The Association between the Utilisation of Empowerment Strategies and Clients’ Changes of Self in the Field of Intimate Partner Abuse: from the Perspective of Social Workers

(1)

The Association between the Utilisation

of Empowerment Strategies and Clients’

Changes of Self in the Field of Intimate

Partner Abuse: From the Perspective

of Social Workers

Li-yu Song*

*Correspondence to Li-yu Song, Ph.D., No. 64, Sec. 2, Zhi Nan Rd, Wenshan District, Taipei City, Taiwan, 11605, Republic of China. E-mail: liyusong@nccu.edu.tw

Abstract

Over the past decade, protection service and treatment for victims of intimate partner vio-lence (IPV) have been the major area of social work practice in Taiwan. This study examined the impact the services could bring to the victims with the focus on clients’ changes of self. Especially, the association between the utilisation of empowerment strategies and the changes was explored. Five groups of contextual variables were also taken into account, and the relative importance between the utilisation of empowerment strategies and the contextual variables was tested through model comparisons. Data were collected from 243 social workers through a survey. The findings revealed multiple domains of clients’ changes of self. The utilisation of empowerment strategies was the most important correlate of the changes. The unique contribution of this variable to clients’ changes of self was 13.9 per cent, whereas it was much lower for the contextual variables (7.2 per cent). The contextual variables, especially life satisfaction, service provision and causes of vio-lence, contribute to such changes through the utilisation of empowerment strategies. Workers’ life satisfaction was the second most important correlate of such changes. The facilitations on the utilisation of empowerment strategies, workers’ life satisfaction and more services to the victims of IPV are discussed.

Keywords: Utilisation of empowerment strategies, changes of self, intimate partner abuse, life satisfaction

Accepted: June 2013

British Journal of Social Work (2013) 1–22 doi:10.1093/bjsw/bct134

at National Chengchi University on December 4, 2013

http://bjsw.oxfordjournals.org/

(2)

Introduction

Intimate partner violence (IPV) is a devastating experience for the victims and could cause tremendous physical and psychological harm to them (Wiehe, 1998). Over the past decade, protection service and treatment have been enforced to this population, and the manpower in this field has increased in Taiwan. In each city and county, the Center for Domestic Violence Prevention and Intervention (DVPI) has been established and operated by a multidisciplinary team. The health department handles the perpetrators that are ordered by the court for mandatory treatment, whereas social workers mainly serve the victims and provide emergency man-agement and supportive services for the remedy of trauma and reconstruc-tion of their lives. Social workers use case management as the mechanism to assess service users’ needs, make referrals, coordinate and monitor the ser-vices, and evaluate the outcomes. The question is what kind of impact would the services bring to the victims? The primary concern of the stakeholders is to prevent the reoccurrence of violence and to ensure the safety of the victims. It is an understandable goal; however, it has proven a challenge as perpetra-tors of IPV have complicated psycho-social and psychiatric histories (Stover et al., 2009). Many perpetrators have witnessed family violence or were victims of abuse as children. Some of them have personality disorders, which are difficult and take a long time to treat. The review ofStover et al. (2009) showed that rates of recidivism in most perpetrator and partner-focused treatment are approximately 20 – 30 per cent within six months, regardless of the intervention strategy used. They suggested that incorpor-ation or coordinincorpor-ation of advocacy and counselling for victims and children, along with perpetrator intervention, may yield the best overall outcomes for family impacted by IPV. The negative impacts of IPV on victims are pro-found, including depression, post-traumatic stress disorder (PTSD), suicide and other anxiety disorder (Lundy and Grossman, 2001). However, most IPV victims stayed with the perpetrator due to economic reasons, learned helplessness, feeling entrapped in the relationship or Stockholm syndrome (Brewster, 2002).Wathen and MacMillan (2003)suggested that, in addition to reoccurrence of violence, other types of outcomes should be examined, such as the impact of clinically based intervention (e.g. counselling, social ser-vices, etc.) and whether the victims’ mental health improves. Their suggestion concurred with the conceptual framework proposed byMears (2003), in that the intermediate outcomes (e.g. improved mental health and family function-ing, more effective decision-making skills and abilities, etc.) need to be achieved first before the long-term goal (e.g. reduced re-victimisation) could be reached. Thus, this study focused on the mental health or well-being of the women who experienced IPV as the outcome.

The intermediate outcomes mentioned above could include the personal, interpersonal or environmental indicators. Conceptually deducted from the

(3)

framework, perceived growth could be one of the indicators, as it might involve mental health, more effective decision-making skills and abilities, etc. The literature has shown that growth against the adversity is possible and common. The perceived growth tends to be demonstrated in three general domains: changes in perception of self, changed relationships with others and a changed philosophy of life that includes a deeper appreciation of life along with new life directions and priorities (Calhoun and Tedeschi, 1998). The review ofFrazier et al. (2001)revealed that, across studies, an average of 50 – 60 per cent of the respondents endorsed some degree of growth or positive changes given the encounter of traumatic events. A quali-tative study of forty-four women who experienced sexual violence revealed that they developed self-reliance over time and could make something good out of something bad (Draucker, 2001). The demonstration study of the strengths perspective bySong and Shih (2010)revealed that the victims of IPV experienced changes of self based on the narratives of the participants. To make the growth possible, empowerment has been considered as a key practice element during the recovery journey for the victims of IPV. For example,Brewster (2002)maintained that ‘evaluation plan should be put into place to assess their effectiveness in empowering women and pre-venting battering’ (Brewster, 2002, p. 43). Empowerment is defined as holding a positive attitude towards one’s own self-efficacy, mastery of envir-onment, and the possibility and ability to influence the outside world, along with taking individual or collective action to fight against oppression and dis-crimination (Adams, 2003;Chadiha et al., 2004;Payne, 1997).Mears (2003) maintained that the victims themselves could be change agents; through the strategies of empowerment, intermediate outcomes could be reached. However, up to today, no study has directly examined this theoretical statement. The importance of empowerment practice has been emphasised in the field of IPV (Peled et al., 2000). It is especially critical for the victims in Taiwan, since many families are still holding patriarchal perspectives and Chinese are relationship-oriented. The female victims of IPV might endure the pain and suppress their feelings to maintain the harmony and integrity of the family (Yang and Hwang, 1991). In Taiwan, some of the case managers did provide supportive services and counselling, and utilise empowerment strat-egies to enhance the victims’ self-efficacy, to strengthen their linkage with others, and to increase the stability of life through employment or support from the informal social network. However, the level of these types of self-enhancement services varies among case managers, since they might just focus on the safety and surviving issues under high caseload. Thus, the major purpose of this study was to explore whether the utilisation of em-powerment strategies is associated with clients’ changes of self from social workers’ perspective.

In addition, the contextual factors of professionals’ practice were taken into account based on the perspective ofPayne (1997).Payne (1997)maintained that social work was a socially constructed profession, and was always

(4)

constructed from three elements: worker, client and context (social service or-ganisation and the society). Thus, the investigator conceptualised these ele-ments as contextual factors and hypothesised that they would not only determine the practice of social workers, but also have to do with the changes of clients. The contextual factors covered in this study include person-al, organisationperson-al, service delivery factors, empowerment-related training and the causes of violence. Specifically, the study aimed (i) to explore whether the utilisation of empowerment strategies would be associated with clients’ changes of self while the contextual factors were controlled for and (ii) to determine the relative importance of the two groups of variables, namely the utilisation of empowerment strategies and the contextual factors, in relation to changes of self through model comparisons.

The utilisation of empowerment strategies as a key element to perceived growth

Mears (2003) proposed a framework of reducing domestic violence re-victimisation. Within the model, victims themselves could act as part of the change agents. Through the strategies of empowerment, education, sanction, treatment and services, intermediate outcomes could be reached. Among the individual outcomes, the ‘decision-making skills and abilities’ and ‘enhanced mental health’ connote perceived growth.Song and Shih (2010)analysed the changes of victims of IPV and conceptualised them into ‘changes of self’, including three layers of the recovery of self: from the awareness of the existence of self (enhanced sense of self, e.g. able to accept myself), to the affirmation of capacity and potential of self (affirmation of self, e.g. know how to better express myself) and further to the demonstration of the strengths of self to enjoy a better life world (action and realisation of self, e.g. enhanced ability of using resources, being able to set goals and making an effort to reach them). Each of them reflected the various levels of recovery as intrapersonal, interpersonal and transpersonal, respectively. Since the concept and content are based on the experiences of the victims of IPV, this study utilised these items as the measure of changes of self.

Empowerment can be demonstrated in intrapersonal, interpersonal and social political levels (Song, 2006). The investigator (Song, 2008) has devel-oped empowerment strategies through synthesising the literature and the data from a qualitative study. The findings yielded sixteen strategies, thirty-eight tactics and sixty-eight detailed tactics. The list is thorough and fruitful concerning the specific strategies under Chinese culture (see Table1for the dimensions). Therefore, this study adopted the list of strategies and constructed a scale to measure the extent of utilisation.

Empirically, no study to date has directly examined how the use of em-powerment strategies is associated with perceived growth in the field of IPV. In psychiatric service, the empowering participation experience in a

(5)

consumer-run organisation was found associated with recovery. An empow-ering participation experience is defined as involvement in leadership roles and contribution to organisational functioning (Brown et al., 2008). The find-ings ofCorrigan (2006)revealed that participation in peer support (the proxy of consumer-operated services) was associated with recovery and empower-ment, although the association was low. However, the correlation between empowerment and recovery was not tested in his study.

Deduced fromMears (2003)and the existing literature, the investigator hypothesised that the utilisation of empowerment strategies would have a positive association with clients’ changes.

The contextual factors of changes

Payne (1997)maintained that social work was a socially constructed profes-sion, and was always constructed from three elements: worker, client and context. In creating a profession, many social forces are involved, such as changing social needs, the influence and needs of related occupations, polit-ical and legislative changes, and academic development of the profession. These forces affect the expectations on what professionals should do. Deduced from Payne’s (1997)argument, the following contextual factors could influence the service delivery and the potential impacts on clients: (i) workers’ professional training, work experience and characteristics; (ii) clients’ needs; (iii) the value and philosophy, resources and service

Table 1 Bivariate analysis between the utilisation of empowerment strategies and clients’ change of self

Utilisation of empowerment strategies Association

Total score r ¼ 0.56***

General empowerment strategy (Cronbach’s a ¼ 0.95)

F1: Facilitating alliance & action through group action r ¼ 0.26*** F2: Facilitating choice & action—reframing & providing options r ¼ 0.50***

F3: Recognising own ability—strengths perspective r ¼ 0.56***

F4: Activating own ability—peer influence r ¼ 0.42***

F5: Facilitating understanding of alternatives & choices—discourse r ¼ 0.48***

F6: Raising collective consciousness r ¼ 0.18**

F7: Facilitating action—strength perspective r ¼ 0.40***

F8: Activating own ability—strengths & support r ¼ 0.45***

F9: Facilitating collective action—positive experiences r ¼ 0.34***

F10: Recognising own ability—reframing r ¼ 0.40***

Specific strategies under Chinese culture (Cronbach’s a ¼ 0.95)

F1: Use family support r ¼ 0.27***

F2: Better use of traditional beliefs and relationships r ¼ 0.26***

F3: Better use of social authority for advocacy r ¼ 0.15*

F4: Incremental empowerment and maintaining harmony r ¼ 0.33***

Cronbach’s a for entire scale ¼ 0.95 * p % 0.05; ** p % 0.01; *** p % 0.001.

(6)

model of an organisation; (iv) and government legislation, regulations and cultural values. In this study, the first, second and third factors were dealt with. The fourth factor is the prevailing and influential structure, which is treated as the background of the social work profession in Taiwan.

This study was part of a larger study on the service delivery model and effectiveness on women who experienced IPV. Previously, the investigator had conducted analysis on the correlates of changes of self and examined the mediating effect of empowerment by using the data collected from the clients (Song, 2012). The model included the demographic variables, social support, coping strategy, variables related to partner abuse, social service, in-tensity of contact and empowerment. The findings relevant to this study are that empowerment was directly associated with changes of self and was the most important correlate. The mediation effect of empowerment was con-firmed. In addition, a good relationship with social workers directly and indir-ectly correlated with changes of self. Intensity of service contact and the control tendency of the perpetrator were also significant correlates.

In addition, the investigator had examined the extent and correlates of the utilisation of empowerment strategies using the data from social workers (Song, 2011). Cultural consideration and interpretation were emphasised in that study. In the multivariate analysis, life satisfaction, control tendency of the perpetrator, service provision and intensity of contact (two to three times a month) were found significant, with life satisfaction being the most important one. This study further examined the association between the util-isation of empowerment strategies and clients’ changes of self from the

Figure 1 Conceptual framework for the analysis

(7)

evaluation of social workers by taking into account the contextual variables. The conceptual framework is depicted in Figure1. The investigator hypothe-sised that both the utilisation of empowerment strategies and contextual vari-ables would be directly associated with clients’ changes of self. Also, based on a previous study (Song, 2011), the contextual variables might have indirect effects on clients’ changes of self through their association with the utilisation of empowerment strategies.

Method

Research design

This study aimed to explore the association between the utilisation of empowerment strategies and clients’ changes of self from workers’ perspec-tive. It was approved by the review board of the National Science Council in Taiwan for the quality of this study and research ethic (NSC 96–2412-H-004 -022-SS2). The participants were informed about the purposes of this study and the anonymity of the data utilisation, and that their responses would not affect their rights and benefits in the agency. Each participant signed a consent form on a voluntary basis. A survey was conducted to gather the data of variables in Figure1. In this study, cause of violence and the dependent variable, clients’ changes of self, were worker’s cognitive evaluation based on their overall clinical experiences with victims of IPV. They were not meant to represent clients’ evaluation. It was certainly not an ideal way of measurement from a scientific standard and might create more measurement error in terms of client changes. To link worker variables with client variables, experimental design would ensure better internal validity. Such design, however, could only be done on a few cases and it requires that the investigator have control over the operation of the service, which is less feasible in the field of domestic violence in Taiwan. Direct observation of the practice is also deterred due to the fact that the protection of clients’ anonymity and confidentiality are very much empathised in Taiwan. Besides, such observation might take place only in small scale even though it is possible. Considering the exploratory nature of this study, the feasible design for the investigator was survey, as it could capture information from a larger sample of participants. The investigator assumed that usually clients have little awareness of which empowerment strat-egies the workers utilised. On the other hand, social workers constantly observe and informally evaluate clients’ changes while working with clients. As a matter of fact, most workers rely on professional judgement in determin-ing the termination of a case without usdetermin-ing any standardised measure. Workers’ evaluation may not be as valid as a client’s own; however, it has its own value in practice and still reveals some information about their clients. Moreover, scales or multiple indicators were used in the measurement of all the variables in question, which could increase the reliability and validity.

(8)

Participants

This study used a cross-sectional design of survey on social workers in the field of partner abuse. The investigator contacted the Center for DVPI in each city and county and asked for their co-operation on the survey. There are twenty-four centres already operating around Taiwan. Some of the services are con-tracted out to the private sector. As a result, twenty-three centres and eight private organisations agreed to participate in this study, and the total number of social workers available was 316. A sample of 265 social workers was recruited on a voluntary basis. It took three months to conduct the survey and yielded 243 completed and valid questionnaires. Among them, only twenty-six participants in this study were workers in the private sector. The response rate represented was 91.7 per cent of 265. Each participant was given a gift voucher (worth $NTD200 or $USD6) to a convenience store.

Variables and instrument

The data were collected through a self-administered questionnaire. Partici-pants’ confidentiality and anonymity were ensured in the introduction of the questionnaire. In this study, one dependent variable and five groups of in-dependent variables were included.

Dependent variable

Changes of self: Sixteen items were designed, based on the findings ofSong and Shih (2010), to capture this concept. It was a Liker-type scale, with four re-sponse categories: never (1), seldom (2), sometimes (3) and often (4). Principal factor analysis using oblique rotation was conducted to explore the factor struc-ture. As a result, two factors were extracted (eigenvalue ^ 1) with 54.63 per cent variance explained. All the factor loadings were greater than 0.42. The first factor (twelve items) combined two of the original dimensions developed bySong and Shih (2010), which were affirmation and actualisation of self. The second factor (four items) confirmed the original dimension of enhanced sense of self. The internal consistency (Cronbach’s a) for the entire scale was very high (0.92). The a for both factors were also satisfactory (0.90 and 0.82). The results showed that the scale has good psychometric properties. A summation score was computed and used in the analyses.

Independent variables

The list of independent variables and the internal consistency (Cronbach’s a) for some of the variables for the current study can be seen in Table2.

(9)

1. Empowerment strategies and tactics extracted by a study bySong (2008)

were constructed into a four-point Likert-type scale. The response categor-ies were: never (1), seldom (2), sometimes (3) and often (4). The scale com-prised sixty-eight items, including fourteen specific strategies under Chinese cultural. The factorial construct validity has been demonstrated and the factor structure was as depicted in Table1with three items deleted (Song, 2011). The internal consistency of the scale is satisfactory (see Table1for Cronbach’s a). The mean summation score was computed out of sixty-five items to indicate the level of the utilisation of empowerment strategies.

Table 2 Bivariate analysis between contextual variables and change of self

Independent variables Association

Personal factors

Total work tenure r ¼ 0.13*

Life satisfaction (Cronbach’s a ¼ 0.79) r ¼ 0.30***

Organisational factors

Role opportunity and support (Cronbach’s a ¼ 0.79) r ¼ 0.04

Service resources (Cronbach’s a ¼ 0.87) r ¼ 0.21**

Professional support (Cronbach’s a ¼ 0.86) r ¼ 0.14*

Total caseload r ¼ 0.05

Cases contacted within past three months r ¼ 0.03

Service delivery factors Frequency of contact (%)

2 – 3 times a week r ¼ 0.04

Once a week r ¼ – 0.03

2 – 3 times a month r ¼ 0.14a

Once a month r ¼ – 0.07

Once every two months r ¼ 0.04

Once every three months r ¼ – 0.03

Ways of contact (%)

Telephone contact r ¼ – 0.05

Interview in agency r ¼ – 0.05

Outreach r ¼ – 0.11

Level of service provision (Cronbach’s a ¼ 0.58) r ¼ 0.30***

Causes of violence

Differences in opinion r ¼ 0.19**

Poor coping mechanism r ¼ 0.20**

Control tendency of perpetrator r ¼ 0.23***

Behavioural problem of the perpetrator r ¼ 0.05

Violent tendency of the perpetrator r ¼ 0.06

Conflict due to extramarital affair r ¼ 0.15*

Empowerment-related training Attend conferences No (n ¼ 77) t ¼ – 2.13* Yes (n ¼ 152) Attend seminars No (n ¼ 68) t ¼ – 2.16* Yes (n ¼ 159 Read books/articles No (n ¼ 57) t ¼ – 2.44* Yes (n ¼ 172) * p % 0.05; ** p % 0.01; *** p % 0.001;a_{p ¼ 0.055.}

(10)

2. Personal factors included two variables: total work tenure (by month) and life satisfaction. Life satisfaction was measured by a scale of seven items developed bySong (2006). The scale captured the level of satisfaction on various life aspects of living: status, work, finance, interpersonal relations, children’s status, self-competence and external environment. Items were phrased in ways such as: ‘I am satisfied with my living status.’ The partici-pants rated each item among four categories: strongly disagree (1), disagree (2), agree (3) and strongly agree (4). The external construct validity of this scale was ensured by its significant correlation with level of empowerment (g ¼ 0.66) and the Cronbach’s a was 0.80 (Song, 2006).

3. The organisational factors included four variables. First, role opportunity and support were measured by five items (Song, 2006). The items measured the opportunity of social workers to express their own opinion in role allo-cation, clear statement of responsibility, level of discretion power, receiving support when needed and mutual help among workers. The response cat-egories were: never (1), seldom (2) and often (3). The a in the previous study was 0.83 (Song, 2006).

Second, service resource was measured by six types of service. Those six types were: consultation (e.g. legal, medical), support (e.g. support group), assistance (e.g. legal, financial), shelter, rehabilitation and counselling, and courses for learning and growth. The participants rated the sufficiency of resources among five levels, ranging from very insufficient (1) to very suf-ficient (5).

Third, eight items were developed to measure professional support from different aspects. These items were: internal and external supervision for case treatment, the director’s attitudes toward content, depth of treatment and outcomes of treatment; sufficiency of manpower, administrative support and training opportunity. The response categories ranged from very insufficient (1) to very sufficient (5). The internal consistency was very satisfactory.

The fourth variable was caseload of each worker, including total caseload and cases contacted within the past three months.

4. Service delivery factors were measured by three indicators. (1) Frequency of contact. The participants rated among their clients the percentage of contact two to three times a week, once a week, two to three times a month, once a month, once every two months and once every three months, respectively. (2) Ways of contact. The participants rated the percentage of times they con-tacted clients by telephone, meeting in agency or outreach, respectively. (3) Level of service provision. The participants rated the extent of the six types of service (as listed in the service resources variable) that they pro-vided to clients, with four response categories: never (1), seldom (2), some-times (3) and often (4). The a for the six items was 0.58. This unsatisfactory result might indicate that the services provided were individualised and not

(11)

necessarily correlated. The mean summation score was used to indicate the extent of service provision.

5. Empowerment-related training. Three items were designed for this variable. Participants were asked whether they had attended empowerment-related conferences or seminars, or had read related articles or books during the past three years. The response category was yes/no. If the answer was yes, then they were further asked to indicate the amount of participation. Since the latter part was only applicable for those who had participated in such kind of training, the number of valid cases ranged from 110 to 144. To retain more cases for analysis, only the three dichotomous items were used in the analyses. 6. Causes of violence were measured by six types (twelve items): (1) differ-ences in opinion (conflict out of discordance, poor communication), (2) poor coping mechanism (ineffective coping methods, poor emotional control), (3) control tendency of the perpetrator (wanting to control the partner, emotional dependency, lack of trust towards the partner, chauvinist attitude), (4) violent tendency of the perpetrator, (5) behavioural problem of the perpetrator (loss of control due to substance abuse or mental illness) and (6) conflict due to an extramarital affair. The participants were asked to rate how frequently each cause of violence occurred among the clients they served based on their experiences. The four response cat-egories for each item were: never (1), seldom (2), sometimes (3) and often (4). While violent tendency and conflict due to a marital affair were mea-sured by a single item, the other four types were indicated by a summation of various items.

Data analysis Correlation analysis

To test the bivariate correlation between each independent variable and the dependent variable, a t-test was used for dichotomous independent variables (empowerment-related training), a Spearman correlation was used for ordinal independent variables (violent tendency of the perpetrator and con-flict due to marital affair) and a Pearson correlation was used for a continuous independent variable.

Multiple regression analysis

Simultaneous method of entry was utilised to examine the significant corre-lates of the utilisation of empowerment strategies while other independent variables were taken into account. In addition, to examine the relative im-portance of the utilisation of empowerment strategies and the contextual variables, model comparisons were performed. First, Model 1, with only the utilisation of empowerment strategies, was entered into the equation of

(12)

regression. Second, Model 2 entered only the indicators of contextual vari-ables into the equation. Third, both groups of independent varivari-ables were entered into the equation (Model 3). The unique contribution of the utilisa-tion of empowerment strategies was calculated by using the R2of Model 3 minus that of Model 2. The unique contribution of the contextual factors was using the R2of Model 3 minus that of Model 1. The joint contribution of both groups of independent variables was using the R2of Model 3 minus the unique R2of the utilisation of empowerment strategies and that of con-textual variables.

Results

Sample characteristics

The mean age of the participants was 31.28 years (SD ¼ 7.18, range ¼ 23 – 56) and the majority of them were female (86.8 per cent). Most of them (88.5 per cent) held a college degree and 10.7 per cent of them had a master’s degree. About two-thirds of the participants (67.5 per cent) earned a monthly salary between $NTD30,000 and $NTD39,999, followed by $NTD40,000 – $NTD49,999 (16.5 per cent), and $NTD50,000 and above (2.5 per cent). The mean of total work tenure was about five years (62.58 months, SD ¼ 66.54). The average work tenure in the field of partner abuse was three years (36.31 months, SD ¼ 37.74), with 52.5 per cent less than two years. On average, participants had worked in the current agency for 40.83 months (SD ¼ 56.85), with 52.7 per cent less than two years. Regarding the sample representativeness, there are no official statistics available con-cerning the characteristics of social workers in the field of domestic violence for comparisons. As compared with the sample of a survey on 130 social workers and social administrators from public sectors in Taiwan (Chang et al., 2003), the participants in this study were much younger, comprising more females, earned less and more of them had a college degree and fewer had a master’s degree. No data on work tenure were available in their data for comparison. Nevertheless, the field of domestic violence is known to be highly demanding with a high turnover rate. That might be the reason why the workers tended to be younger and earned less than the workers in other public sectors.

Changes of self

The participants perceived clients experienced more change on ‘enhanced sense of self’ (mean ¼ 3.11, SD ¼ 0.61) than on ‘affirmation and actualiza-tion of self’ (mean ¼ 3.06, SD ¼ 0.41). The mean rank among the items ranged from 2.59 to 3.33. Higher mean rank was observed on the following

(13)

items: enhanced ability of using resources (3.33), have learned how to take better care of themselves (3.32), have learned how to let go and live with the problem (3.21), are able to accept themselves (3.16), know how to better express themselves (3.16), enhanced coping ability (3.12), willing to change and action increased (3.12), no more self-blame for the violence that happened to them (3.00), understand the value of their life (2.97). The lowest mean rank was observed on ‘feel that they are capable of helping others’ (2.59) and ‘able to expand their life domain and interpersonal relationship’ (2.87).

Correlations between the utilisation of empowerment strategies and clients’ changes of self

The total score of the utilisation of empowerment strategies had a medium high correlation with clients’ changes of self (r ¼ 0.56) (see Table1). Among the factors of empowerment strategies, ‘recognizing own ability—strengths per-spective’ had the highest association with the changes of self (r ¼ 0.56), ‘facili-tating choice and action—reframing and providing options’ being the second (r ¼ 0.50), ‘facilitating understanding of alternatives and choices—discourse’ being the third (r ¼ 0.48), ‘activating own ability—strengths and support’ being the fourth (r ¼ 0.45) and ‘activating own ability—peer influence’ being the fifth (r ¼ 0.42). These are all related to the intrapersonal empowerment and strengths perspective. The association between ‘raising collective con-sciousness’ and changes of self was low (r ¼ 0.18), which might have to do with the fact that it was the least-used strategy (mean ¼ 1.69, SD ¼ 0.66). The utilisation of specific strategies under Chinese culture exerted medium low association with changes of self, with ‘incremental empowerment and main-taining harmony’ being the highest (r ¼ 0.33) among the items (see Table1).

Correlations between contextual variables and changes of self The bivariate analyses revealed that twelve contextual variables were corre-lated with changes in self, including the two personal factors (total work tenure and life satisfaction), service resources and professional support in organisational factors, level of service provision in service delivery factors, four of the causes of violence (differences in opinion, poor coping mechanism, control tendency of the perpetrator and conflict due to extramarital affair) and three empowerment-related training variables (see Table2). The associa-tions were medium to weak. Among them, life satisfaction and level of service provision each had the highest association (r ¼ 0.30) with changes of self. Those who had attended empowerment-related conferences and courses and read articles or books on empowerment tended to reported more

(14)

clients’ changes of self. All the indicators of frequency of contact and ways of contact were not significant.

Multiple regression analysis on changes of self

To retain more valid cases, only the significant variables in the bivariate analyses were included in the regression analyses, namely thirteen independ-ent variables. ‘Conflict due to extramarital affair’ was an ordinal variable. Thus, two dummy variables were created by using the response category ‘sometimes’ as the reference group. The reason was that there was significant difference between the ‘sometimes’ group and the ‘often’ group in the one-way ANOVA. The dependent variable, ‘changes of self’, was normally distributed based on the Kolmogorov–Smirnov test (p ¼ 0.200, skewness ¼ –0.524, kurtosis ¼ 1.044). One outlier was deleted based on the result of the initial ana-lysis. Thus, 219 cases were retained for anaana-lysis. No serious multicollinearity existed among the independent variables with the tolerance coefficient ranging from 0.58 to 0.91.

As can be seen in Table3, the utilisation of empowerment strategies alone explained 30 per cent of the variance in changes of self (Model 1). When only the contextual variables (Model 2) were entered, 23.4 per cent of the vari-ance in changes of self was explained. Model 3, with both the utilisation of em-powerment strategies and the contextual variables, yielded 37.3 per cent explained variance in changes of self. Thus, the unique variance explained by the utilisation of empowerment strategies was 13.9 per cent (37.3– 23.4 per cent), whereas the unique variance explained by the contextual vari-ables was 7.3 per cent (37.3–30.0 per cent) and the joint contribution by the two groups of variables was 16.1 per cent (37.3–13.9–7.2 per cent) (see Figure1). Moreover, Table4shows that the contextual variables had signifi-cant association with the utilisation of empowerment strategies (R2¼ 0.267; p , 0.05), which indicated that the contextual variables might contribute to clients’ changes of self through their effects on the utilisation of empowerment strategies. And the main effects came from the four significant variables: life satisfaction, service provision, poor coping mechanism and control tendency of perpetrator. Social workers’ life satisfaction and more service provision positively correlated with the utilisation of empowerment strategies. When the causes of violence were due to the perpetrators’ poor coping mechanism or control tendency, social workers tended to used more empowerment strat-egies and in turn might be conducive to clients’ changes of self.

When all thirteen variables were simultaneously entered into the equation (Model 3), only two variables appeared to be significant: the utilisation of em-powerment strategies and life satisfaction. Both exerted positive effects on clients’ changes of self, with the utilisation of empowerment strategies being the most important correlate based on the standardised regression co-efficient (beta ¼ 0.435), life satisfaction being the second (beta ¼ 0.182).

(15)

Table 3 Results of multiple regression analysis on clients’ change of self (valid cases ¼ 219)

Independent variables b SE of b Beta t-value p-value

Model 1: Empowerment strategies

Constant 1.249 0.191 6.532 0.000

Empowerment strategies 0.670 0.069 0.548 9.654 0.000

R2_{¼ 0.300; Adjusted R}2_{¼ 0.297; F(1, 217) ¼ 90.21; p ¼ 0.000}

Model 2: Contextual variables

Constant 0.307 0.411 0.749 0.455

Personal factors

Total work tenure 0.000 0.000 – 0.045 – 0.669 0.504

Life satisfaction 0.341 0.085 0.268 3.996 0.000 Organisational factors Service resources 0.045 0.044 0.072 1.005 0.316 Professional support – 0.036 0.055 – 0.048 – 0.650 0.516 Service delivery Service provision 0.111 0.082 0.096 1.363 0.174 Causes of violence Differences in opinion 0.110 0.065 0.107 1.674 0.096

Poor coping mechanism 0.100 0.069 0.102 1.441 0.151

Control tendency of perpetrator 0.099 0.067 0.110 1.477 0.141

Conflict due to extramarital affair

Seldom (1) 0.109 0.070 0.112 1.561 0.120 Often (1) 0.167 0.074 0.160 2.263 0.025 Empowerment training Attend conferences – 0.043 0.077 – 0.044 – 0.554 0.581 Attend seminars 0.118 0.080 0.118 1.477 0.141 Read books/articles 0.099 0.069 0.094 1.430 0.154 R2_{¼ 0.234; Adjusted R}2_{¼ 0.185; F(13, 205) ¼ 4.81; p ¼ 0.000}

Model 3: Empowerment strategies 1 contextual variables

Constant 0.056 0.374 0.149 0.881

Empowerment strategies 0.532 0.079 0.435 6.723 0.000

Personal factors

Total work tenure 0.000 0.000 – 0.048 – 0.782 0.435

Life satisfaction 0.233 0.079 0.182 2.941 0.004 Organisational factors Service resources 0.045 0.040 0.072 1.113 0.267 Professional support – 0.023 0.050 – 0.030 – 0.447 0.655 Service delivery Service provision 0.005 0.076 0.005 0.072 0.943 Causes of violence Differences in opinion 0.089 0.059 0.088 1.505 0.134

Seldom (1) 0.087 0.063 0.090 1.378 0.170 Often (1) 0.131 0.067 0.125 1.944 0.053 Empowerment training Attend conferences – 0.021 0.070 – 0.022 – 0.298 0.766 Attend seminars 0.083 0.073 0.082 1.136 0.257 Read books/articles 0.063 0.063 0.060 1.001 0.318 R2_{¼ 0.373; Adjusted R}2_{¼ 0.330; F(14, 204) ¼ 8.66; p ¼ 0.000}

(16)

Worker’s life satisfaction had both direct effects and indirect effects (0.085) via the utilisation of empowerment strategies on changes of self (see Table5). Service provision, perpetrators’ poor coping mechanism and perpe-trators’ control tendency each also had indirect effects on clients’ changes of self. The indirect effects were calculated by multiplying the beta of the variable on the utilisation of empowerment strategies with the beta of the util-isation of empowerment strategies on clients’ change of self (0.435). For example, the indirect effect of life satisfaction on clients’ changes of life was 0.085¼ 0.195× 0.435. In terms of total effect, the utilisation of empower-ment strategies was the most important correlate of clients’ changes of self,

Table 4 Results of multiple regression analysis on the utilisation of empowerment strategies (valid cases ¼ 219)

Independent variables b SE of b Beta t-value p-value

Constant 0.472 0.328 1.438 0.152

Total work tenure 0.000 0.000 0.006 0.095 0.924

Life satisfaction 0.204 0.068 0.195 2.984 0.003 Organisational factors Service resources 0.000 0.035 – 0.001 – 0.012 0.991 Professional support – 0.025 0.044 – 0.041 – 0.574 0.566 Service delivery Service provision 0.199 0.065 0.210 3.044 0.003 Causes of violence Differences in opinion 0.038 0.052 0.045 0.721 0.472

Seldom (1) 0.040 0.056 0.051 0.726 0.469 Often (1) 0.069 0.059 0.080 1.160 0.247 Empowerment training Attend conferences – 0.041 0.062 – 0.052 – 0.663 0.508 Attend seminars 0.067 0.064 0.081 1.042 0.299 Read books/articles 0.068 0.055 0.079 1.218 0.225 R2_{¼ 0.267; Adjusted R}2_{¼ 0.220; F(13, 205) ¼ 5.74; p ¼ 0.000.}

Table 5 The significant correlates of clients’ changes of self: direct and indirect effects

Independent v.

Utilisation of

empowerment strategies Clients’ changes of self

Effects Directa _Direct _Indirect

Life satisfaction 0.195 0.182 0.085

Service delivery

Service provision 0.210 0.091

Causes of violence

Poor coping mechanism 0.154 0.067

Control tendency of perpetrator 0.173 0.075

Empowerment strategies NA 0.435

a_{The direct effect in the table is standardised regression coefficient.}

(17)

followed by life satisfaction, service provision, control tendency of the per-petrator and poor coping mechanism of the perper-petrator.

Discussions and implications

Worker-perceived clients’ changes of self on multiple domains In this study, social workers reported multiple domains in clients’ changes of self, which confirmed the synthesis ofCalhoun and Tedschi (1998)and the findings ofSong and Shih (2010). Social workers reported more changes on clients’ enhanced sense of self, such as being able to accept themselves, no more self-blame for the violence, understanding the value of their life. Such enhancement is essential as the victims of IPV might start to regain their self-esteem, dignity and value of their life, and in turn find the strengths to fight against violence and search for other options for life. Growth on af-firmation of self in the women’s experiences of IPV was also frequently observed, such as self-care, coping ability, to express self, etc. These are related to the growth of ability and self-assertiveness. Relatively speaking, social workers reported less growth on actualisation of self, such as being pro-active towards life, setting and achieving goals, and expanding interpersonal relationships, etc. To further facilitate such changes, more services and more use of empowerment strategies on interpersonal and social political levels might be necessary.

The critical avenue to changes—the utilisation of empowerment strategies

The essential role of the utilisation of empowerment strategies in facilitating clients’ changes of self was supported by the analysis. It was the most import-ant correlate while the contextual variables were taken into account. It served as the direct correlate (R2¼ 13.9 per cent) and also the mediator of context-ual variables in relation to perceived clients’ changes (R2¼ 16.1 per cent). Such findings confirmed the theoretical framework proposed by Mears (2003). This study revealed specific empowerment strategies that had higher correlation with clients’ changes of self. These strategies are related to the strengths perspective and reframing in narrative analysis. The utilisa-tion of the strengths perspective could help clients in recognising and activat-ing their own ability, facilitatactivat-ing understandactivat-ing of alternatives and choices, and further facilitating actions. Reframing and providing options were con-ducive for clients to recognise their own ability, make choices and take action. The utilisation of peer influences to encourage clients was also helpful to changes of self.

(18)

Among the specific strategies under Chinese cultural, incremental em-powerment and maintaining harmony could facilitate more clients’ changes as the Chinese emphasise both personal right and responsibility in society (Yang and Hwang, 1991;Yip, 2004). Chinese people are relationship-oriented. They care very much for their status within society; thus, most of them tend to be submissive to the norms of society and sensitive to the defined social roles and obligations (Yang and Hwang, 1991). Chinese philosophy also emphasises the complementary and dynamic balance between opposites—in other words, the everlasting rhythm and harmony of the person, social environment, nature and the universe. Thus, Chinese people might not take radical actions towards oppression or regressive isolation. Usually, they endure the oppression and pain and in the mean time find an indirect way to make changes (e.g. via influ-ence of another authoritative figure or special relationship) or wait for the op-portunity to come. The women who experienced violence might face patriarchal suppression from the perpetrator, their in-laws or the family of origin. The empowerment strategies need to take into account how the clients perceive their role in and relationship with the social network. Since maintaining their status and harmony within the society and social network is very important, the empowerment process usually takes time to awaken the client’s sense of autonomy, reframe the situation, learn new coping strat-egies and how to better express themselves, make coalitions with other family members, and increase their own resources and be more independent. Thus, the empowerment process needs to be incremental.

The findings revealed the strong association between the use of the strengths perspective and narrative analysis and worker-perceived clients’ changes of self, with reframing and peer influence emphasised. Thus, the professional knowledge and implementation of these approaches might be necessary to fa-cilitate clients’ changes of self, which are worthy of consideration in future training in the field of domestic violence. The data of this study showed that the percentage of participants had attended empowerment-related confer-ences and seminars, or read related articles or books during the past three years ranged from 67.5 to 75.8 per cent. The mean number of participation for each type was 2.56 conferences, 17.62 hours of seminars, and 4.63 articles and 2.15 books. The results revealed the potential needs of such training in terms of percentage of participants and quantity. In addition, attention should be put on the content of training and the application of relevant knowl-edge and skills in practice.

Importance contextual correlates: life satisfaction, service provisions and causes of violence

The findings of this study support the social construct theory of the social work profession maintained by Payne (1997). Three contextual variables appeared to be important correlates of clients’ changes through the

(19)

correlation with the utilisation of empowerment strategies: workers’ life sat-isfaction, service provision and causes of violence (control tendency, poor coping mechanism). Social workers’ life satisfaction was the second most im-portant correlate of clients’ changes of self. It had direct effects and indirect effects on such changes. The finding confirms the argument byAdams (2003) in that self-empowerment of social workers seems necessary before empow-ering other people. Obviously, social workers’ thoughts, feelings and situ-ation would affect how they work with clients. Social service appears to be a mutually beneficial process. Especially, ‘Chinese culture emphasizes culti-vating one’s inner self as well as caring for others and contributing to society; thus an individual self and his moral judgement have become embedded in society (Yang and Hwang, 1991)’ (Song, 2011, p. 1034). It implies a happy person would have the strength and desire to inspire others and make the world a better place. Therefore, human service organisations need to imple-ment employee assistant programmes (EAP), as workers in the field of do-mestic violence usually face tremendous pressure from heavy workload, vicarious trauma and threats to clients’ safety. The EAP might include self-care and adjustment, self-empowerment, counselling, emotional and instru-mental support when needed. Based on the investigator’s observation in practice and research finding (Song and Shih, 2010), the strengths perspective is conducive to workers’ self-empowering and recovery. Thus, the supervisors in this field might consider adopting a strengths-based supervision and to fa-cilitate the applications of the model onto workers themselves as well as clients, although the effects need to be further examined. Caseload needs to be reduced for such an approach to be pervasively implemented.

The findings revealed that more service provision (types and extent) might increase the opportunity of using empowerment strategies, in turn enhancing clients’ changes of self. Thus, it seems logical that services themselves are vehicles to exert influence on and increase the chance of contacts with clients, and the chance of using empowerment strategies might also increase. Moreover, the utilisation of empowerment strategies was based on clients’ needs. The higher social workers rated perpetrators having control tendency or poor coping mechanism (ineffective coping or poor emotional control) as the causes of violence, the higher the chance they used empowerment strat-egies. The study byDutton and Starzomski (1993)revealed that the perpetra-tors with borderline personality correlate strongly with wives’ reports of psychological abuse but correlate less with physical abuse. Intermittent undermining of the significant other, manipulation and masked dependency are part of the essential characteristics of borderline personality (Dutton and Starzomski, 1993); thus, the investigator speculates that these perpetrators might tend to use psychological abuse, which would have a negative impact on victims’ self-esteem and cause sense of powerlessness. However, the mechanism still needs further study to clarify.

(20)

Contributions and limitations

This is the first study that empirically examined the association between the utilisation of empowerment strategies and clients’ changes of self from social workers’ perspective in Taiwan. It confirmed the critical and essential role of this variable. However, by using a survey design and data collected based on the ratings of social workers, this study has the following limitations: (i) the findings only imply association instead of causal relationship between the vari-ables; (ii) it was not a direct evaluation of the impact of an empowerment pro-gramme on clients; (iii) the recruitment of participants was on a voluntary basis, thus the external validity is minimised; (iv) the measurement of causes of violence and clients’ changes of self were aggregate and cognitive evaluation based on practice experiences and observations, which could increase meas-urement error. The sources of error might stem from recall error, lack of accur-ate information and misperception or misattribution, which varies among workers. The error might be large; however, the investigator had no control of it. Especially, the attribution of causes of violence has been controversial over patriarchal perspective versus system perspective or personality disorder (Dutton, 2008). As can be seen, there are six types including twelve causes listed in the measurement section of this paper. The ratings depended on the level of social workers’ knowledge, personal values and grasp of information. Nevertheless, this is a limitation that is hard to avoid when surveying a large sample of social workers. Besides, the previous study based on data collected from clients also confirmed the association between empowerment and changes of self (Song, 2012). Adding these evidences together might increase the confidence on internal validity of the findings. The investigator hopes that the results of this exploratory study could provide some preliminary infor-mation for future study. The contribution of this study would be on calling for more attention on facilitating the utilisation of empowerment strategies, implementing an employee assistant programme to enhance workers’ life sat-isfaction and providing more services to the victims of IPV as they might be conducive to clients’ changes of self.

Acknowledgements

This study was funded by the National Science Council (NSC 96 – 2412-H-004 -022 -SS2). The author is very grateful for the funding and for all the partici-pants in this study. In addition, the author appreciates very much the help from Dr Chaiw-yi Shih, for the review and suggestions on the manuscript.

References

Adams, R. (2003) Social Work and Empowerment, 3rd edn, New York, Palgrave MacMillan.

(21)

Brewster, M. P. (2002) ‘Domestic violence theories, research, and practice implications’, in A. R. Roberts (ed.), Handbooks of Domestic Violence Intervention Strategies, London, Oxford University Press.

Brown, L. D., Shepherd, M. F., Merkle, E. C., Wituk, S. A. and Meissen, G. (2008) ‘Under-standing how participation in a consumer-run organization relates to recovery’, American Journal of Community Psychology, 42, pp. 162– 78.

Calhoun, L. G. and Tedeschi, R. G. (1998) ‘Beyond recovery from trauma: Implications for clinical practice and research’, Journal of Social Issues, 54(2), pp. 357 –71.

Chadiha, L. A., Adams, P., Biegel, D., Auslander, W. and Gutierrez, L. (2004) ‘Empower-ing African women informal caregivers: A literature synthesis and practice strategies’, Social Work, 49(1), pp. 97 – 108.

Chang, R., Lin, W. I. and Wang, Y. T. (2003) A Comparative Study on Social Work Profes-sional System among the World and Survey on the Needs, Utilisation, and Coping Methods for Difficulties of Social Work Manpower in Taiwan, Taipei, Taiwan, Ministry of the Interior.

Corrigan, P. (2006) ‘Impact of consumer-operated services on empowerment and recovery of people with psychiatric disabilities’, Psychiatric Services, 57(10), pp. 1493 – 6. Draucker, C. B. (2001) ‘Learning the harsh realities of life: Sexual violence,

disillusion-ment, and meaning’, Health Care for Women International, 22, pp. 67 – 84.

Dutton, D. G. (2008) ‘Reflections on thirty years of domestic violence research’, Trauma, Violence, & Abuse, 9(3), pp. 131 – 43.

Dutton, D. G. and Starzomski, A. J. (1993) ‘Borderline personality in perpetrators of psy-chological and physical abuse’, Violence & Victims, 8, pp. 327 – 37.

Frazier, P., Conlon, A. and Glaser, T. (2001) ‘Positive and negative life changes following sexual assault’, Journal of Consulting and Clinical Psychology, 69, pp. 1048 – 55. Lundy, M. and Grossman, S. (2001) ‘Clinical research and practice with battered women’,

Trauma, Violence, & Abuse, 2(2), pp. 120 – 41.

Mears, D. P. (2003) ‘Research and interventions to reduce domestic violence revictimiza-tion’, Trauma, Violence, & Abuse, 4(2), pp. 127 – 47.

Payne, M. (1997) Modern Social Work Theory, 2nd edn, London, MacMillan Press Ltd. Peled, E., Eisikovitz, Z., Enosh, G. and Winstok, Z. (2000) ‘Choice and empowerment for

battered women who stay: Toward a constructivist model’, Social Work, 45(1), pp. 9 – 25. Song, L. (2006) ‘The development and validation of an empowerment scale’, Social Policy

& Social Work, 10(2), pp. 49 – 86.

Song, L. (2008) ‘The strategies and tactics of empowerment practice: The exploration of Taiwan experiences’, Social Policy & Social Work, 12(2), pp. 123 – 94.

Song, L. (2011) ‘The extent and correlates of the utilisation of empowerment strategies: Survey on social workers in the field of partner violence’, British Journal of Social Work, 41(6), pp. 1106– 37.

Song, L. (2012) ‘Service utilization, perceived changes of self, and life satisfaction among women who experienced intimate partner abuse: The mediation effect of empower-ment’, Journal of Interpersonal Violence, 27(6), pp. 1112 –36.

Song, L. and Shih, C. (2010) ‘Recovery from partner abuse: The application of the strengths perspective’, International Journal of Social Welfare, 19, pp. 23 – 32.

Stover, C. S., Meadows, A. L. and Kaufman, J. (2009) ‘Interventions for intimate partner violence: Review and implications for evidence-based practice’, Professional Psychology Research and Practice, 40(3), pp. 223 – 33.

Wathen, C. N. and MacMillan, H. L. (2003) ‘Interventions for violence against women: Scientific Review’, JAMA, 289(5), pp. 589 – 600.

(22)

Wiehe, V. R. (1998) Understanding Family Violence: Treating and Preventing Partner, Child, Sibling, and Elder Abuse, Thousand Oaks, CA, Sage Publications, Inc. Yang, K. and Hwang, K. (eds) (1991) The Psychology and Behaviors of Chinese, Taipei,

Taiwan, Laurel Publication, Inc.

Yip, K. (2004) ‘The empowerment model: A critical reflection of empowerment in Chinese culture’, Social Work, 49(3), pp. 479 – 87.