perfectly fair, some Wal-Marts are more deserving of being bombed than others.)
Since driving is far riskier than air travel, the extra automobile traffic generated by increased airport security screening measures has been estimated to result in 400 or more extra road fatalities per year. In comparison, the number of people killed worldwide outside of war zones since 2001 by al-Qaeda or by its look-alikes and wannabes stands at some 200-300 per year.
Exercises in security theater can have counterproductive effects in the case of terrorism. One preliminary study finds that visible security elements like armed guards, high walls, and barbed wire made people feel less vulnerable to crime. However, when these same devices are instituted in the context of dealing with the threat of terrorism, their effect is to make people feel tense, suspicious, and fearful apparently because they implicitly suggest that the place under visible protection is potentially a terrorist target. In other words, they supplied exactly the effect terrorists hope to induce themselves.
...the annual cost [of federal homeland security expenditures] ranges from $64 million to $600 million (or even more) per life saved, greatly in excess of the regulatory safety goal of $1-$10 million per life saved. Not only do these expenditures clearly and dramatically fail a cost-benefit analysis, but their opportunity cost, amounting to $32 billion per year, is considerable. It is highly likely that far more lives would have been saved if the money (or even a portion of it) had been invested instead in a wide range of more cost-effective risk mitigation programs. For example, an investment of $200,000 per year in smoke alarms will save one life, and similar examples can be found in other risk reduction measures or regulations.
Many more interesting tidbits at the link above. See also Schneier's commentary. The conclusion of one section could be applied to the whole field of homeland security:
Any analysis that leaves out such considerations is profoundly faulty, even immoral.
My feeling has always been that terrorist is a useless label that betrays a political agenda on the part of the speaker. When you get down to the nitty gritty, all you have are insurgents or criminals. And it's easy to tell the difference: just look at who you call in to deal with the problem. If it is the military, you've got yourself some insurgents. If it's the police, you're dealing with criminals.