The differences between sex offenders who victimise older women and sex offenders who offend against children

ABSTRACT Objectives: Within the literature on sex offending, much attention is paid to the distinction between those sex offenders who offend against adults and those who offend against children. In contrast, there is a paucity of research into sex offenders who offend specifically against elderly or older victims. Method: A detailed interview and psychometric tests were conducted with a sample of 28 sex offenders who had been convicted of a sexually motivated offence against an older female. These data were compared to a sample of 23 child sex offenders. Results: Results indicate that amongst other significant differences between these sub-groups, men who offend against older women are generally younger, are more violent, and are more likely to use a weapon and cause injury and death compared to child sex offenders. The men who offended against children were more likely to think about and plan their offending, spend more time with the victim pre and post offence, admit sexual arousal during the offence, and admit to a sexual motivation for the offence. Conclusions: This study suggests that men who sexually offend against older women and men who sexually offend against children are distinct groups. Treatment and risk management strategies should take this into account. Further exploration of this sub-group of offenders is recommended to help inform treatment and risk management strategies for sex offenders who offend against older people.


Introduction
As the number of older people increases in society due to an increase in health care and general standard of living, interest in issues related to older people has increased. The fastest population increases will be seen in the numbers of those aged 18 and over in the UK (ONS, 2012). An increase in interest in the study of older adults has included studies into the impact of crime; as people live longer the prevalence of crime against these older people may increase (Safarik, Jarvis, & Nussbaum, 2002). Older people already report high levels of discrimination based on their age (Age UK, 2016), and to neglect to focus on their victimisation, when there is much focus on crimes against children, would add to their disadvantage. When referring to crime against the elderly there appear to be two categories, that of general crime and that of elder abuse (e.g. Age UK, 2016). Elder abuse frequently focuses on the home and institutions as the source of abuse, and approximately 5% of older people in the UK are estimated to be abused in private households each year (Age UK, 2016). With regard to both categories of crime, the focus is often on violent, psychological, and financial abuse (Defra, 2011). Sexual abuse against older people, specifically against older women, in the past was more rarely identified. However, in more recent years the interest and acknowledgement of all offences against older people including abuse has grown, and sexual abuse is now recognised as an important feature of that abuse (Ball, 2005;Jeary, 2005).
Although sex crime against older people may be perceived as a low frequency event, it is likely that this is actually a higher frequency event than it is generally believed to be (Cooper & King, 2006). Three per cent of calls to a helpline in 2004 were related to sexual abuse of older people (Defra, 2011), which is likely to underestimate the scale of the problem. The reasons for the neglect of older sex crime victims may well link to classic theories of 'rape myths' whereby rape victims are seen as young and attractive, assaulted by strangers (Anderson & Docherty, 2008). Furthermore, where a crime against an older victim is reported, such myths may contribute to the complainant's allegation being disbelieved or even discredited (Burgess, Dowdell, & Brown, 2000), and this is further complicated by older victims being at risk of their disclosure being ignored and attributed simply to dementia (Burgess & Clements, 2006). Consequently, despite increase interest in this area, reporting of such offences is likely to underestimate true rates of this type of crime, and official rates are difficult to determine due to the manner in which public response surveys are reported, with figures for the UK not being at all clear (Walby & Myhill, 2001). In relation to older victims specifically, it is suggested that sexual victimisation is underreported even more so than the underreporting of sex offending in general (Burgess & Clements, 2006). Lea, Hunt and Shaw (2011) even cited scepticism as to the need for studies in this area due to the perceived infrequency of this type of offending. However, older people who are offended against may suffer from similar, and even more devastating physical and psychological consequences due to their victimisation. For instance, genital damage has been found to be more frequent and more severe in elderly sex assault victims (Eckert & Sugar, 2008;Poulos & Sheridan, 2008). Furthermore, offences against older people have been found to manifest as similar offences to those perpetrated against younger people (Teaster & Roberto, 2004). Given the nature of the offending and the consequences on victims, the current authors argue that this area warrants research as much as any other area of sexual offending.
The difficulty in examining sex offending against older victims stems from the relative lack of existing research, meaning that researchers have little to start from, including the usual difficulties in assessing prevalence rates of sex offending against any victim group (Falshaw et al., 2003;Rice, Harris, Lang, & Cormier, 2006). Added to this, there is the difficulty of researchers accessing appropriate samples and engaging older victims or the perpetrators in research, and when this does occur, the sample is usually opportunistic and may not represent older sex offenders as a group.
In searching to understand sex offending, it is relevant to consider well-established rape theory and its application to elderly victimisation. Groth (1978) applies his work to the rape of older females. He argues that the older woman appears to symbolise an authority figure over whom the offender wants control, and/or an actual woman against whom he wants to retaliate against or achieve revenge. Groth, Burgess, andHolstrom (1977, 1979) argue that sexuality becomes the means to express power and anger, to hurt and humiliate the victims, and that the sexual behaviour serves non-sexual needs and motives. Groth (1978) purports that vulnerability and accessibility play a more significant role in victim selection than physical attractiveness or alleged provocativeness. This part of Groth's theory links in with the later work of Burgess, Dowdell, and Prentky (2000) who suggest three types of offenders within older adult sex abusers (OASAs). One of these groups is the type where the offender has a job in a caring role such as within nursing homes, although Burgess suggests that the job role is sought specifically to offend, whereas Groth's theory would indicate that the offending is linked more to accessibility. Groth's theories suggest that rape is more an issue of hostility than sexual desire, especially in relation to older female victims who are weak and vulnerable and less likely to offer resistance. However, Groth offers no explanation as to why a sexual assault occurs, nor does he offer an explanation for offenders who would disclose a sexual interest in older people. His model may therefore exclude a number of offenders. It is possible that older female victims of sexual assaults may be chosen because of the offender's sexual preference. Elderly females have particular qualities that other victim groups do not, such as the feel of the skin or other aspects about their appearance. Some sexual offenders may find these physical characteristics sexually arousing, or they may be aroused by other factors such as this victim group's relative vulnerability. They may have this in common with child sexual abusers (CSAs), who could also be aroused by the vulnerability of their chosen victim group. , who proposed three groups, attempt to examine OASAs as a separate group. It is this approach that moves the literature in this area forward, because this group is being recognised as potentially being distinguishable from other groups of sex abusers such as child sex abusers or those who rape younger adults.
The three groups who offend against older women proposed by  are: those who seek a job caring for older adults, older men who live with older women (e.g. in a residential home), and offenders who victimise older women who live independently. Although these may well be distinguishable groups, this does not provide a bigger picture nor aids us with causality. There are some studies that examine the characteristics of OASAs, for instance, Cartwright and Moore (1989) detailed the characteristics of 21 rapes against older adults, and Collins and O'Connor (2000) examined 10 sex assault crimes against this victim group. These studies have limited generalisability for various reasons including sample source and size; however, more comprehensive studies of these offenders enhanced our understanding of OASAs by attempting to propose typologies in a manner consistent with attempts to typologies general adult rapists (Burgess et al., 2007). In order to more comprehensively understand OASAs, researchers do however need to be aware of differences between them and other groups, and comparative studies have helped with this.
Some studies have compared OASAs with rapists of younger adults, such as Muram, Miller and Cutler (1992) who found differences in location and victimÀperpetrator relationship. OASAs offended against the victim in their home at a much higher rate than the comparison group, and OASAs had a higher rate of victimisation of a stranger. Despite some differences being found, these were few, and from this one could suggest that OASAs are not particularly distinct. However, it could also be argued that a different comparison group may have revealed more differences, such as a child sex abuser comparison group. Attempts to examine similarities and differences between OASAs and other groups are important, not least because this can inform treatment and risk management of these offenders, but also so that any differences may be used to help protect older people from victimisation. In the UK, sex offenders are offered a Sex Offender Treatment Programme (SOTP) in prisons and in the community. This SOTP is designed around areas that the literature deems to be relevant to risk and the same treatment is offered to all sex offenders, regardless of victim group. Sex offenders are a heterogeneous group and in relation to risk and treatment, and this needs careful consideration and research, as it could well be the case that SOTP based on general risk factor literature does not target factors relevant to OASAs. Sub-groups of sex offenders are often neglected within risk assessment research, within which it is often assumed that issues such as reconviction statistics apply to all sex offenders . Sub-groups of offenders are often neglected and this can be due to the changing nature of sex offending such as newer online offences, and the difficulties in examining sub-groups of offenders who may or may not choose to consent to engage in research activities. If OASAs are different to other types of offenders, this has implications for the treatment offered to them, with important issues regarding appropriateness of the design of the treatment. The treatment targets of the current SOTPs may not be relevant (or as relevant) for OASAs, potentially leaving their risk factors unaddressed, therefore leaving older adults at risk of victimisation. Therefore, the safe management of the risk of OASAs when in the community could be impacted upon if we have increased knowledge of this group and how they may differ to other sub-groups of sex offenders. Given that risk of serious harm is high when the victim is elderly (Poulos & Sheridan, 2008), enhanced understanding of this type of offending is very important.
Overall, sex offending against elderly women is a problem in society and these victims deserve relevant research to be conducted as much as any other victim group. Offending against elderly men also occurs  but it is acknowledged in the literature that this tends to occur with males as the offender, and females as victims (Ramsey-Klawsnick et al., 2007). The current research aims to help fill the gap in the research into OASAs by comparing a group of OASAs with a group of CSAs, using a sample of UK prisoners. This study will include direct contact with OASAs, and will examine differences using both interview and psychometric assessment in order to provide a more comprehensive picture when compared to less comprehensive studies that have conducted analyses on only crime databases or file information (e.g. Burgess, Dowdell & Prentky, 2000;Ramsey-Klawsnick et al., 2008).

Ethics
This study received approval from the relevant ethics committees.

Participants
It is not known how large the population of sex offenders who have offended against older people is in the UK, as these data were not collected on a national level. Seventy adult male sex offenders who had sexually offended against an older female were contacted within Her Majesty's prisons of England and Wales. Forty-two of this group did not reply or refused to participate, leaving a group of 28 who participated. Thirty child offenders were initially contacted and 7 declined the offer of participation, leaving a group of 23. All participants were serving a determinate or life sentence in a variety of prison establishments around England. There were no dropouts after the contact phase of the study.
Twenty-eight (of these men had either been convicted of a sexual offence against an older woman or had been convicted of a sexual offence against an older woman where there was evidence of a sexual motivation. This is the OASA group. The term 'older' referred to any female victim aged 50 or over, or where there was a 25-year age gap between the offender and victim, so as to include those who may have purposely sought out an older person to sexually offend against. Twenty-three participants had each been convicted of at least one sexual offence against a child (CSA group). This resulted in a total sample of 51 sex offenders. There was a significant difference in age for OASA (M D 37.25, SD D 9.47) and CSA (M D 44.78, SD D 7.82) conditions; t (49) D 3.05, p D 0.004, meaning that the CSA group was significantly older than the OASA group.
All information to inform the study, e.g. demographic information, was obtained from prison files and directly from the participants during interview. The relevant permissions meant that limited demographic information was available.

Procedures
A variety of prison establishments were contacted and potential participants who had offended against either at least one child, or at least one elderly victim were identified. A call for participants was distributed to these prisoners. Those who expressed an interest were then visited, and a full explanation was given with the option to withdraw from the study at any time. Full written information was given and all participants signed a consent form. All participants were then interviewed using a semi-structured format (taking 1À1.5 hours) alongside completion of a battery of psychometrics (described in detail below). Participants were thanked and debriefed.

Materials
The battery included psychometrics designed to identify sexual interests, levels of anger, attitudes towards women, and levels of empathy. All questionnaires were scored in accordance with the psychometric scoring guides.
Measure of empathic tendency (MET; Mehrabian & Epstein, 1972) This is a 33-item questionnaire used to measure emotional empathy. Originally designed as a true or false response measure, this was later adapted a 7-point scale and combined with the 12-item Personal Reaction Inventory to assess social desirability. The focus of the empathy questionnaire is to measure emotional empathy, whereas other empathy questionnaires measure empathy with perspective taking which is a cognitive skill.
Burt rape myth acceptance scale (BRMAS; Burt, 1980) This is a 19-item scale that measures the endorsement or rejection of beliefs sometimes used to justify rape. Burt (1980) found that high scores on the BRMA scale correlated with sexual conservatism (r D .39), adversarial sexual beliefs (r D .40), and acceptance of violence (r D .50) and reported a Cronbach's alpha of .88 for the BRMAS.
The reaction to provocation scale (NAS; Novaco, 2003) The NAS is a 3-point scale measuring an individual's ability to become angry and respond aggressively to a situation. It is divided into two parts: A and B. Part A is divided into three domains: cognitive, arousal, and behavioural. Novaco reports that the scale has an internal consistency (a D .97) with a test retest reliability of (r D .86).
The multiphasic sex inventory (MSI; Nichols & Molinder, 1984) The MSI psychometric is designed to assess a wide range of psychosexual characteristics of a sexual offender. There are 20 sub-scales with 6 of these being validity scales. The remaining 14 sub-scales include 3 scales of sexual deviance, 5 scales related to atypical sexual behaviour, 4 scales of sexual dysfunction, a sexual knowledge scale, and a treatment attitude scale. The authors report the total reliability for all items on the MSI to be (r D .86).

Semi-structured interview
The semi-structured interview consists of 19 sets of questions considering areas such as historical events, sexual history, substance use, preparations for offending, and consequences of offending. Some of the questions have a number of options; others require the participant to produce answers. The semi-structured interview (SSI) is not standardised and was designed to elicit the information required for the present study.

Treatment of data
All psychometric tests were scored using the appropriate scoring guides. The PASEQ and the SSI were assigned values for each question and this information.
The data analysis was carried out using SPSS. Means and standard deviations are documented for all psychometrics in Tables 1 and 2. Two independent samples were used in this study and the psychometric data were analysed using a Man-nÀWhitney non-parametric test due to non-normal distribution of data. The SSI results were analysed using Pearson chisquare due to the application of coding.

Descriptive statistics
The mean age of the victim for OASA group was 69.86 with a standard deviation of 12.38. The mean victim age of the victim for CSA group was 9.17 with a standard deviation of 2.90.

Psychometric measures
None of the data were normally distributed. For the psychometric measures, a significant difference was found between groups for emotional empathy, such that OASAs demonstrated significantly lower scores than CSAs. No significant between group differences were found for any of the NAS subscales or total score. Seven of the 18 subscales of the MSI had significant differences.
The mean and standard deviations for all psychometric measures are displayed in Tables 1 and 2 with statistical significance between the OASA and CSA groups identified with an asterisk.

Semi-structured interview
Twenty-three of the 49 areas considered demonstrated significant differences between groups. All 5 of the 'victim characteristic' items, none of the 14 'past history of offender' items, 1 of the 6 'offence lead up' items, 12 of 18 'offence characteristics' items, and 5 of the 6 'post offending' items significantly differed between groups. Although not a statistically significant difference, stealing from the victim was a feature of some OASA offending (28.60%), and was not a feature of any of the CSA offending (0%). Table 3 identifies the characteristics of the CSA and OASA groups by domain. Significant differences are highlighted.

Discussion
The current study applied a semi-structured interview and psychometric testing to OASAs and CSAs. Some areas were found to significantly differ between these two sub-groups of offenders, suggesting that these two groups are distinct. First, the OASA group were significantly younger than the CSA group, indicating that a young age may be an aggravating risk factor for this type of offence. Although there were few differences between some of the more interpersonal areas of the offenders' past, such as friendship and relationships, there was an important difference between the length of time the OASAs thought about offending before actually committing the offence. In total, 35.7% of the OASA group admitted a previous attraction to, or offending against, someone of this age range before. If they had offended previously, this supports the difference between OASA and CSA ages, because this would have lowered the mean age even further had the offender been included in the study at the first time at which they offended. The fact that many OASA participants admitted to some previous attraction to older victims or previous offending suggests a more chronic, than acute, interest in sex offending against the older female victim group. However, similar to the interpersonal aspects of the offenders' past, there were no between group differences in issues such as employment status or drug or alcohol problems in the lead up to the offence. That is not to say there were no problems in these areas, because in fact there were more offenders with drug than alcohol problems in both groups.
The offence characteristics, victim characteristics, and post offence behaviour were the areas that were found to differ most between OASAs and CSAs. In terms of offending the OASA group reported spending less time thinking and planning the offence, some acting quite spontaneously. This type of offender is included in Knight and Prentky's (1990) model; the opportunistic and impulsive offender whose offending is triggered by context and antecedent factors. The OASA group were also more prepared to carry and use a weapon during the offending and seriously harm or kill the victim, even CThe MSI manual indicates that only the MSI 'lie' scales that relate to the offender's offence type (e.g. child victim, rape) should be examined, therefore the child molest lie scale is irrelevant to the OASA group and the rapist scale is irrelevant to the CSA group and therefore significant differences between groups on these two subscales is not interpreted. though many of the victims were elderly and already very vulnerable due to their age, meaning that they are more likely to be harmed simply by being victimised (Eckert & Sugar, 2008;Poulos & Sheridan, 2008). As already discussed, this level of aggression perpetrated by the OASA accords with the causal theories that emphasise the role of anger and aggression in sexual offending against adults. Groth (1978) also notes the extreme physical force used by this group of offenders, with 60% using intense violence. Similarly, in this study 60.7% seriously harmed or killed their victim, with this giving the picture of OASAs as being more impulsive, opportunistic, and violent than CSAs.
OASAs were found to be less likely to offend against a family member than CSAs and more likely to offend against strangers or acquaintances, and almost 90% of the OASAs offended against the victim within an hour of meeting them. Strangers are therefore most at risk from OASAs compared with CSAs who are more likely to know their victim, in line with authors who discuss offence behaviours such as grooming the child (e.g. Burn & Brown, 2013). This links in with the current finding that CSAs used less violence in their offending than OASAs, because strangers may be less likely to comply regardless of age, whereas a groomed victim may have been influenced enough by the offender's prior contact with them to comply with the offending. Indeed, the CSAs reported using persuasion more than the OASAs did.
More OASAs offended against the victims in their homes, but less specifically targeted that particular victim and less OASAs attended the offence location specifically in order to offend compared to CSAs. Theft from the victim was a feature of OASA offending, and not a feature of CSA offending. This suggests that the location of offence was more pertinent to offending for the OASAs than the actual victim. This is further supported by the CSA group planning a time to offend at a significantly higher rate than the OASA group, with this together suggesting that the OASA group's offending is more impulsive and less planned than the offending of CSAs. Previous studies have found that sex offences against older victims occur in the victim's home residence, however this has also been found to be within care home facilities lived in by the elderly (Stermac, Del Bove, Brazeau, & Bainbridge, 2006).
Although it has been reported that offences against older people have been found to manifest as similar offences to those perpetrated against younger people (Teaster & Roberto, 2004), the present study found that there were some differences. Although there was no significant difference between frequency of anal or vaginal penetrative offences, OASAs were less likely to force the victim to masturbate them and were less likely to force oral sex than CSAs. This may link with the manifestation of other characteristics of the offence such as the duration of the offence, and CSAs possibly having more compliant victims due to grooming behaviour. This is consistent with other studies that have found differences in the nature of sex offences of OASAs, such as Lea, Hunt and Shaw (2011); however, these authors compared OASAs with those offenders who rape younger adult women. Vaginal assault was most common for OASAs and CSAs, consistent with Thomas, Titley and Esquibel (2015) who although focussing on discussion of mid-life victims, did compare sexual offence types across age groups.
Almost two thirds of the OASAs reported that their victims resisted during the offending, which was significantly higher than was reported by the CSAs. However, other studies have identified the vulnerabilities of older victims, such as physical disabilities, mental health issues, or general poor functioning as resulting in lower likelihood of resistance Teaster et al., 2000), which is not consistent with the findings of the present study. It may be that post-hoc justifications for offending, and the level of force used against such vulnerable victims, influenced offender self-report of resistance. This could also have been influenced by impression management strategies being applied by the offender when interviewed. Future studies could gather this information from official records of the offence rather than relying on offender self-report, and an impression management scale could be used. Differences were found also with the level of sexual interest and arousal reported by OASAs and CSAs. Significantly fewer OASAs reported having an erection during the offence, with the same pattern of reporting regarding ejaculation. This is consistent with a related post-offence behaviour of fantasising about the offence, with significantly fewer OASAs reporting such fantasies. This supports a hypothesis that OASAs may have been less driven by sexual deviance/paraphilic interests than CSAs. When considered in the context of the results reported above, this suggests that OASAs offend on a more opportunistic basis, with less planning than CSAs, less victim targeting, and lower levels of sexual deviance. The victims may happen to be selected on a non-planned basis but they are more likely to be seriously harmed by the significantly higher levels of violence used by OASAs than CSAs, and also possibly due to their existing vulnerabilities. This does pose a useful finding of the present study, as opportunism suggests to some degree that the self-protection and security of older adults may be a factor that leaves them less at risk of sexual abuse. Therefore, consideration of physical security and methods that older people may have in place to alert others if they are in distress may help to reduce victimisation by opportunistic OASAs.
The OASA group displayed no more beliefs justifying rape than the CSA group using the BRAMAS (Burt, 1980). The mean score for both the OASA and CSA groups was below the nonclinical mean scores for the BRMAS. This is somewhat contradictory to the older causal theories of sexual offending, which tend to emphasise negative attitudes and distorted beliefs towards victims. Well established rape theories suggested by Groth (1978Groth ( , 1979 use terms such as 'degrade', 'humiliate', and 'revenge' and although Cohen, Seghorn, and Galamas (1969) place more emphasis on the sexual motivation, they too endorse negative beliefs towards women using terms such as 'humiliate' within their postulations. Knight and Prentky also argue that one of their types of offender is motivated in part by distorted cognitions. This is not evident in the scores on the BRMAS (Burt, 1980), however, this particular psychometric is very transparent as to what it is measuring, which may have biased responding in the current study. Furthermore, it is possible that some of the offenders could have addressed or successfully treated some of these attitudes whilst in prison, with this factor not having been controlled for in the present study.
The OASA group had a significantly lower mean score on the MET (Mehrabian & Epstein, 1972), indicating that they were less able to empathise on an emotional level with others. Empathy deficits can therefore be considered as important treatment target for OASAs. Even though some authors suggest that empathy does not have broad empirical support as a risk factor for sexual offending (Mann & Barnett, 2013), it may be that OASAs as a distinct group do require improved empathy skills in order to reduce risk. Although results that inform how to directly protect older people are important, identification of relevant treatment to inform the SOTP offered to offenders can help to reduce the risks posed to older people in our communities. Although the impact of sex offender treatment is not certain and this area lacks high quality evaluation studies (L€ angstr€ om et al., 2013), these present findings could be used to personalise treatment for OASAs, who may, for example, have more of a need to examine empathy within psychological treatment programmes.
The NAS (Novaco, 2003) scores did not identify significant differences in anger between the OASA and CSA groups. These results were surprising as some causal theories of sexual offending emphasise the role of aggression and anger in sexual offences committed against women. Groth (1978Groth ( , 1979 clearly defines the role of anger and hostility in sex offences and Cohen (1969) focuses on aggression as a key motivation in his theory of sexual offending. Furthermore, Knight and Prentky emphasise anger and aggression as a motivation in three out of their four types of sexual offender. It was therefore anticipated that OASAs may score significantly higher on an anger scale such as the NAS, but this did not occur. However, this was in contrast with OASAs narrative responses during interview, which did identify a significant difference regarding the offenders' feelings about the victim and within the offence. OASA identified a higher rate of negative emotions towards the victim including feelings such as hate or anger. This group also significantly used or carried a weapon during the offence and had a much higher rate of seriously harming or killing their victim compared to CSAs. This is more consistent with the causal theories. Whilst the NAS is a well validated tool accepted within the forensic field as a measure of components of anger, it may be that because the questions relate to general feelings of anger and aggressive behaviour rather than offence focused questions, the NAS was unable to measure offence or victim related anger/ aggression effectively. Given this finding, future similar studies could use more than one anger measure, e.g. the NAS and the STAXI-II (Spielberger, 1999), and they could apply a deception or impression management scale, as is recommended when assessing forensic samples (Schamborg, Tully & Browne, 2015).
The MSI (Nichols & Molinder, 1984) identified several significant differences between the two groups. Some of these differences were on the child molest scale and the exhibitionism scale, where differences between groups of men who had sexually offended against adults and men who had offended against children would be expected. However, there were other differences that add to our understanding of OASAs. The OASA group scored significantly higher on the sexual deviance and development scale and the cognitive distortions and immaturity subscales than CSAs. The cognitive distortions and immaturity scale is designed to measure the degree to which an offender is distorting his thinking and behaviour, and this higher level of distortion by OASAs may have some link with the poorer empathy scores for OASAs as measured by the MET. This indicates that cognitive distortions and empathy may be more important treatment targets for OASAs than CSAs. The OASA group did not score significantly higher than CSAs on MSI scales such as the sado-masochism subscale, which would have been expected based on what is discussed above relating to some of the causal theories of adult rape. Both Cohen's and Knight and Prentky's theories of sexual offending consider sadistic motivation in their models, yet this was not reflected in the MSI outcomes within the current study. Again the MSI is a very transparent as a psychometric scale, and the questions relating to sadistic behaviour are easily identifiable. Offenders may have been very cautious about indicating any kind of sadistic interests even though Nichols and Molinder argue that even a successfully treated offender with such interests will endorse these items. It may be that the MSI failed to identify elements such as sadism for this reason, or it may have been that issues such as the length of the scale (over 300 items) resulted in inconsistent or random responding on the part of respondents. When reviewing qualitative information about the offences of the OASA group, it is clear that a number of these offences contained elements that were so violent that they can be considered to be sadistic, such as torturing the victim. Utilising psychometric testing alongside other sources of information such as interview and official records is therefore highlighted as a recommendation for future research into the differences between OASAs and other types of offenders.
As discussed above, studies such as this that use interview and psychometric assessment are limited by the nature of offender self-report. A deception scale could be applied in future studies; however, in the present study research participants had nothing practical to gain from engaging in the project. This may have encouraged openness. Future studies could include a CSA, OASA and a younger adult rape victim group, to compare offender subtypes, and could also include a non-offending control to compare psychometric test outcomes to. Furthermore, it is important that when an offender is received into prison professionals record if an offender is an OASA, or not, so that appropriate treatment and risk management considerations can be made. This would not only assist treatment and risk management, but would also aid further research into the differences not just between OASAs and other types of sex offender.
This present study defined an older victim as being aged 50 or above, and included those offenders where there was a 25-year age gap between offender and victim. The definition of 'elderly' or 'older' adult may differ between studies, making them non-comparable. Future research should therefore clearly define their definition of elderly, and it may be worthwhile to conduct sub-group analysis, e.g. victim aged 50-60, 60-70, and so on, where data allow for this. This may further add to our understanding of OASAs, in order to help us to better protect older people from sex crime.

Conclusions
This current paper has enhanced understanding regarding those who sexually offend against the older female population. Accepting the study's relatively small sample size compared to the possible population size, it is of note that differences were found in the nature of offending, characteristics of the offence, victim characteristics, and post-offence behaviour. The OASA profile appears to be a younger offender who is more opportunistic than the CSA. Older victims are more at risk in their own home than other places, and this is particularly worrying given that the home can be considered to be a place where one might expect to be safest. The opportunistic offending against older women highlights the importance of practical safety measures and vigilance on the part of older people (and their carers where appropriate) in order to help maintain personal safety and reduce risk of victimisation. OASAs reported lower sexual arousal, with lower empathic ability, and according to traditional rape theories they may be more similar to the 'angry' rapist than other types. Sex crimes against older women were found to be more violent, with increased likelihood of serious physical harm, which highlights how important this phenomenon is to research and understand. This offender group should be studied further to help treatment providers know what areas to target, and importantly, to help protect older people from victimisation.

Author note
All three authors of this paper contributed equally in different ways and therefore receive equal author status.

Disclosure statement
No potential conflict of interest was reported by the authors.