Cancer, Correlation, Causation, and Counter-Terrorism

My first wife died of breast cancer in 1999. Over the seven-year course of her illness, I was forced to become something of an expert in the disease. While I no longer devote a great deal of time to the subject, I still keep a general eye on developments in the field; so I read with interest a series of New York Times articles on preventing cancer, including installments on diet, exercise, stress and the immune system, possible environmental causes, and cancer genetics. (I assume that there will be further articles in the series.) Quite apart from the subject matter – which is fascinating and important in itself – I was struck by the repeated emphasis on the difficulty of establishing the true causes of cancer, and the pitfalls in recommending lifestyle changes which have often turned out to be useless in preventing illness.

The classic tool used to search for the causes of cancer is statistical correlation: researchers examine and question large groups of people, attempting to find the significant differences between those who come down with cancer and those who do not. In some cases, this technique has been phenomenally successful – for example, cigarette smoking was identified as the single greatest cause of lung cancer long before science was able to determine exactly how smoking causes cancer. But for other potential factors such as diet, exercise, and stress, many years of correlation-based studies have failed to come up with much: according to the latest studies, neither eating lots of fiber, fruits, and veggies, nor exercising, nor chilling out, has anything like the dramatic effect on cancer risk that avoiding smoking does.

There are several reasons why it is so difficult to find effective ways of avoiding cancer. Without going into too much detail (you do remember that this is a blog about Israeli politics and the Mideast conflict, don’t you?), I’ll focus on two of them:

1. Most epidemiological studies are retrospective rather than prospective (see also here and here, “Study Objective, Direction and Timing” definitions 5 and 6) – meaning that they are limited by the quality of medical records and individual memory. While prospective studies avoid problems like recall bias and can yield much more precise and reliable information, they are expensive to run and can take decades to complete. For practical reasons, then, researchers first conduct retrospective studies to find interesting correlations, and only afterwards design prospective studies to see if changes in particular factors really do have an effect on cancer incidence. Most of the time, the latest and greatest healthy-living advice “based on new research” is actually based on retrospective studies, and so must be taken with a grain of reduced-sodium salt substitute.



2. Studies involving lifestyle issues and cancer must attempt to extract simple, practical propositions from an extremely complex reality. It can be almost impossible to untangle the many factors that may help to cause or prevent cancer. For example, it’s known that breast cancer rates are much lower in Japan than in the West, but second- and third-generation American immigrants of Japanese ancestry contract breast cancer about as often as other American citizens. This indicates that some environmental or lifestyle factors are involved, but given the huge number of potential variables – from micronutrients to food-storage methods to exposure to sunlight – it’s nearly impossible to isolate the ones that make a difference.

Of course, as scientific understanding of cancer improves, it’s easier for scientists to identify the factors that are likely to have an impact on cancer incidence; this doesn’t eliminate the need for epidemiological research, but it does allow the research to be more focused – yielding better answers in less time, we hope.

*          *          *

Terrorism, like cancer, is a complicated phenomenon with extremely complex – and, so far, incompletely understood – causes; but just as many of us latch onto simple “magic formulas” for avoiding cancer or recovering from it, we find it hard to resist simplistic thinking about terrorism. This is particularly true here in Israel, where terrorism, justifiably or not, reigns as Issue Number One in our political discourse. While we are arguably the world’s most sophisticated society in terms of our resilience in coping with terror attacks, and can boast one of the world’s most effective security establishments, our politicians and columnists (along with a fair portion of the electorate, I suppose) often discuss terrorism in dishearteningly primitive terms.

In reality, the level of terrorism at any given time is influenced by a wide variety of factors, including (but not limited to) the following:

• motivation (which may be economic, nationalistic, religious/ideological, social/personal, organizational, financial, or just plain perverse);


• capability (access to explosives and weapons, expertise, leadership, and so on);


• political considerations in participant and “spectator” societies;


• state sponsorship;


• accessibility of targets;


• deterrence;


• and a large dose of dumb luck, which may at different times favor the terrorists or their victims.

In confronting Palestinian terrorism, Israel has a number of options:

• It can attempt to reduce the Palestinians’ motivation to commit attacks.


• It can try to reduce the terror organizations’ capabilities by preventing weapons smuggling, killing or imprisoning terrorist leaders, or attacking terror organizations’ physical infrastructure (what there is of it).


• It can improve Israeli resilience to terrorism by supporting relevant educational programs, tailoring government rhetoric to put terrorism in perspective, and encouraging responsible media coverage.


• It can attempt to eliminate state sponsorship of terrorism by exposing it to the international community.


• It can physically separate potential terrorists from potential targets.


• It can attempt to create deterrence on the individual, organizational, and social/national levels.


• It can pray for a continued supply of dumb luck.

While many of these approaches have merit, few of them have a clear record of success in reducing the number and severity of “successful” terror attacks against Israel. Only physical separation (towards which I have always had mixed feelings) and “targeted killings” (which are politically costly and have only a temporary effect) seem genuinely effective to me in this regard. At the same time, much of Israeli political discourse about terrorism revolves around deterrence – which I think has more to do with our emotional reactions to terrorism than with any likelihood of success. (I plan to address the challenges of deterring terrorism in future posts. You’ve been warned.) Even when we speak about eliminating or reducing the terrorists’ capabilities – with our calls for “destroying the terrorists’ infrastructure” and our endless repetitions of the outstandingly fatuous “Let the IDF win!” – we seem to be focused as much on our need to feel empowered as on any probability that the policies we advocate will actually work.

In addition to the inordinate complexity of terrorism and the emotions that interfere with our ability to confront it rationally (analogous, perhaps, to “recall bias” in cancer studies), we are hampered by our inability to perform “prospective studies” to see what approaches work best under what conditions. For example, it has been frequently observed (usually by Netanyahu himself and his allies) that there was a relatively low level of terrorism during Binyamin Netanyahu’s term as Prime Minister, from 1996 to 1999. The implication we are supposed to draw is that Netanyahu was uniquely tough, so much so that the terrorists were afraid to carry out attacks while Israel was in his capable hands. I find this argument less than fully convincing; it seems to me that other factors were likely responsible for the reduced terrorism during the Netanyahu years. (I’ll admit that I haven’t made a detailed study of counter-terrorism under Netanyahu, so my thoughts shouldn’t be taken too seriously on this subject; I do suspect, though, that at least part of the reason for the comparatively low level of attacks during the Netanyahu years was that there was essentially no peace process to disrupt.) But since we can’t “play back” the same years with different policies in place to see what would have happened to us under different leadership, we can’t draw firm conclusions; we are stuck trying to connect nebulous causes with nebulous results.

If we – as politicians, voters, and would-be pundits – want to make rational decisions about counter-terrorism strategies, we need to realize the difficulties of conducting “research” on the subject: We have no General Theory of Terrorism, only a sketchy idea of how all the many contributing and inhibiting factors interact, tremendous emotional pressures towards choosing activist (and often violent) counter-terrorist measures, and no way to test different approaches scientifically. Like anyone trying to navigate in a fog, we would be well advised to cultivate calm, caution, and humility.