Psychiatrists, psychologists, and other mental health professionals who have studied obsessive-compulsive disorder have always been keenly interested in identifying its most fundamental psychological cause. The main problem with OCD is, after all, tormenting thoughts.

Freud thought the culprit was repressed unconscious conflicts from childhood. These conflicts, bottled up and hidden, were released in adulthood in transmuted forms as obsessions and compulsions. This theory has been proven wrong. After Freud, a leading theory held that people with OCD were overly perfectionistic. The person who washed his hands again and again did so in order to wash them perfectly. Studies show, however, that while OCDers are indeed often overly perfectionistic, this trait is even more common in the sufferers of other mental disorders. Another commonly held theory is that OCD sufferers have a special problem with uncertainty. They worry and obsess in order to gain a measure of certainty that makes them comfortable (and, of course, they rarely find it). This theory does, in fact, explain a good deal; and the therapeutic tactic of helping OCDers become more comfortable in the face of uncertainty is a widely used and valid clinical approach. But again, research suggests that problems with uncertainly are not specific for OCD.

In 1985, Oxford psychologst Paul Salkovskis suggested the idea that the critical factor in the development of clinical obsessions is an exaggerated sense of personal responsibility for harm that may occur to self or others. At first, it seems obvious, right? Of course obsessive-compulsives check the stove because they feel responsible for preventing danger—why else would they do it? Yet Salkovskis’s idea nicely explains a number of common clinical observations.

OCD sufferers, for instance, almost never obsess about purely chance events, such as being caught in an earthquake or a hurricane. Why? Because the obsessional patient plays no role in the occurrence or the prevention of such happenings. Obsessional individuals do, on the other hand, readily develop symptoms when they are put into situations where obvious harm may occur as a result of their actions. A striking example is the frequent onset of obsessions and compulsions in women after the birth of a first child. Approximately 20 percent of all females with obsessive-compulsive disorder suffer its onset (almost always related to the safety of their babies) at this time of unparalleled assumption of personal responsibility.

Another example: It has often been observed that people with obsessive-compulsive disorder experience a “holiday” from their disorder when they find themselves in a new environment—that is, until they become accustomed to the place. Once I admitted to the hospital a student who suffered from severe checking compulsions. He did fine for a few days; in fact, our staff wondered why I had admitted him at all. Then he started setting off fire alarms, ear-piercing whistles that sent our mental health unit into turmoil. Why did he begin compulsively setting off the alarms when he did? My patient later explained, apologetically, that obsessions dealing with harm happening to friends were a long-term problem. When he had arrived on the unit, however, he had felt calm because he didn’t know anybody. In a few days, however, his feelings began to change.

“I began to make friends,” he told me, “and I felt responsible for their safety.” It was then that he was driven by obsessions of fires to set off the alarms.

Several dozen well-designed experimental studies have lent strong support to Salkovskis’s theory. An article in Behavior Research and Therapy summarizes the results of studies using psychological tests to compare OCD sufferers with other groups of people. Obsessionals, it is clear, have many more “inflated responsibility” beliefs than “normals,” or than people suffering from depression, panic disorder, generalized anxiety, or phobias. These beliefs include both general assumptions (“I am often close to causing harm”) and specific interpretations of intrusive thoughts (“Since I’ve had this thought, I must want it to happen”). In no other mental disorder do people feel so accountable for danger.

This inflated sense of personal responsibility has been demonstrated experimentally in a number of ways. One consistent finding is that when an obsessional person is alone, both anxiety and the tendency to perform compulsions increase. Roz Shafran of the University of British Columbia, for instance, asked a group of people with contamination obsessions to touch a toilet seat and then refrain from washing. They were to perform the task alone, then in the presence of an examiner. Shafran found that obsessive-compulsives experienced much less anxiety and fewer urges to perform compulsions when another person was present—that is, when they felt less individual responsibility for the outcome.

In another study, psychologist Edna Foa of the University of Pennsylvania, a longtime leader in research in anxiety disorders, compared the responses of OCD sufferers, phobics, and a “normal” control group to differing levels of imagined danger: high risk (“You see that a person sitting alone in a diner is choking”), medium risk (“You see some nails on a road”), and no significant risk (“You see a piece of string on the ground”). In the last two situations, OCD patients felt more responsible for outcome, and experienced more anxiety and more urges to check that harm would not occur. Foa concluded that these findings indicate “strong support for the hypothesis that inflated responsibility is an important factor in obsessive-compulsive disorder.”

A related finding is that OCD sufferers appear to experience a unique information-processing error. Normally, people can readily distinguish between causing harm and failing to prevent it—a presumption that is extensively reflected in our legal code. To hit a person with your car, for instance, is worse than failing to keep someone from being struck. Obsessionals, however, do not appear to make this distinction. For them, researchers at Oxford University have recently concluded, “failing to try to prevent harm to self or others is the same as having caused the harm in the first place.”