The psychology of kink: a survey study into the relationships of trauma and attachment

style with BDSM interests

S Ten Brink1,Msc, V Coppens1, 2,Msc, PhD, W Huys1,Msc, M Morrens1, 2, MD, PhD.

1
Collaborative Antwerp Psychiatric Research Institute, Faculty of Medicine
and Health Sciences, University of Antwerp, R3.22, Building R, Campus Drie Eiken,
Universiteitsplein 1, 2610, Antwerp, Belgium;

2
University Psychiatric Hospital Duffel, Stationsstraat 22c, 2570 Duffel, Belgium;

Abstract

Introduction: BDSM is a prevalent type of sexual expression that refers to Bondage

and Discipline (BD), Dominance and Submission (DS), and Sadism and Masochism (SM). It

has been speculated that a substantial percentage of BDSM practitioners have experienced

(sexual) trauma in the past. Attachment style is an additional factor resulting from early life

dynamics that has been suggested to potentially influence BDSM interests. This study will

investigate to what extent BDSM interests are related to trauma and attachment style, while

differentiating between BDSM community practitioners and private practitioners.

Methods: A group of 771 BDSM-practitioners and 518 non-practitioners from the

general population completed a survey in 2017 assessing BDSM interests as well as the Brief

Trauma Questionnaire to gauge traumatic events and the Relationships Questionnaire to

assess attachment style.

Results: Community BDSM-practitioners and private practitioners reported higher

levels of physical abuse in adulthood but no significant differences emerged for other

traumatic experiences including childhood physical abuse or unwanted sexual trauma.

Surprisingly, BDSM-practitioners had more secure and at the same time more anxious-

preoccupied attachment styles compared to non-practitioners. Besides, secure attachment style

 1

was associated with dominance, whereas the anxious-avoidant attachment style was

associated with submissiveness. Intensity of BDSM interest was predicted by secure

attachment style, gender, sexual orientation and living area.

Discussion: Thus, our findings do not support the hypothesis of BDSM being a

maladaptive coping mechanism in response to early life dynamics.

Social policy implications: BDSM practices deserve perception as normal sexual

practice free from stigmatization rather than deviant behavior.

Key words: attachment, BDSM, kink, sadomasochism, trauma

 2

Introduction

BDSM is a type of sexual expression that refers to Bondage and Discipline (BD),

Dominance and Submission (DS), and Sadism and Masochism (SM). With mutual consent,

sex or play partners typically make use of imposed physical restraints, assignments and (real

or played) punishments (BD) and have or implement fantasies about power relationships and

power roleplay (DS). They hereby enjoy causing or experiencing intense sensory stimuli like

pain (SM). More often than not, practitioners can have one of the 3 specific power exchange

roles during BDSM interactions: Dom – the person who is in control and assumes the

dominant role, Sub – the person who gives up control and assumes the submissive role or

Switch, where the practitioner alternates between dominant and submissive roles depending

on the kind of practices and situations (Wismeijer & Assen, 2013).

Previously, the prevalence of BDSM in the general population has been reported to lie

between 2 and 65%; a broad range presumably due to use of differing definitions of BDSM

and other methodological aspects (Renaud & Byers, 1999; Richters et al 2008; Masters,

Johnson & Kolodny 1995; De Neef et al., 2019). Recently, our group demonstrated in a large

sample (n=1027) that 46.8% of the general population had ever performed at least one

BDSM-related activity, and an additional 22% indicated having (had) fantasies about it

(Holvoet et al., 2017). About 10% engaged in these activities on a regular basis. These high

prevalence rates show that at least certain levels of BDSM-interest are present within a

majority of the general population.

Surprisingly however, although experienced by almost half the general population,

sadistic and masochistic interests and behaviors are still categorized as paraphilic disorders

within the Diagnostic and Statistical Manual of Mental Disorders, 5 th edition (DSM-V).

Inevitably, this medicalization of BDSM activities propagates strong stigmatization of BDSM

practitioners, recurrently inducing feelings of shame and guilt and consequential

 3

psychological problems or even suicide ideations (Roush, et al., 2016).

With regard to the origination of these sexual interests, it has been speculated that a

substantial percentage of BDSM practitioners have experienced (sexual) trauma in the past

(e.g. Bekes, Perry, & Robertson, 2017). Indeed, some authors frame BDSM interests as

maladaptive coping mechanisms resulting from early life events. Finkelhor & Browne (1985)

argued that sexual abuse induced dysfunctional development of a child’s sexuality (sexual

feelings and attitudes), leading to the creation of unusual emotional and cognitive

representations of sexual activities. Freud labeled these “inappropriate” sexualized behaviors

as a repetition compulsion (Freud 1905), a psychoanalytic phenomenon that describes a

pattern whereby people endlessly repeat patterns of behavior (like reenacting an event or

putting oneself in situations where the event is likely to happen) which were repressed and

traumatic in their childhood (Buckingham, 2002). This concept could explain the link

between early sexual abuse and “inappropriate” sexual behaviors later in life. Alternatively,

Califia (1983) postulated that masochism or submission potentially function as healing

processes for remedying old wounds.

An association between trauma and masochism was indeed demonstrated by Frias and

colleagues (2017), who showed that masochistic women with borderline personality disorder

had higher childhood sexual abuse rates compared to their non-masochistic peers. This is also

in line with higher prevalence rates of self-reported sexual abuse in BDSM club members

(Nordling et al., 2000). Notwithstanding, only 9.6% of the SM club members in this study

actually reported childhood sexual abuse, which is a small minority. In addition, it has been

the only study that has found this difference between practitioners and the general population.

The Australian Study of Health and Relationships (ASHR) found contrary results in a national

representative sample wherein psychological distress and sexual functioning was examined.

This study found that 2% of sexually active men and 1.4% of sexually active women had
 4

engaged in BDSM activities within the past year, and found no difference in past sexual abuse

history, or levels of distress compared to other members of the general population (Richters et

al, 2008). Nevertheless, in a broader context, some research did demonstrate associations

between trauma and sexual behavior in general. For example, Meston, Herman & Trapnell

(1999) found a relationship between early sexual abuse and adult sexual behavior like

frequency of masturbation, range of sexual fantasies, masochism, promiscuity, and

voyeurism. Later studies also found associations between trauma and health-risking sexual

behaviors in girls (Smith et al., 2006) and gay men (Kalichman et al., 2004). As a result, these

studies may suggest that (early) trauma can affect certain sexual behaviors later in life,

however the connection between trauma and BDSM practices asks for more research-based

elaboration. Furthermore, the actual relationship between BDSM interests and the adoption of

specific BDSM-roles (dominant/submissive/switch) and trauma in the general BDSM

community has not been investigated.

Attachment style is an additional factor resulting from early life dynamics that

potentially may influence BDSM interests, as it has been demonstrated that attachment style

could reliably predict some sexual attitudes and behaviors, with especially an anxious

attachment style seemingly predicting drug use during sexual contact, unsafe sex, and

negative attitudes towards condoms, in late adolescence (Feeney et al., 2000). Attachment is

defined as a lasting relationship between a person and the people around them with whom

they regularly interact. This emotional bond lays the foundations of identity formation, self-

regulation/self-dependence and interpersonal beliefs and behaviors later in life. Bowlby’s

attachment theory states that a good development of the emotional bond necessitates

sensitivity and reactiveness to the child (Bowlby, 1977). Bartholomew and Horowitz (1991)

defined four attachment styles for adults: 1. secure attachment style; 2. anxious-preoccupied

attachment style; 3. avoidant attachment style; and 4. anxious-avoidant (disorganized)

 5

attachment style. Furthermore, Gentzler and Kerns (2004) found that an avoidant attachment

style was associated with practices of casual sex while an anxious attachment style was more

related to practices of unwanted but consensual sex. Similarly, Szielasko et al. (2013) found

that an avoidant attachment style was associated to a higher lifetime number of sexual

partners while ambivalence attachment predicted invasive and coercive sexual behaviors.

Only one study (Wismeijer & Van Assen, 2013) looked into the relationship between

attachment styles and BDSM preferences. They found no support for the persistent

assumption that BDSM is associated with maladaptive attachment processes. It should be

noted that the few studies looking into the psychological dynamics of BDSM interests

typically include participants recruited from BDSM-themed internet forums or from BDSM-

clubs. Nevertheless, about 85% of the BDSM practitioners only engage in these sexual

activities in the privacy of their own homes (Holvoet et al., 2017). As such, research until now

has suffered from a strong selection bias. It may be rewarding to differentiate between

practitioners recruited from the BDSM community and private BDSM practitioners. The

present study will thus investigate to what extent BDSM interests are related to trauma and

attachment style in BDSM community practitioners and private practitioners.

METHODS

Study Design

Data were gathered between February 2017 and March 2017 by means of an online

questionnaire, querying 1) demographic variables, 2) 54 items about interest and practice

experience with specific BDSM activities, 3) context of BDSM activities, 4) trauma history

and 5) attachment styles. For a detailed description of the survey’s structure and content, see

Holvoet et al. 2017. This survey was distributed amongst the general population via iVOX, a

market research and polling agency with access to a panel of 150,000 Belgian citizens
 6

representative of the general population (www.ivox.be; n=1027), and within the online

BDSM community (through BDSM-specific forums; n=251) to which the survey was

presented using Survey Monkey (www.surveymonkey.com). The duration time to complete

the survey was about 15 minutes. The study was approved by the ethical committee board of

UZA/UA (University (Hospital) of Antwerp).

Participants

Survey completers (n=1289) were subdivided into 4 BDSM interest level groups: “No

interest” (NI, scoring ≤ 3 on all items; n=326); “Fantasy” (BDSM-F; score either 4 or 5 on

any item; n=192) and 2 BDSM Practice groups (score of ≥ 6 on any item; n=771). The BDSM

Practice group was further subdivided in a BDSM-Private Practice (BDSM-PP; n=559) and a

BDSM-Community Practice (BDSM-CP; n=212) group based on whether practitioners

performed their activities solely at home (negative response on all BDSM community event

attendance items; BDSM-PP) or within an existing BDSM Community (positive response on

at least 1 BDSM community event attendance item; BDSM-CP).

Participants in the NI and BDSM-F groups were all volunteers from a general

population study in Flanders. Participants from the 2 BDSM Practice groups were either

recruited from the general population or via Fetlife (an online BDSM community website),

four participating BDSM-associations (Kajira, VZW Steel Moon, Fetish Café, Club 78), as

well as other local online BDSM forums.

Materials

The questionnaire was developed in collaboration with several Belgian BDSM

organizations: Kajira Gent (student association Ugent for students with an interest in BDSM),

VZW Steel Moon (Antwerp), Fetish Café (Antwerp), and Club 78 (Oostham). Interests and
 7

experiences with BDSM related activities, fetish-related activities and general socio-

economical factors like age, gender, study level, etc. were surveyed (see Holvoet, et al., 2017

for more detailed description about the questionnaire).

The Brief Trauma Questionnaire

The Brief Trauma Questionnaire (BTQ) (Schnurr et al., 2002) is a validated 10-item

self-report questionnaire derived from the Brief Trauma Interview (BTI; Schnurr, et al., 1995)

to scrutinize the presence and severity of past trauma. The first seven items, respectively

items on 1) serving in a war zone, 2) serious car accident, 3) major natural/technological

disaster, 4) life-threatening illness, 5) physical punishment before the age of 18 by parent,

caretaker or teacher, 6) other physical violence and 7) unwanted sexual behavior towards the

participant were included in the analyses. It should be noted with item 6 that participants were

asked about if they were ever attacked, beaten, or robbed after the age of 18 by whoever,

including friends, family and strangers. No specification was made in the questionnaire

whether this was consensual or not. Severity of the specific traumas was measured by two

additional yes or no questions.

The Relationships questionnaire

For evaluation of attachment styles, a translated Dutch version of the Relationships

Questionnaire (RQ) was used (Bartholomew & Horowitz, 1991). This self-report screening

instrument is designed to obtain continuous ratings of each of the following four attachment

patterns: (i) secure attachment style, (ii) anxious-avoidant attachment style, (iii) anxious-

preoccupied attachment style, and (iv) dismissive-avoidant attachment style.

Statistical analysis
 8

Statistical analyses were performed using SPSS v22.0. Multiple group comparisons

with normally distributed data were done by the multivariate analysis of (co)variance method

(Pillai’s trace). For not normally distributed data the Kruskal-Wallis test was used. Analysis

of variance tests and Mann-Whitney tests were applied for single group comparisons

depending on whether the data was normally distributed or not. Contingency analyses (Chi-

square test) were used for group comparisons with nominal variables. Exploring associations

between variables was done by use of Pearson’s and Spearman Rho correlation methods,

depending on the normality of the data. Finally, a linear regression model (method=enter) was

applied to see which variables had the most predictive value for the overall outcome of

BDSM interests.

RESULTS

Demographics

In total, 1440 individuals responded to the survey; of these, 1289 completed the survey with

251 completers recruited via online platforms and 1038 completers from the general

population approached by iVox. Of the completers 326 individuals indicated having no

affinity with BDSM (no interest (NI) group), either on fantasy or on practice level and 15%

proclaimed having had BDSM related fantasies (n=192; BDSM-F group). Surprisingly,

almost 60% (n=771) of the completers have put at least 1 BDSM related activity into

practice, be it at home (BDSM private practice (BDSM-PP); 43% of our total sample) or

within a BDSM related community event (BDSM community practice (BDSM-CP); 16% of

our total sample). From the private practice group (BDSM-PP) 90,7% of the participants

(from the general population) completed the survey using iVox, and 9,3% completed the

survey with Survey Monkey. For the community practitioners group (BDSM-CP) 6,1% of its

participants made use of iVox, and 93,9% via Survey Monkey (See Table 1). For a detailed
 9

overview of the demographics, see Coppens et al. (2019).

-- INSERT TABLE 1 ABOUT HERE –

BDSM summary scores

Based on a principal component analysis with the 54 BDSM-items, four BDSM

components were created (see Holvoet et al., 2017): submissiveness (SUB; including items as

‘being hit’, ‘kneeling before partner’, or ‘movement restriction’), dominance (DOM;

including items as ‘hitting a partner’, ‘let partner address you with title’, ‘blindfolding

partner’), a component representing voyeurism/visual stimuli (VOY; including items as

‘watching people getting hit’ or ‘fire play’), and finally a component including items on the

use of attributes (ATT, including ‘use of medical attributes and ‘penetration using big

objects’). We refer to Holvoet et al (2017) for further information on these subscales.

For each component and their corresponding items, summary scores were made in

order to create a total score for each underlying sub-domain of BDSM. By summing up the

summary scores of the 4 BDSM components (DOM, SUB, VOY and ATT) a total score for

BDSM (BDSM_total) was created as well.

As expected, by use of multivariate analysis and Pillai’s trace, a significant group

effect for the 4 groups (V = .79, F (12, 3852) = 113.67, p < .001) was found for each of the

BDSM summary scores (TOT, SUB, DOM, VOY and ATT). Post-hoc Bonferroni analyses

showed that all group comparisons were significant (NI < BDSM-F < BDSM-PP

< BDSM-CP; all comparisons p ≤ .001) for each summary score (see figure

1).

--INSERT FIGURE 1 ABOUT HERE –

 10

Distribution of BDSM identities within the 2 BDSM Practice groups

With regard to distribution of BDSM identities (dominant/submissive/switch/other) within the

BDSM Practice groups (PP & CP), 299 participants from the 771 BDSM practitioners

indicated to have an affiliation with a certain BDSM identity type. The 472 remaining

participants indicated experience with one or more BDSM related practices, but did not self-

identify as BDSM practitioner and did not indicate having a certain BDSM identity type.

Among these BDSM identities, a majority self-identified as submissive (39,5%), with

dominant and switch identities more or less equally represented (28,8% and 31,1%

respectively). Only 2 participants (.7%) self-identified as ‘other’. Due to the heterogeneous

nature of this group as well as its small sample size, the BDSM identity type ‘other’ was not

included for further analyses (see also Coppens et al, in prep).

Chi-square test results showed a significant effect for the distribution of gender among the

three BDSM identity types (χ2 (2) = 47.32, p < .001). Of the men, 42,9% indicated being

dominant, whereas 27,1% indicated being submissive and 29,9% as switches. As such, there

were significantly more dominant men and less submissive men (standardized residual > ±2).

Among the female participants, a majority of 59,3% were submissive, another 32,2% self-

identified as switches and only 8,5% were dominant. For women the finding was reversed,

hence there were significantly more submissive and less dominant women than expected

(standardized residual > ±3).

Occurrence of trauma for the four groups (NI, BDSM-F, BDSM-PP, BDSM-CP)

Subjects were questioned whether they had experienced trauma related to war/combat zones

(WAR), car accidents (CAR), natural/nuclear disasters (NAT), physical illness/disease (DIS),
 11

physical beatings or attacks during childhood (PBC) and adulthood (PBA), or if they had

suffered unwanted sexual contact (USC) in their lifetime. See table 2 for the occurrence rates.

--INSERT TABLE 2 ABOUT HERE –

Comparing the overall trauma scores for each trauma item among all groups, Kruskal

Wallis tests showed significant differences for the trauma variables PBA (H(3) = 13.96, p

< .01), USC (H(3) = 8.67, p < .05) and the total trauma score (TOT trauma; H(3) = 13.15, p

< .01), but not for any of the other trauma scores (including PBC). Post-hoc analyses with

Mann-Whitney tests were done to follow-up this finding. A Bonferroni correction with a

critical value of .0083 was applied as well to report significant group differences. When

looking at the total trauma score (TOT trauma) post-hoc analyses revealed that community

practitioners of BDSM (BDSM-CP) had a higher total trauma score than participants in the NI

(U = 29678.50, r = -.13, p < .0083) and marginally significantly higher than BDSM-F group

(U = 17500.00, r = -.13, p = .009). As for Physical Beatings in Adulthood (PBA), controls

had experienced less physical violence at a later age (18,8%) than practitioners of BDSM in

private setting (BDSM-PP: 27,3%; U = 82934.50, r = -.10, p < .0083) as well as in

community/club setting (BDSM-CP: 32,7%; U = 29969.00, r = -.15, p < .001). Concerning

the trauma related to unwanted sexual contacts, no group differences could be found, although

club practitioners (BDSM-CP) tended to have experienced more sexual trauma (23,4%) than

non-practitioners with fantasies (BDSM-F; 12,4%; U = 18427.50, r = -.13, p = .011), but this

was not significant (given critical value of .0083). No other group comparisons for the trauma

variables PBA, USC, and TOT trauma were significant, see Figure 2.

--INSERT FIGURE 2 ABOUT HERE –

 12

Severity of trauma for the four groups (NI, BDSM-F, BDSM-PP, BDSM-CP)

For those individuals having reported presence of DIS and PBA trauma, additional

analyses by use of Kruskal-Wallis test were done on the severity of the trauma. A significant

group difference was only found for the DIS trauma (H(3) = 8.46, p < .05), but not for the

PBA trauma. Post-hoc analyses with Mann-Whitney tests were done to follow-up this finding.

A Bonferroni correction with a critical value of .0083 was applied as well to report significant

group differences. None of the separate group comparisons for the DIS trauma were

significant.

In conclusion, community practitioners of BDSM had a higher PBA (32,7% vs.

18,8%) and TOT trauma score than controls, but did not significantly differ on any of the

other trauma measures. Private practitioners only differed from controls on the trauma item

PBA (27,3% vs. 18,8%). Neither private nor community practitioners of BDSM did differ in

their degree of experienced trauma from the control group. Despite the significant results, a

couple of things should be taken into consideration. First, the effect sizes were small (-.30 > r

< .30). Secondly, because no specification was made whether the reported PBA trauma was

consensual or not, a possible confounding might be present in the significant results between

practitioners and controls: practitioners of BDSM could have answered the PBA trauma

question with ‘yes’ when actually thinking about beatings during BDSM play.

Associations between BDSM and trauma scores

Associations between trauma and BDSM scores were calculated for the BDSM

practice groups (BDSM-PP & BDSM-CP). Within the BDSM-CP group, the TOT BDSM-

score was significantly related to the unwanted sexual contact trauma item (USC: rs = .25, p <

.001). Exploratory follow up analyses showed that specifically the BDSM component
 13

submissiveness (SUB) was significantly but mildly associated with USC trauma (rs = .20, p

< .01). No other associations between BDSM- and trauma scores for the practitioners groups

were found.

Within the BDSM-CP group USC trauma was more reported by women (38.2%) than

men (11%) and this difference was significant (U = 3823.50, z = -4.61, p < .001). Other

findings were that bi- and pansexuals experienced more USC trauma than heterosexuals (U =

2946, z = -2.31, p < .05 and U = 618, z = -2.37, p < .05 respectively); bi- / pan- / asexuals

more than homosexuals (U = 157.50, z = -2.06, p < .05 / U = 31.50, z = -2.34, p < .05 / U =

4.50, z = -2.12, p < .05); people with the lowest education level more than averagely or highly

educated individuals (U = 236.50, z = -3.33, p = .001 and U = 531, z = -3.56, p < .001

respectively); and submissives and switches more than dominants (U = 1764.50, z = -2.47, p <

.05 and U = 1176, z = -3.92, p < .001).

BDSM practitioners have a more secure attachment style than non-practitioners

To investigate the link between BDSM interests/practices and attachment style for the

four groups, each subject’s most prominent attachment style was determined by means of the

Relationships Questionnaire (RQ).

Multivariate analysis of variance showed a significant effect of group condition on the

overall attachment styles (V = .06, F(12, 3729) = 6.39, p < .001). Specifically, the test of

between-subjects effects showed that the four groups significantly differed from each other

for the secure (F(3) = 12.44, p < .001), anxious-avoidant (F(3) = 4.19, p < .01), and the

anxious-preoccupied attachment style (F(3) = 8.11, p < .001). No significant group effect was

found for the dismissive-avoidant attachment style. Post-hoc Bonferroni analyses were done

for comparing the groups on each of the four attachment styles, see Figure 3.
 14

--INSERT FIGURE 3 ABOUT HERE –

Surprisingly, participants in the BDSM-CP group had a more secure attachment style

than participants in the NI group (p < .001, SE: .14), BDSM-F group (p < .001, SE: .16) and

the BDSM-PP group (p < .001, SE: .13). Additionally, these community practitioners of

BDSM also had a less anxious-avoidant attachment style than participants from the groups,

BDSM-F (p < .05, SE: .18) and BDSM-PP (p < .05, SE: .14). For the anxious-preoccupied

attachment style findings were different. Here, participants from the BDSM-CP group had a

significantly more anxious-preoccupied attachment style than the participants from the NI

group (p < .001, SE: .14) and BDSM-PP group (p < .05, SE: .13). Also, participants from the

BDSM-F group were significantly more anxious-preoccupied than participants from the NI

group (p < .05, SE: .15). Private practitioners (BDSM-PP) did not differ in any of the four

attachment styles from controls.

Impact of gender and BDSM identity on attachment style

Gender also seemed to have an overall significant effect on all four attachment styles

(MANCOVA - Pillai’s Trace: V = .03, F (4, 1237) = 10.88, p < .001). The test of between-

subjects showed that gender had independent significant effects for the secure (F(1, 1240) =

4.40, p < .05), anxious-avoidant (F(1, 1240) = 11.53, p = .001), anxious-preoccupied (F(1,

1240) = 9.63, p < .01), and dismissive-avoidant (F(1, 1240) = 8.77, p < .01) attachment style.

Within the total sample, male participants had a more secure (F(1, 1243) = 6.94, p < .01),

anxious-preoccupied (F(1, 1243) = 13.33, p < .001) and dismissive-avoidant (F(1, 1243) =

9.13, p < .01) attachment style, and a less anxious-avoidant (F(1, 1243) = 11.73, p = .001)

attachment style than female participants. On group level this difference between men and

women in attachment style was only found in the BDSM practice groups and not in the NI
 15

and BDSM-F group. When controlling for gender the group effect on all four attachment

styles hardly changed and significant differences remained for the secure (p < .001), anxious-

avoidant (p < .01), and anxious-preoccupied (p < .001) attachment style. The gender “other”

was not included in the previous analyses.

Associations between BDSM and attachment styles

Looking for associations between BDSM-scores and attachment styles by use of

Pearson’s correlation method, positive significant relations between secure attachment style

and the BDSM factors, DOM (r = .206, p < .05) and VOY (r = .211, p < .05) were found

within the BDSM-CP group. High scores on dominant practices and visual stimuli during

BDSM play tended to be associated with a more secure attachment style for participants who

practiced BDSM in clubs. Another positive association within the BDSM-CP group was

found for the anxious-avoidant attachment style and the SUB factor (r = .200, p < .01) of

BDSM. As such, participants with a more anxious-avoidant attachment style tended to

execute more submissive practices.

BDSM identities and attachment styles

Using Pillai’s Trace, the multivariate analysis of variance test showed an overall

significant group effect of BDSM identity for the attachment styles (V = .11, F(8, 576) = 4.36,

p < .001) within the subgroup of BDSM practitioners (BDSM-PP; BDSM-CP). The findings

were that participants significantly differed in terms of their secure attachment style (F(2,

290) = 4.66, p = .01) and anxious-avoidant attachment style (F(2, 290) = 11.86, p < .001).

Post-hoc Bonferroni analyses showed that Dom participants had a significantly more secure

attachment style than Sub participants (SE = .22, p = .01) and at the same time a significantly

less anxious-avoidant attachment style than Sub (SE = .24, p < .001), and Switch participants
 16

(SE = .26, p < .001). No differences were found in attachment styles between Sub and Switch

participants, see figure 4.

--INSERT FIGURE 4 ABOUT HERE –

When comparing controls (NI group) with BDSM community practitioners (BDSM-

CP group) who identified themselves with one of the three BDSM identity types

(Dom/Sub/Switch), differences were found for the secure (F(3, 492) = 11.34, p < .001),

anxious-avoidant (F(3, 492) = 3.82, p = .01), and anxious-preoccupied attachment style (F(3,

492) = 7.56, p < .001). Post-hoc Bonferroni analyses showed that controls had; a less secure

attachment style than Dom (p < .001), Sub (p < .01), and Switch (p < .05) practitioners; a more

anxious-avoidant attachment style than Dom participants (p < .01); and a less anxious-

preoccupied attachment style than Dom (p < .01), Sub (p < .05), and Switch (p < .01)

practitioners of BDSM. In conclusion, all BDSM identity types had a more secure attachment

style than controls, wherein dominance had the strongest association. Thereby, Dom

participants also had a less anxious-avoidant attachment style than other BDSM-identity types

or controls. No differences between all three BDSM identities and controls were found for the

dismissive-avoidant attachment style, see figure 5.

--INSERT FIGURE 5 ABOUT HERE –

Predicting intensity of BDSM interest based on attachment and trauma

Finally, in the subgroup of practitioners (BDSM-CP + BDSM-PP) an evaluation was

made whether total BDSM-scores could be predicted by attachment style, trauma scores but

also by the variables gender, sexual preference, education levels and living area. A linear
 17

regression model (method=stepwise) was constructed, yielding a highly significant model

(F=52.539; p < .001) with a R2 = .219. Sexual preference (p < .001), gender (p < .001), living

area (p <.001) and finally higher secure attachment style (p <.001) were significant predictors

for the intensity of the BDSM interest and practices. Non-heterosexual orientation, being male

and living in an urban region each predicted higher BDSM levels. Interestingly, none of the

trauma measures contributed to the model.

DISCUSSION

This study investigated to what extent BDSM interests are related to early life

dynamics such as trauma and attachment style, in both BDSM private practitioners as well as

those practitioners that were part of the outdoor BDSM community. It was found that

practitioners from the BDSM community reported more physical abuse during adulthood but

not during childhood, as well as more diseases and, although not significantly, there was a

trend towards more unwanted sexual contacts in their lifetime. BDSM community

practitioners had higher total trauma scores compared to controls and equally reported more

adult physical abuse (PBA), whereas in private BDSM practitioners only higher levels of

PBA were found. No differences between groups were found for the severity of these

traumatic events. Also, there were no significant differences in traumatic events between

private and community practitioners. Although no significant group differences emerged for

the unwanted sexual traumatic events, mild significant associations between trauma and

BDSM were found in the community practice group (BDSM-CP), which seemed mostly

driven by the factor submissiveness (SUB). Such associations were not present in the private

practice group. Levels of intensity of BDSM interest and practices were predicted by a secure

attachment style, gender, living area and sexual orientation.

Surprisingly, when gauging attachment styles, BDSM-practitioners from the

 18

community had an overall more secure and anxious-preoccupied attachment style than

controls and private practitioners of BDSM and a less anxious-avoidant attachment style than

non-practitioners having BDSM fantasies and private practitioners of BDSM. These findings

are in line with Wismeijer and Van Assen (2013) who demonstrated increases in both

adaptive and maladaptive attachment styles in BDSM practitioners compared to controls.

Interestingly, in our sample, private practitioners of BDSM did not differ in any of the four

attachment styles from controls, suggesting that BDSM community members and private

practitioners have different attachment styles, and may represent different subgroups within

the general population. The contradictory result of the BDSM-CP group being more secure

and anxious-preoccupied at the same time leaves room for interpretation. Further exploration

within the BDSM community group showed no association between BDSM interests and the

anxious-preoccupied attachment style, which shows that factors other than intensity of BDSM

interests explain this difference between groups. Solely the secure attachment style was

associated with the intensity of BDSM interests: secure attachment was associated with

dominance and voyeurism. In addition, comparing the BDSM identities with controls, this

greater amount of secure attachment was found in all three subgroups of BDSM identities, but

was the most prominent in dominants.

It has been suggested (Freud, 1905) that BDSM activities are driven by an

unconscious negative reenactment pattern in which people seek out destructive situations and

relationships that have familiarity with their past trauma. Alternatively, BDSM-practices have

also been suggested (Califia, 1983) to be a coping mechanism that somehow facilitates trauma

healing. In the present study, there is a slight non-significant increase in prevalence of

unwanted sexual contacts among the community practitioners group which may be in line

with this notion, although it should also be noted that a vast majority of them (76,6%) did not

report any traumatic experiences of this type in their past. As such, simply framing BDSM-
 19

practices as a coping style for sexual trauma would not be an adequate justification model.

Indeed, Nordling et al (2006) already accentuated the complexity of sadomasochistic sexual

behavior and the significance of a broad range of social influences on it, and argued that a

singular association between childhood abuse experiences and later sadomasochistic sexual

behavior would probably not be found. Nevertheless, 23,4% of the BDSM community group

had experienced sexual contacts they categorized as being unwanted, whereas this was only

the case for 16,2% of the controls. Women on one hand and the submissives and switches on

the other in the BDSM group seemed to drive this difference. The finding that community

practitioners (BDSM-CP) had a higher total trauma score than non-practitioners (NI &

BDSM-F), supports a possible link between trauma and BDSM. An alternative explanation

for the slight difference in prevalence of unwanted sexual contacts between the two samples

might have less to do with the actual sexually abusive event and more with its subjective

perception. BDSM-practitioners emphasize a lot on consent, safety and personal boundaries

within BDSM-play (Faccio et al., 2014). Therefore, they might have a lower threshold in

labelling/classifying certain behaviours as sexually transgressive than non-practitioners. Data

from Klement and colleagues (2017) demonstrate that BDSM practitioners reported

significantly lower levels of sexism, rape myth acceptance, and victim blaming than non-

BDSM controls.

It should be noted that both private and community practitioners had experienced

more physical beatings during adulthood (PBA) than controls, hence not during childhood

(PBC). The link between childhood trauma and BDSM interests was therefore rather not

confirmed in this study. Besides, no association was found between PBA trauma and the

intensity of BDSM practices in our community sample, despite of the BDSM practice groups

(BDSM-PP & BDSM-CP) having experienced more PBA trauma than controls. These

findings postulate a small influence of physical beatings on BDSM intensity and practices.
 20

Furthermore, it is unclear if those physical beatings were experienced within the context of

BDSM-play, which could then explain this difference between BDSM-practitioners and

controls. Elaborating on whether physical beatings happened inside or outside the BDSM

community is recommended for future studies. Again, the significant difference in PBA may

also be related to the emphasis on consent and personal boundaries as stated before.

The finding that BDSM practitioners were overall more secure in their attachment

styles than controls, resonates with sociological theories about BDSM which rather view

these practices as a recreational leisure rather than as a pathological practice. This could also

be more in line with the studies of Cowan (1982) and Stekel (1953) that showed that BDSM-

practitioners were found to be more successful and well-rounded persons. Maybe a more

secure attachment style is needed in order to participate in BDSM-play and in particular for

the dominant role. Given the kind of activities that are put into practice, a lot of trust is needed

with each participant, regardless of the role they have. This exchange of trust is generally

easier for securely attached individuals than for insecurely attached people.

It could also be that the relationship between secure attachment and dominance is

bidirectional in nature. Even though the basis of attachment is formed in the early stages of

someone’s life, it may be subject to changes due to new experiences. Therefore, it is plausible

for people with the dominant role to get more confidence and a more secure attachment style

within relationships because of the trust they are receiving from their submissive BDSM-

partners. Adopting the dominant role and performing the activities that come with it might

enhance or facilitate more confidence and secure feelings about oneself in the same way as an

assertiveness training could enhance relationships between people. But these hypotheses are

for future research to further explore.

To summarize, both associations between BDSM interest and, albeit weakly, trauma

on one hand and attachment style on the other were found. However, despite the demonstrated
 21

link between trauma and BDSM, the final model showed that none of the trauma items were a

good enough predictor for the intensity of BDSM interests, which should be taken into

consideration. Rather, sexual preference, living area, gender and secure attachment entered as

valid predictors for the level of BDSM interest.

Moreover, this is a first study that makes a distinction between private and community

practitioners of BDSM in exploring the relationship of BDSM and trauma/attachment.

Interestingly, community practitioners significantly differed from private practitioners in 3

out of 4 attachment styles. Hence, they had a more secure and anxious-preoccupied

attachment style and a less anxious-avoidant attachment style than private practitioners. This

could be due to the different contexts in which BDSM-practices are played out. Being able to

practice it more often and being part of a community could enhance the secure attachment

style practitioners experience with others. Since each BDSM identity group had a more secure

attachment style than controls, the conclusion that BDSM interests and practices come from

insecure attachment styles can’t be drawn. Rather, indications for the inverse relationship

were shown. Further, no differences were found between private and community

practitioners concerning their degree of experienced trauma, which again shows little

evidence for a link between trauma and BDSM practices.

From the perspective from social policy, our findings argue against framing BDSM-

practices as coping mechanisms for experienced (sexual) trauma or as maladaptive coping

mechanisms in general. Although these practices have been subject to stigmatization,

medicalization and even criminalization, our data, together with a large body of scientific

evidence argue much more for acceptance of BDSM practices within the realm of normal

sexual interests and behavior. Active interventions aimed at tackling the stigma surrounding

these sexual interests are needed, including the distribution of adequate information.

There may be some limitations in this study. The fact that the survey in this study was
 22

a self-report measure could lead to biased results. It is possible that there was a tendency

within the BDSM-population to portray themselves differently with the aim of changing the

fixed beliefs about people who practice BDSM. BDSM practitioners had higher education

levels and lived in more populated regions may also be a source of bias. Another limitation

concerns the different way the people in the two groups are approached for this study. We

made use of a social media approach for the BDSM group, whereas the control group were

contacted by a research bureau. Moreover, only those BDSM participants were included that

were active in online communities, and as such, do not necessarily reflect all BDSM

practitioners, thus hampering generalizability towards all subjects with BDSM interests. In

addition, the BDSM-group may also have included participants from the Netherlands, as the

invitation and the survey were presented in Dutch, a language spoken both in Belgium and

The Netherlands. As such, cultural differences between countries and different policies in the

countries concerning BDSM-practices should be considered as well. Finally, we only assessed

psychical and sexual unwanted contacts, but did not investigate other emotional and

psychological traumatic early life experiences.

To conclude, this study found an association between BDSM interest and experienced

trauma, mainly driven by physical beatings in adulthood, as well as a marginally significant

association with unwanted sexual trauma in BDSM-practitioners. In addition, people who

practiced BDSM had a more secure attachment style than non-practitioners. The intensity of

BDSM interest was predicted by a secure attachment style, sexual preferences other than

heterosexuality, gender (i.e. higher scores in males), and urban living area but not by

traumatic experiences. Taken together, these findings argue against the hypothesis of BDSM

being a maladaptive coping mechanism in response to early life dynamics. Future research is

highly recommended to elaborate on these findings.

23
 24

References

 Bartholomew, K., Horowitz, L. M. (1991). Attachment styles among young adults: A

test of a four-category model. Journal of Personality and Social Psychology, 61, 226-

244.

 Békés, V., Perry, J. C., & Robertson, B. M. (2017). Masochism: A mixed-method

analysis of its development, psychological function, and conceptual evolution. The

Psychoanalytic Review, 104, 33-63.

 Bowlby, J. (1977). The making and breaking of affectional bonds: I. Aetiology and

psychopathology in the light of attachment theory. The British Journal of Psychiatry,

130, 201-210.

 Califia P. (1983). A secret side of lesbian sexuality. In T. Weinberg & G. Kamel

(Eds.), S and M: Studies in sadomasochism (pp. 129-136). Buffalo, NY: Prometheus.

 Coppens, V., Ten Brink, S., Huys, W., Fransen, E., Morrens, M. (2020). A Survey on

BDSM-related Activities: BDSM Experience Correlates with Age of First Exposure,

Interest Profile, and Role Identity. Journal of Sex Research 57, 129-136.

 Cowan L. (1982). Masochism: A Jungian view. Dallas, TX: Spring Publications.

 De Neef , N., Coppens, V., Huys, W., Morrens, M. (2019). Bondage-Discipline,

Dominance-Submission and Sadomasochism (BDSM) From an Integrative

Biopsychosocial Perspective: A Systematic Review. Sexual Medicine 7, 129-144.

 Faccio, E., Casini, C., & Cipolletta, S. (2014) Forbidden games: the construction of

sexuality and sexual pleasure by BDSM ‘players’. Culture, Health & Sexuality, 16,

752-764.

 Feeney, J. A., Peterson, C., Gallois, C., & Terry, D. J. (2000). Attachment as a

predictor of sexual attitudes and behavior in late adolescence. Psychology & Health,

14, 1105-1122.
 25

 Finkelhor, D., & Browne, A. (1985). The traumatic impact of child sexual abuse: A

conceptualization. The American Journal of Orthopsychiatry, 55, 530-541.

 Freud, S. (1905). Three essays on the theory of sexuality. Standard edition, 7 (ed. and

trans. J. Strachey). London: Hogarth Press.

 Frías, Á, González, L., Palma, C., & Farriols, L. (2017). Is there a relationship

between borderline personality disorder and sexual masochism in women? Archives of

Sexual Behavior, 46, 747-754.

 Gentzler, A. L., & Kerns, K. A. (2004). Associations between insecure attachment and

sexual experiences. Personal Relationships, 11, 249-265.

 Grant, J., & Crawley, J. (2002). Transference and projection : mirrors to the self (p.

38). Buckingham, Open University Press.

 Holvoet, L., Huys, W., Coppens, V., Seeuws, J., Goethals, K., & Morrens, M. (2017).

Fifty shades of Belgian gray: The prevalence of BDSM-related fantasies and activities

in the general population. The Journal of Sexual Medicine, 14, 1152-1159.

 Kalichman, S. C., Gore-Felton, C., Benotsch, E., Cage, M., & Rompa, D. (2004).

Trauma symptoms, sexual behaviors, and substance abuse: correlates of childhood

sexual abuse and HIV risks among men who have sex with men. Journal of Child

Sexual Abuse, 13, 1-15.

 Klement, K. R., Sagarin, B. J., & Lee, E. M. (2017) Participating in a culture of

consent may be associated with lower rape-supportive beliefs. The Journal of Sex

Research, 54, 130-134.

 Masters, W. H., Johnson, V. E., & Kolodny, R. C. (1995). Heterosexuality. New York:

Perennial.

 Meston, C. M., Heiman, J. R., Trapnell, P. D. (1999). The relation between early

abuse and adult sexuality. The Journal of Sex Research, 36, 385-395.
 26

 Nordling, N., Sandnabba, N. K., & Santtila, P. (2000). The prevalence and effects of

self-reported childhood sexual abuse among sadomasochistically oriented males and

females. Journal of Child Sexual Abuse, 9, 53-63.

 Renaud, C. A., & Byers, E. S. (1999). Exploring the frequency, diversity, and content

of university students’ positive and negative sexual cognitions. The Canadian Journal

of Human Sexuality, 8, 17-30.

 Richters, J., de Visser, R. O., Rissel, C. E., Grulich, A. E., & Smith, A. M. A. (2008).

Demographic and psychosocial features of participants in bondage and discipline,

‘sadomasochism’ or dominance and submission (BDSM): Data from a national

survey. The Journal of Sexual Medicine, 5, 1660-1668.

 Roush, J. F., Brown, S. L., Mitchell, S. M., & Cukrowicz, K. C. (2017). Shame, guilt,

and suicide ideation among bondage and discipline, dominance and submission, and

sadomasochism practitioners: examining the role of the interpersonal theory of

suicide. Suicide and Life-Threatening Behavior, 47, 129-141.

 Schnurr, P. P., Spiro, A., Vielhauer, M. J., Findler, M. N., & Hamblen, J. L. (2002).

Trauma in the lives of older men: Findings from the normative aging study. Journal of

Clinical Geropsychology, 8, 175-187.

 Schnurr, P. P., Vielhauer, M. J., & Weathers, F. W. (1995). Brief Trauma Interview.

Unpublished interview.

 Smith, D. K., Leve, L. D., & Chamberlain, P. (2006). Adolescent girls' offending and

health-risking sexual behavior: The predictive role of trauma. Child Maltreatment, 11,

346-353.

 Stekel, W. (1953). Sadism and masochism: The psychology of hatred and cruelty (tr.

E. Gutheil). Volume 1 and 2. New York: Liveright. (Original work published 1929)

 Szielasko, A. L., Symons, D. K., Price, L. E. (2013). Development of an attachment-

 27

informed measure of sexual behavior in late adolescence. Journal of Adolescence, 36,

361-370.

 Wismeijer, A. A., & Assen, M. A. (2013). Psychological characteristics of BDSM

practitioners. The Journal of Sexual Medicine, 10, 1943-1952.

 28

Private Communit
No Interest Fantasy
practice y practice p
Number of completers 326 (25%) 192 (15%) 559 (43%) 212 (16%)
44.55 ± 37.91 ± 37.19 ± 40.83 ±
Age 13.14 13.78 11.73 13.64 <.001*
Male 39.9% 54.7% 44.7% 56.6% <.001#
Gender Female 59.8% 45.3% 54.7% 42.5%
Other 0.3% 0.0% 0.5% 0.9%
Urban 16.0% 13.5% 19.3% 31.6% <.001#
Living
Suburban 26.1% 28.6% 24.5% 59.9%
Area
Rural 58.0% 57.8% 56.2% 40.1%
Heterosexual 93.6% 90.1% 87.1% 63.2% <.001#
Homosexual 3.1% 6.3% 6.4% 4.7%
Sexual
Bisexual 2.1% 2.6% 5.4% 25.0%
Preference
Asexual 0.9% 1.0% 0.2% 0.9%
Other 0.3% 0.0% 0.9% 6.1%

Table 1: Demographic parameters on socioeconomic status and sexual

preference of people with no BDSM interest, people fantasizing about

BDSM and practitioners who enact at home (private practice) or in a

community setting (community practice). Data marked by * are analysed

by ANOVA, data marked by # are analysed by contingency analysis and

deemed true effects if standardized residual ≥2 .

 29

BDSM-groups NI BDSM-F BDSM-PP BDSM-CP

Traum
a n % n % n % n %
items
War zone
No 303 98,1 182 97,8 526 96,9 198 94,3
Yes 6 1,9 4 2,2 17 3,1 12 5,7
Car accident
No 261 84,5 164 85,4 463 85,3 170 81,3
Yes 48 15,5 22 11,8 80 14,7 39 18,7
Natural/technological
disaster
No 289 93,5 177 95,2 515 94,7 197 93,8
Yes 20 6,5 9 4,8 29 5,3 13 6,2
Life-threatening illness
No 288 93,2 173 93,0 510 93,8 183 87,6
Yes 21 6,8 13 7,0 34 6,3 26 12,4
Physical punishment < 18
years
No 275 89,0 161 86,6 467 86,0 174 82,5
Yes 34 11,0 25 13,4 76 14,0 37 17,5
Physical violence in
adulthood
No 251 81,2 140 75,3 394 72,7 140 67,3
Yes 58 18,8 46 24,7 148 27,3 68 32,7
Unwanted sexual contact
No 259 83,8 162 87,6 436 80,3 160 76,6
Yes 50 16,2 23 12,4 107 19,7 49 23,4

Table 2. Occurrence rates for the seven types of traumatic events for each of the four BDSM

interest level groups (NI: neither fantasies nor practices; BDSM-F: fantasies without

practices; BDSM-PP: practices in private setting; BDSM-CP: practices within a BDSM

community/club) used in this study.

 30

Figure 1. Mean scores of each BDSM summary score for all BDSM intensity level groups

(NI: neither fantasies nor practices; BDSM-F: fantasies without practices; BDSM-PP:

practices of BDSM in private setting; BDSM-CP: practices of BDSM in community/club

setting).
 31

Figure 2. Mean rank scores of the Kruskal-Wallis test and Mann-Whitney test results of all

group comparisons for the variables PBA, USC and the total trauma score. A Bonferroni

correction with a critical value of p = .0083 as significance level was applied. (NI: neither

fantasies nor practices; BDSM-F: fantasies without practices; BDSM-PP: practices of BDSM

in private setting; BDSM-CP: practices of BDSM in club setting).

 32

Figure 3. Mean attachment styles scores and MANOVA test results for all group comparisons

using post-hoc Bonferroni analyses. (NI: neither fantasies nor practices; BDSM-F: fantasies

without practices; BDSM-PP: practices of BDSM in private setting; BDSM-CP: practices of

BDSM in community/club setting).

 33

Figure 4. Mean attachment style scores and post-hoc Bonferroni test results (MANOVA) of

comparing the BDSM identities within the BDSM practice groups (BDSM-PP & BDSM-CP).
 34

Figure 5. Mean attachment style scores and post-hoc Bonferroni test results (MANOVA) of

the comparisons between controls and all BDSM identities from the community practice

group (BDSM-CP).