DISCUSSION

The present study found evidence that childhood accidents that produce unconsciousness are associated with

~ [Page 522] ~

pedophilia and with lower levels of cognitive functioning. These associations were statistically significant for accidents that occurred before the age of 6, but not for accidents that occurred between the ages of 6 and 12.

These results suggest that neuro-developmental perturbations in early childhood might increase the risk of pedophilia. That would have the further implication that erotic age-preference is not fully determined in utero, and that later events can influence a man’s relative attraction to adults versus children.

A secondary finding was that maternal history of psychiatric treatment — a reasonable proxy for the occurrence of psychiatric problems — was also associated with pedophilia. Maternal psychiatric history did not, however, correlate with lower levels of cognitive functioning.

These results raise several possibilities.

The first is that pedophilia may have multiple causes. That would make pedophilia analogous in one regard to homosexuality, for which the available evidence strongly suggests multiple causes

(Blanchard, 2001; Blanchard et al., 2002; Blanchard & Ellis, 2001; Cantor, Blanchard, Paterson, & Bogaert, 2002).

The second is that pedophilia may be influenced by genetic factors, which are manifested in women as an increased risk of psychiatric problems, and in their sons, as an increased risk of erotic interest in children.

The third possibility is that different etiologic path-ways to pedophilia may have different associated features.

Thus, for example, men whose pedophilia was ultimately caused by a head injury may have lower levels of cognitive functioning — or perhaps a different pattern of cognitive deficits — than men whose pedophilia was largely determined in utero by genetic factors.

The obvious limitation of this study

is the self-report nature of the childhood and family history information. There are at least two problems that must be considered under this heading.

The first is that of literal accuracy.
It is questionable whether all subjects who reported that before age 6, they experienced an accident that left them unconscious for at least half an hour had, in fact, been under age 6 when the accident occurred, had been left in a state that would meet clinical criteria of unconsciousness, and had remained in this state for 30 min or longer.

It is also questionable whether the rate of head injuries reported (10.2% among the pedophiles) is reasonable and believable. The latter question is difficult to judge on general grounds. On the one hand, a rate over 10% seems extraordinarily, if not suspiciously, high in a group of patients who were not referred because of problems in reasoning, memory, motor skills, or similar complaints. On the other hand, a rate this high might not be excessive if pedophilia is itself a type of neuro-developmental disorder.

Therefore the plausibility of this absolute value can be decided only by examining other pedophilic samples.

In summary,

it is safer to interpret the patients’ self-reports, not as historically accurate accounts of childhood head injuries, but rather as indications that the subject had some childhood injury involving a blow to the head — an injury serious enough for him to remember it directly or for it to have become part of family lore.

The question of literal accuracy is not critical for purposes of this study, provided that between-groups differences in self-report actually reflected between-groups differences in rates of head injury, and provided that intentional or unintentional distortions of the facts were roughly equal for the pedophilic and non-pedophilic groups.

This, however, brings up the second problem:

Did the pedophiles report more head injuries than the non-pedophiles because they truly had more head injuries, or rather because they were disposed to claim more head injuries?

One might hypothesize

that many pedophiles claimed head injuries that never occurred, or grossly exaggerated memorable but minor occurrences, as a way of diminishing responsibility for their sexual conduct toward children.

This is a reasonable hypothesis, but there are equally reasonable arguments against it.

The men in this study who were not phallometrically diagnosed as pedophiles also had actions to account for; a proportion of them had sexual offenses against adults, and another proportion had sexual offenses against children, even though they were not phallometrically diagnosed as pedophiles. Why should this group have felt significantly less motivated to explain or excuse their behavior?

Another argument against the hypothesis that the present results were simply artifacts of response bias is the finding that the phallometrically diagnosed pedophiles did not indiscriminately endorse all questionnaire items that suggested neuro-developmental problems or family histories of psychiatric illness.
Of the eight items we examined, they differentially endorsed only two:

accidents with unconsciousness before age 6, and

mother’s history of psychiatric treatment.

At this point, the objection might be raised

that two of eight items is what could be expected by chance. It should be noted that this is not an argument that the results were caused by response bias but rather an argument that there were no real results at all.

	This statistical argument is blunted by the fact that similar results were found in two independent samples.
	Also weighing against it is the fact that, of two head injury items with identical wording except for the timing of the injury, the pedophiles differentially endorsed the one that also correlated, in this sample, with lower intelligence and lesser education.

[Page 523]

Another approach

to assessing the general plausibility of the present data is to consider whether our finding that lower cognitive functioning correlates with head injuries in earlier rather than later childhood is consistent with other research.

Unfortunately, researchers disagree regarding age of injury effects. It was once believed that peri-natal or neo-natal brain damage had milder consequences compared with later brain damage

(e.g., Kennard, 1938).

In contrast, current researchers often predict more positive outcomes for injuries sustained at later ages when neural connections are more mature. Empirical findings are conflicting, with studies to support either perspective

(for reviews, see Aram & Eisele, 1992; Chapman & McKinnon, 2000).

Additionally, differences in how researchers categorize age and define outcomes likely contribute to inconsistencies. For instance, although studies including both children and adults have linked earlier ages of injury with better language production outcomes

(Bates et al., 2001)

and global recovery rates

(Overgaard et al., 1973),

studies of pediatric populations have found that children injured at younger ages have worse semantic memory outcomes

(Levin et al., 1996)

and higher mortality rates

(Mahoney et al., 1983).

Studies have also reported that age of injury interacts with injury type

(Bates et al., 2001; Chapman & McKinnon, 2000)

and severity

(Levin et al., 1996).

Long-term follow-up studies are few, although one longitudinal study found that functioning was worst among those adults who had sustained head injuries in the earliest years of life

(Koskiniemi, Kyykkä, Nybo, & Jarho, 1995).

Thus, in the current state of knowledge, it is not possible to evaluate the likely validity of the present data by comparing them to definitive findings in the general literature on childhood head injuries.

There is, to sum up to this point,

no way to eliminate response bias as a possible explanation of the present findings — the EPES questionnaire was simply not designed to address this type of issue — but there is also no compelling reason to regard response bias as the most likely explanation.

We can therefore move on to consider

what interpretations of the data are possible if pedophiles do, in fact, experience more head injuries before age 6.

The simplest and most obvious interpretation of such a relation is that childhood head injuries increase the risk of pedophilia. There is, however, at least one alternative interpretation that cannot be ruled out at this time: Some third variable — one that is chronologically and causally prior — both increases a boy’s risk of pedophilia and also renders him prone to childhood accidents.

This type of relation is readily illustrated with an analogy. Childhood head injuries correlate with gender: Boys are more likely to sustain head injuries than are girls

(e.g., Henry, Hauber, & Rice, 1992; Kraus & McArthur, 1996).

It is clear that head injuries do not turn a child into a boy. A more reasonable explanation is that prenatal events that induce phenotypic masculinity (e.g., androgen exposure) also increase the child’s activity level and thus increase the child’s risk of getting into accidents.

In the present case, it is possible that neuro-developmental perturbations in utero increase the individual’s risk of pedophilia and also increase the risk of some behavioral or perceptual problem that leads to accident-proneness.

One example of such a problem is attention deficit hyperactivity disorder (ADHD). Numerous studies have documented that children with ADHD have accident proneness and elevated rates of injuries

(see Barkley, 1996, for review).

The prevalence of pre-morbid ADHD and behavioral problems among children with head injuries is higher than that for control samples of children

(Brown, Chadwick, Shaffer, Rutter, & Traub, 1981; Gerring et al., 1998),

and, among children hospitalized for injuries, those with a pre-morbid diagnosis of ADHD are more likely to have sustained head injuries and to be severely injured compared to children without pre-morbid ADHD diag-oses

(DiScala, Lescohier, Barthel, & Li, 1998).

Thus, if pedophiles were to have elevated rates of ADHD, they would be expected to have greater rates of childhood head injuries than the general population.

Figure 3 illustrates the two different models of the relation between childhood head injury and pedophilia:

	the first model, in which head injury causes pedophilia, and
	the second model, in which it does not.

Fig. 3.
Alternative models for explaining the observed correlation between childhood head injuries and pedophilia.

[Page 524]

In the first (i.e., upper) model, head injuries cause neuro-developmental problems, which, in turn, increase the risk of pedophilia.
In the second (i.e., lower) model, prior neuro-developmental problems lead to accident-proneness and head injury, on the one hand, and to pedophilia, on the other.
In this model, head injury has no causal influence on pedophilia.

[Page 523 continued]

It should be clear, from the foregoing discussion, that we regard the present study as preliminary in nature and largely heuristic in value. Its limitations with regard to neuro-psychiatric assessment should not completely overshadow its strengths with regard to sexological assessment, namely, two large and non-overlapping groups of phallometrically diagnosed patients.

Satisfactory research in this area requires valid sexological assessment just as much as it requires valid neuro-psychiatric or neuro-genetic assessment, and the present research compares favorably with any previously published in the former regard.

We therefore feel that this investigation provides ample justification for future studies specifically designed to elucidate the statistical association of pedophilia and childhood head injury.

Clarifying the role of head injury is important for understanding the relation between neurodevelopment and pedophilia, even if that clarification means eliminating head injury as a possible cause of pedophilia. It should be noted that useful research in this area will require groups of substantial size.