**5. Method**

#### **5.1 Sample**

Seventy-five subjects, reporting CSA volunteered in our research. The total group consisted of seventeen men and fifty-eight women, ranging from sixteen to sixty-one years of age (M age males = 37.31; M age females = 37.83).

### **5.2 Procedure**

Respondents were recruited through newspaper announcements and through contact persons, working in CSA self-help groups supervised by clinical psychologists, where group members were encouraged to both share and cope with their traumatic past by receiving support and understanding from their fellow group members. It should be noted that it was much harder to find male than female subjects. Furthermore the data of the male subjects were collected at another location and time. Hence the sessions consisted of either males only or females only. Before entering the first session a test battery was administered to each gender group consisting of a Checklist for CSA Characteristics, the BVAQ, the DES, and the LSQ.

#### **5.3 Instruments**

*Checklist for Childhood Sexual Abuse Characteristics*, specially developed for this occasion, mainly based on the results and insights derived from in-depth interviews taken by Ensink (1992) and Albach (1993) for their PhD research.

The checklist contains 8 categories:

*1) Perpetrator* (a. father; b. brother or sister; c. other relative (grandfather, uncle, etc.); d. family friend; e. stranger; f. familiar persons other than relatives or family friends); *2) Onset before the age of 10; 3) Duration; 4) Frequency; 5) Assault* (a. sexual harassment; b. masturbation; c. attempted penetration; d. penetration; e. fellatio; f. anal penetration; 6) *Coercion* (a. position of power; b. misleading; c. presents; d. threatening; 7) *Enforcing secrecy* (a. shame on you; b. emotional pressure; c. physical violence); 8) *Memory recovery* (a. having experienced an episode of inability to recall the traumatic event; b. number of years)

If otherwise mentioned all items were coded in a yes/no format. In the category perpetrator for instance the number of perpetrators could be indicated in the subcategory

stems from their perception that the world is a dangerous and threatening place (Feiring et al., 1999). The above seems to be confirmed by case studies, where male victims of sexual abuse were found to have low self-esteem as well (Myers, 1989). Expressions were as being flawed, and shame and guilt for participating in the sexual activities were reported as well, which later may become personality traits. Hunter (1990) reports significantly lower selfesteem scores for male victims of sexual abuse as opposed to non-victims. Dhaliwal et al.

In contrast with the above studies where gender comparisons are more relative than absolute, in the present research possible differences in symptom reporting between the

Seventy-five subjects, reporting CSA volunteered in our research. The total group consisted of seventeen men and fifty-eight women, ranging from sixteen to sixty-one years of age (M

Respondents were recruited through newspaper announcements and through contact persons, working in CSA self-help groups supervised by clinical psychologists, where group members were encouraged to both share and cope with their traumatic past by receiving support and understanding from their fellow group members. It should be noted that it was much harder to find male than female subjects. Furthermore the data of the male subjects were collected at another location and time. Hence the sessions consisted of either males only or females only. Before entering the first session a test battery was administered to each gender group consisting of a Checklist for CSA Characteristics, the BVAQ, the DES, and the

*Checklist for Childhood Sexual Abuse Characteristics*, specially developed for this occasion, mainly based on the results and insights derived from in-depth interviews taken by Ensink

*1) Perpetrator* (a. father; b. brother or sister; c. other relative (grandfather, uncle, etc.); d. family friend; e. stranger; f. familiar persons other than relatives or family friends); *2) Onset before the age of 10; 3) Duration; 4) Frequency; 5) Assault* (a. sexual harassment; b. masturbation; c. attempted penetration; d. penetration; e. fellatio; f. anal penetration; 6) *Coercion* (a. position of power; b. misleading; c. presents; d. threatening; 7) *Enforcing secrecy* (a. shame on you; b. emotional pressure; c. physical violence); 8) *Memory recovery* (a. having experienced

If otherwise mentioned all items were coded in a yes/no format. In the category perpetrator for instance the number of perpetrators could be indicated in the subcategory

an episode of inability to recall the traumatic event; b. number of years)

(1996) report reduced sexual self-esteem as a consequence of sexual abuse.

sexes will be tested using the same constructs and the same instruments.

**5. Method 5.1 Sample** 

**5.2 Procedure** 

LSQ.

**5.3 Instruments** 

age males = 37.31; M age females = 37.83).

(1992) and Albach (1993) for their PhD research.

The checklist contains 8 categories:

strangers. Duration related to the numbers of years the abuse had lasted and frequency related to how many times a week the abuse had taken place. In the category coercion misleading meant that the child was told it was perfectly normal for children to be involved in the kind of sexual activities the perpetrator demanded. Giving presents to the victims supposedly was a strategy to consolidate the bond between perpetrator and child and thereby avoiding betrayal. Threatening meant that the child would no longer benefit from all kind of privileges if it refused to cooperate. Giving the child the idea that he/she was responsible for the abuse by seducing the perpetrator was applied to ensure the abuse was kept a secret. The child was bad and should be ashamed. Exerting emotional pressure happened when the perpetrator convinced the child that he acted out of love and that revealing "our secret" would make the perpetrator very sad and would ruin his life. Physical pressure was another way of preventing the child from revealing the abuse. Finally memory recovery was investigated by asking the subjects whether he/she had ever experienced an episode of inability to recall the traumatic events and if so how many years this period lasted.

*Bermond-Vorst Alexithymia Questionnaire* (BVAQ, Bermond & Vorst, 1993), consisting of five subscales (Vorst & Bermond, 2001) which relate to a reduced ability to: a) verbalize emotional experiences (Verbalizing), b) differentiate between emotional feelings (Identifying), c) reflect upon emotions (Analyzing), d) fantasize (Fantasizing), and e) experience emotional feelings (Emotionalizing).

Reliability and validity of the BVAQ are good (Vorst & Bermond, 2001; Bermond et al., 2007). Assessing the prevalence of alexithymia is based on impairment of the cognitive component (sum total of reduced Verbalizing, Identifying, and Analyzing), as an impaired cognitive component (belonging to Type I and Type II alexithymia) appeared to be related to a broad range of psychological disorders (Moormann et al., 2008; see also the 2007 publication of Bailey & Henry for the relation between Type I & II alexithymia and somatization), while an impaired affective component (sum total of reduced Emotionalizing & Fantasizing) primarily related to activities where expressive behaviour and creative imagination were hampered. For both alexithymia components scores ≥ percentile 70 were used as an indication of impairment while scores ≤ percentile 30 were used as an indication of unimpairement. In the Moormann et al. study of 2008 the cut-off scores for the alexithymia types (based on a student population of 354 subjects) were computed as follows: raw scores ≥54.0 related to an impaired cognitive component while raw scores ≤43.0 related to an unimpaired cognitive component. Regarding the affective component raw scores ≥42.0 related to an impaired affective, while raw scores ≤34.0 related to an unimpaired affective component. Using the above cut-off scores for an impaired cognitive BVAQ component as an indication of the prevalence of alexithymia resulted in an incidence of alexithymia of 20.4% in the Netherlands (Moormann et al., 2008) and of 18.2% in Australia (Bailey & Henry, 2007). On the basis of an albeit small clinical sample of 39, Taylor et al. (1997) suggest a cutoff value of 61 and above on the Toronto Alexithymia Scale (TAS-20). Despite the different instrument for alexithymia the statistics given in Finland (21.0%; Honkolampi et al., 2000), using these TAS-20 cut-off scores, are practically the same as the statistics obtained with the BVAQ in Australia and The Netherlands. Older studies using a cut-off on the earlier 26-item TAS have reported rates of 23% for a normal group in France (Loas et al., 1995) and 18.8% in Canada (Parker et al., 1989). Hence, the cut-off scores used in the present research seem a fair indication for the prevalence of alexithymia.

Gender Differences in the Impact of Child Sexual Abuse on Alexithymia, Dissociation and Self 83

Regarding the category perpetrator only t-tests where the father had been the perpetrator (29% for males, 40% for females) could be computed as in all other cases one of the groups contained empty cells: 18% of the males reported abuse by brother or sister, none of the females; No males reported abuse by other relatives (grandfather, uncle, etc.), while 24% of the females did; No males reported abuse by a family friend, while 6% of the females did; No males reported abuse by strangers, while the average numbers of abusive strangers was 1.12 for females; however 47% of the males reported abuse by people they knew other than

relatives or a family friend, whereas none of the females did.

Table 1. Gender differences in abuse characteristics and age

Because of empty cells only the scores for the father were given in Table 1. None of the differences reached significance. The elevated percentages suggest a severe form of CSA.

*Dissociative Experiences Scale* (DES – Bernstein & Putnam, 1986), a widely used self-report scale, containing 28 items, using a visual analogue scale (from 0-100). A Dutch adaptation was administered, consisting of the 3 subscales, as defined by Ross et al. (1995): a) Absorption-imaginative involvement ("So involved in fantasy that it seems real"), b) Activities of dissociated states ("Finding oneself in a place, but unaware how one got there"), and c) Depersonalization / Derealization ("Not recognizing one's reflection in a mirror; Other people and objects do not seem real")

Summing up all the scores and then dividing them by 28 results in the average DES score. The test retest reliability is good (0.84-0.96) and effectively differentiates patients with Dissociation from other psychiatric groups (Bernstein & Putnam, 1986). The DES can be used as a screening instrument for Dissociative Identity Disorder (DID). However semistructured interviews are needed for a reliable diagnosis. Several cut-off values are used, ranging from 20 (Ross et al., 1991) to 31.3 (Bernstein & Putnam, 1986). In The Netherlands the DES cut-off scores were validated using the outcomes from clinical diagnostic interviews (Boon & Draijer, 1993), which resulted in an optimal cut-off score of 25 for the screening of DID. The mean DES scores in normal populations vary from 3.7 to 7.8, while scores of a group of a-select psychiatric patients vary from 14.6 to 17.0 (in Boon & Draijer, 1993).

*Leiden Self Concept Questionnaire* (LSQ ; Moormann & Duikers, 1984). The items relate to intellectual functioning, social functioning, physical appearance and self-efficacy. It consists of 27 items, first formulated in the present tense (present self), then in the past tense (past self), and finally in the future tense (future self). Only the present self will be discussed in our study. The Cronbach alphas are based on the 1989 publication by Moormann et al., and are given for the sexes separately (male: = .89; female: = .87). Scores ≤ percentile 30 is considered to be indicative of a low self (cut-off score of ≤ 103 for a student population (based on the data of the Moormann et al. study of 2008), cut-off score of ≤ 80 for male and ≤ 63 for female poly hard drug addicts (based on the data of the Moormann et al. study of 1989, and the Bauer et al. study of 1992), and cut-off scores ≤ 98 for the elderly (74-98 years of age), discussed in the data of the Moormann, et al. study of 1997.

#### **5.4 Statistical analyses**

First *Independent-Samples t Tests* will be run on all abuse characteristics (dependent variable) to identify whether the sexes (independent variable) differ on these variables. If so then in all analyses where means are compared it will be controlled for those abuse characteristics where sexes differ.

Regarding **Hypothesis 1** it was decided to use *cut-off scores* to identify the prevalence of alexithymia, dissociation and low self in the CSA group.

**Gender differences** will be tested by again performing *Independent-Samples t* Tests with gender as independent and alexithymia, dissociation and self as dependent variable.

### **6. Results**

#### **6.1 Gender differences in abuse characteristics and age**

The results are given in Table 1 and demonstrate that the sexes didn't differ significantly neither on any of the abuse characteristics nor on age. However there was one trend on anal penetration that was more prominent in male than in female subjects.

*Dissociative Experiences Scale* (DES – Bernstein & Putnam, 1986), a widely used self-report scale, containing 28 items, using a visual analogue scale (from 0-100). A Dutch adaptation was administered, consisting of the 3 subscales, as defined by Ross et al. (1995): a) Absorption-imaginative involvement ("So involved in fantasy that it seems real"), b) Activities of dissociated states ("Finding oneself in a place, but unaware how one got there"), and c) Depersonalization / Derealization ("Not recognizing one's reflection in a

Summing up all the scores and then dividing them by 28 results in the average DES score. The test retest reliability is good (0.84-0.96) and effectively differentiates patients with Dissociation from other psychiatric groups (Bernstein & Putnam, 1986). The DES can be used as a screening instrument for Dissociative Identity Disorder (DID). However semistructured interviews are needed for a reliable diagnosis. Several cut-off values are used, ranging from 20 (Ross et al., 1991) to 31.3 (Bernstein & Putnam, 1986). In The Netherlands the DES cut-off scores were validated using the outcomes from clinical diagnostic interviews (Boon & Draijer, 1993), which resulted in an optimal cut-off score of 25 for the screening of DID. The mean DES scores in normal populations vary from 3.7 to 7.8, while scores of a

group of a-select psychiatric patients vary from 14.6 to 17.0 (in Boon & Draijer, 1993).

of age), discussed in the data of the Moormann, et al. study of 1997.

alexithymia, dissociation and low self in the CSA group.

**6.1 Gender differences in abuse characteristics and age** 

penetration that was more prominent in male than in female subjects.

**5.4 Statistical analyses** 

where sexes differ.

**6. Results** 

*Leiden Self Concept Questionnaire* (LSQ ; Moormann & Duikers, 1984). The items relate to intellectual functioning, social functioning, physical appearance and self-efficacy. It consists of 27 items, first formulated in the present tense (present self), then in the past tense (past self), and finally in the future tense (future self). Only the present self will be discussed in our study. The Cronbach alphas are based on the 1989 publication by Moormann et al., and are given for the sexes separately (male: = .89; female: = .87). Scores ≤ percentile 30 is considered to be indicative of a low self (cut-off score of ≤ 103 for a student population (based on the data of the Moormann et al. study of 2008), cut-off score of ≤ 80 for male and ≤ 63 for female poly hard drug addicts (based on the data of the Moormann et al. study of 1989, and the Bauer et al. study of 1992), and cut-off scores ≤ 98 for the elderly (74-98 years

First *Independent-Samples t Tests* will be run on all abuse characteristics (dependent variable) to identify whether the sexes (independent variable) differ on these variables. If so then in all analyses where means are compared it will be controlled for those abuse characteristics

Regarding **Hypothesis 1** it was decided to use *cut-off scores* to identify the prevalence of

**Gender differences** will be tested by again performing *Independent-Samples t* Tests with

The results are given in Table 1 and demonstrate that the sexes didn't differ significantly neither on any of the abuse characteristics nor on age. However there was one trend on anal

gender as independent and alexithymia, dissociation and self as dependent variable.

mirror; Other people and objects do not seem real")

Regarding the category perpetrator only t-tests where the father had been the perpetrator (29% for males, 40% for females) could be computed as in all other cases one of the groups contained empty cells: 18% of the males reported abuse by brother or sister, none of the females; No males reported abuse by other relatives (grandfather, uncle, etc.), while 24% of the females did; No males reported abuse by a family friend, while 6% of the females did; No males reported abuse by strangers, while the average numbers of abusive strangers was 1.12 for females; however 47% of the males reported abuse by people they knew other than relatives or a family friend, whereas none of the females did.


Table 1. Gender differences in abuse characteristics and age

Because of empty cells only the scores for the father were given in Table 1. None of the differences reached significance. The elevated percentages suggest a severe form of CSA.

Gender Differences in the Impact of Child Sexual Abuse on Alexithymia, Dissociation and Self 85

Table 2. Differences between male and female respondents on alexithymia, dissociation and

Females scored significantly more alexithymic than men on the cognitive component, particularly on Identifying and Verbalizing emotions (on Analyzing the sexes did not differ significantly). Neither on the affective component nor on its subscales Emotionalizing and Fantasizing gender differences was significant. With the exception of the subscale 'Activities of Dissociated States', also called Dissociative amnesia, women scored significantly higher on all dissociation measures as well. Furthermore CSA women had a significantly lower self-concept. From the above it is concluded that that the impact of CSA is greater for

Subjects in our sample have been victims of severe CSA and the results on alexithymia, dissociation, self and memory recovery confirm the notion that severe CSA has a devastating effect on psychological health in adulthood. The incidence of alexithymia (40%) was found to be substantially higher than the statistics published in non-abused populations in various different countries (around 20%; see the BVAQ description in the instruments part). Furthermore it was found that 47% of males and 78% of the females had average DES scores > 25, indicating DID. Finally CSA respondents reported a negative self. The outcomes above can be criticized on a common problem with prevalence indices for psychological disorders, e.g. where the exact cut off value should be located to ensure a valid diagnosis of alexithymia, dissociation or self. To some extent cut off scores always remain arbitrary and subjective. Even when sensitivity and specificity requirements are fulfilled the choice of the criterion remains an arbitrary decision. Furthermore the reliability on the reported abuse characteristics, a common problem in all retrospective studies using self-report

self (\*: p<0.05, \*\*: p<0.01, \*\*\*: p<0.001)

**7.1 Impact of CSA on alexithymia, dissociation and self** 

questionnaires, can be questioned as well.

women than for men.

**7. Discussion** 

The average abuse lasted 3-4 years (ranging from 1-30 years), and happened 3-4 times a week. Furthermore the prevalence of the various assaults (ranging from 16% to 100%) indicates severe abuse when considering:


The finding that 71% of the males and 78% of the females reported that the abuse had started before the age of 10 suggests that we are really dealing with *childhood* sexual abuse.

Because neither on age nor on any abuse characteristic significant gender differences could be demonstrated, there was no further need for controlling for these variables in the upcoming statistical analyses where gender differences were tested on alexithymia, dissociation and self.

#### **6.2 Impact of CSA on alexithymia, dissociation and self**

Considerable evidence was found for **Hypothesis 1**, stating that CSA-victims display:

A *high prevalence of alexithymia* (e.g. scores ≥ 54.0 on the cognitive component of the BVAQ), as 40% of the CSA subjects (see Table 1 in the next chapter by Moormann, Albach & Bermond on alexithymia, dissociation and memory recovery) could be classified as either Type I (14%) or Type II (26%) alexithymia compared with the common guideline that about 1/5th of the normal population is alexithymic (see Instruments for prevalence statistics).

*High dissociation scores* (DES/28 > 25), as the average DES scores of both male (M = 25.79) and female (M = 38.23) CSA respondents exceeded the optimal cut-off value of 25 (see Table 2), indicative of Dissociative Identity Disorder (DID). Regarding the incidence of severe dissociation it was found that 47% of male and 78% of the females had average DES scores > 25

A *negative Self* as the average Self scores for male (M = 88.39) and female (M = 71.06) CSA respondents (see Table 2) were far below the cut-off value for students (N = 1061) corresponding to ≤ percentile 30 (Mean LSQ ≤ 103). Unfortunately norms for the normal population were not yet available. However compared with the elderly (N = 30; 74-98 years of age) the scores of both male and female CSA respondents were also far below the cut-off value of 98, corresponding to values lower than percentile 25. Compared with poly hard drug addicts, known for a negative self, the average mean self scores of both male and female CSA respondents fell in stanine 4 (for norm references see Instruments).

#### **6.3 Explorative part: Gender differences**

In Table 2 the statistics for this paragraph are given.


Table 2. Differences between male and female respondents on alexithymia, dissociation and self (\*: p<0.05, \*\*: p<0.01, \*\*\*: p<0.001)

Females scored significantly more alexithymic than men on the cognitive component, particularly on Identifying and Verbalizing emotions (on Analyzing the sexes did not differ significantly). Neither on the affective component nor on its subscales Emotionalizing and Fantasizing gender differences was significant. With the exception of the subscale 'Activities of Dissociated States', also called Dissociative amnesia, women scored significantly higher on all dissociation measures as well. Furthermore CSA women had a significantly lower self-concept. From the above it is concluded that that the impact of CSA is greater for women than for men.
