Abstract
Using semistructured interviews with 388 women under supervision, this study integrates criminal justice and communication theories by investigating gender responsivity and type of support in messages women receive about employment from supervision agents. Informational support was the most frequent form of supportive communication clients received from their agents, and was the only type of supportive communication clients perceived negatively. Women recalled agents’ messages that varied in their sensitivity to the range of women offenders’ needs (child and family care demands, human capital attainment, mental health issues, and substance abuse recovery). Supportive messages that took into account a variety of problems commonly shared by women on probation and parole had positive effects, whereas supportive messages that were relevant to employment, but failed to consider other needs, had negative effects. Results of this work have implications regarding effective support offered by community supervision agents as they discuss employment.
Prior research has shown that probation and parole agents have unique dual-role relationships with their clients, as they are expected to both help and control the people they supervise (Skeem, Encandela, & Louden, 2003; Skeem, Louden, Polaschek, & Camp, 2007). Morash, Kashy, Smith, and Cobbina (2018) found that agents’ communication pattern contributes to an effective dual-role relationship, which is characterized by caring, fairness, trust, and an absence of toughness. Specifically, a conversational communication pattern that is related to a caring style encourages relationships in which clients are independent, share feelings, and contribute to decision making. A conformity communication pattern that is characterized by the supervision agent’s power over the client and by client compliance with directives is high in toughness and low in caring.
Providing support for the influence of agents’ communication on supervision outcomes, research has found that an authoritarian/conformity pattern of supervision agents’ communication is related to the undesirable outcome of client reactance (i.e., acting contrary to directives or suggestions in an effort to restore freedoms). In contrast, conversational communication has the desirable outcomes of increasing client self-efficacy to avoid lawbreaking and lowering recidivism (Smith, Cornacchione, Morash, Kashy, & Cobbina, 2016). Despite this evidence, theory and research developed in the discipline of communication has rarely been used to advance understanding of the effects of different types of messages that people receive from community supervision agents. The aim of the qualitative research described in this article is to examine the content and effect of messages that women receive from supervision agents regarding employment. Whereas studies of dual-role relationships consider the individual level of analysis (i.e., the agent’s communication pattern and relationship style), we examine communication at a microconceptual level, the message received by the client and her assessment of the message’s effects.
Two theoretical frameworks informed the present research. From the discipline of communication, we use the optimal matching model of support, which identifies types of supportive communication best suited for helping people who experience particular types of stressors (Cutrona, 1990). From the discipline of criminology, we use the feminist pathways framework, which acknowledges that women have greater challenges than men who break the law in the areas of mental health, substance abuse, responsibilities for raising children, and human capital deficits (Gehring, 2018; Lynch et al., 2014; Wattanaporn & Holtfreter, 2014). To be consistent with the feminist pathways framework, community supervision should be holistic by addressing the range of needs that are especially common among women. The integrated theoretical frameworks led us to consider discrete messages to determine what type of social support supervision agents provided, to identify examples of messages that are gender responsive by virtue of considering the range of women’s needs, and to link specific message types to positive or negative effects that clients attributed to the messages.
The Focus on Employment
We focus on employment-related communication for three reasons. First, courts and parole boards often require individuals to find and maintain employment as a condition of supervision, and supervision agents are tasked with exerting control over clients to ensure and verify compliance. In many settings, agents also are expected to assist clients in locating and maintaining work. Given the complexity of managing both control and helping objectives, it is useful to document the employment-related messages that clients receive from agents, and to determine what types of messages clients find beneficial.
Second, many women on supervision face gender-specific disadvantages that affect their ability to find and maintain employment, making their job search especially problematic. More than their male counterparts, they have high rates of mental illness, including substance dependence (McPhail, Falvo, & Burker, 2012; Sirdifield, 2012; Staton-Tindall et al., 2015). Mental illness reduces the likelihood of employment, either as a result of symptoms or discrimination faced in the labor market (Cregan, Kulik, & Bainbridge, 2017; Kelloway, 2017). Justice-involved women are also more likely to act as primary or sole caregivers to children and relatives in need of home care (Glaze & Maruschak, 2008). Assisting relatives and providing and procuring child care often limit the time and energy available to find and keep a job (Berman, 2005; Flavin, 2004). Finally, compared with similarly situated men, women on supervision have more human capital deficits (i.e., defined by Boardman, Grove, Perkins, & Shepherd, 2003, as lack of skills or knowledge with economic value). As a result, women lack resume competitiveness in the labor market, and often must participate in additional educational or vocational programs to find stable employment (Bergseth, Jens, Bergeron-Vigwesaa, & McDonald, 2011). Conversations about employment are especially relevant to women’s multiple needs, and therefore serve as an appropriate focus to examine messages to determine whether they address a range of needs.
A third reason for emphasis on the agent–client communication is that studies have found that supervision agents are important sources of social support for women clients (Bui & Morash, 2010; Cobbina, 2010; Holmstrom, Adams, Morash, Smith, & Cobbina, 2017). This may occur because of a lack of supportive social networks. A high proportion of justice-involved women have substance abuse and mental health problems, a combination that is related to small social networks (Skeem, Louden, Manchak, Vidal, & Haddad, 2009). Many justice-involved women have low incomes and limited education, which also are related to their perceptions of a lack of emotional, social, and material support from people in their social networks (Reisig, Holtfreter, & Morash, 2002). Finally, isolation from the networks they had prior to incarceration often has reduced social support for women on parole. One consequence of limited social support is increased difficulty finding employment (Maddy, Cannon, & Lichtenberger, 2015). Given that supervision agents can act as sources of social support, and are also responsible for ensuring their clients meet parole and probation requirements, agents possess a unique opportunity to aid and encourage female clients in their search for work.
Communication Theory
Research on supportive communication at the message level and feminist criminological research on women’s needs provided the basis for research questions. First, we consider the communication theory, the optimal matching model of support (Cutrona, 1990), which conceptualizes social support provided through communication. Common communication forms include emotional, esteem, network, tangible, and informational support. Emotional support consists of comforting and caring communication. Esteem support bolsters a person’s feelings of self-esteem or competence. Network support facilitates participation in a group with members who share common interests or concerns. Tangible support involves a direct offer of services or resources, and informational support includes an offer of advice or guidance. Communicated social support can have a positive, negative, or mixed valence (Burleson, 2003). A positive valence indicates a good outcome, such as increased job search behavior. A negative valence indicates a problematic outcome, often expressed as increased stress or frustration. It is possible that a supportive message can have a mixture of these two effects, such as increased job search behavior and frustration due to lack of success.
The optimal matching model suggests that different stressful situations require different forms of support based on characteristics of the context. Action-facilitating support, such as informational or tangible support, are viewed as less desirable when both actors have limited control over the stressor. In these situations, advice may not seem practical or feasible, and when the provider has limited expertise, the recipient may perceive information as inappropriate or ineffective. When the stress is more severe and both actors have limited control over the stressor, nurturant supports such as emotional and esteem support are more effective (Cutrona & Suhr, 1992). Given the limitations on the employment prospects of people with a felony conviction, and the varying control that supervision agents have over employment outcomes of their clients, identifying the ideal form of support is further complicated (Cutrona, 1990).
Past research has utilized the optimal matching model to describe the nature and effects of supervision agents’ supportive messages regarding clients’ substance use avoidance (Holmstrom et al., 2017). This research found that agents most often provided women with informational support about substance misuse, and they rarely provided tangible and network support. The researchers recommended that agents be trained to provide a greater range of support. No prior research has applied this communication framework to women’s employment-related conversations with supervision agents. However, studies of samples that are not limited to people in the criminal justice system have found positive impacts of supportive communication on job search behavior, which is an important predictor of employment (Holmstrom, Russell, & Clare, 2015; Kanfer, Wanberg, & Kantrowitz, 2001; Russell, Holmstrom, & Clare, 2015). These studies have shown that social support can directly affect job search behavior by providing information about job openings, or by connecting potential employers and employees. Indirectly, it can improve self-efficacy that promotes job search behavior (e.g., creating resumes, networking; Lim, Lent, & Penn, 2016; Maddy et al., 2015). Understanding the types of supportive messages received in agent–client interactions will increase insight into how agent communication can support clients in their job search, and how agents can improve their conversations to maximize job search–related outcomes.
Feminist Criminological Focus on Women’s Needs
Feminist criminologists have emphasized the assessment of common needs of justice-involved women when deciding on appropriate interventions, and researchers working from a variety of theoretical perspectives have supported this idea (Brennan, Breitenbach, Dieterich, Salisbury, & Van Voorhis, 2012; Gehring, 2018; McPhail et al., 2012; Morash 2010; Sirdifield, 2012; Staton-Tindall et al., 2015). The logical follow-up is community supervision that addresses a range of needs. For justice-involved women, these needs are typically not limited to employment, but extend to substance abuse, mental illness, child and family care, and lack of vocational skills. Although research shows the effectiveness of gender-responsive treatment programs (for reviews see Fagan & Lindsey, 2014; Gobeil, Blanchette, & Stewart, 2016; Grella, 2008), evidence of gender responsivity in messages from supervising agents has not been studied. The present research fills this gap by examining gender-responsive communication, as indicated by the consideration of women’s multiple needs outside of employment, in messages women receive from agents regarding work.
As mentioned previously, many of the common problems likely to occur for justice-involved women are pertinent to difficulties with employment. Research suggests that addressing the realities of women’s lives through gender-responsive interventions is fundamental to improved supervision outcomes (Duwe & Clark, 2015; Messina, Grella, Cartier, & Torres, 2010; Wright, Van Voorhis, Salisbury, & Bauman, 2012).
Research Focus
This work utilizes communication theory and a feminist criminological approach to better understand the ways in which agents communicate with their clients. The following questions involve the supportive and gender-responsive messages women receive that are related to employment, and consider the effects of these messages to better understand criminal justice practices and their effects on women.
Method
Setting and Data
Data were collected in Michigan, where the Michigan Department of Corrections (MDOC) has responsibility for supervision of felony probationers and parolees. Supervision agents specialize in working with women. In high-population counties, agents have women-only probation or parole caseloads. In less populous counties, they have mixed caseloads of women on probation and parole or caseloads that include all women under supervision plus some men. Michigan uses a risk and needs assessment, COMPAS (Northpointe, Inc, 2012), which includes a supplement designed to assess needs especially common among women.
The MDOC’s goal is to manage people who have been convicted of a crime in the community if there is no threat to public safety. Counties receive grants for efforts to reduce prison admissions and improve reentry, and the MDOC’s budget supports prerelease and aftercare services for people in prison (Holtfreter & Wattanaporn, 2014). As with a study of probation and parole in any location, specific findings from the present research may not generalize to other settings, for this study, particularly those that emphasize punishment.
The university institutional review board (IRB) approved all protocol for collecting and analyzing data. Data were from the first interview of a longitudinal study of 402 Michigan women on probation or parole. It included the women’s descriptions of the agents’ communication regarding employment and issues especially common for women in the criminal justice system (i.e., limited education, mental illness, substance abuse, parenting) and their descriptions of the effects of this communication. Because the sample only included drug-involved women, they were typically rated high needs/risk, so agents met with them at least bimonthly for about 20 min each time, had phone contacts, and made periodic home visits. Based on communication research showing that brief exposure to patient–doctor interactions can affect individuals (Smith, Atkin, Skubisz, Nazione, & Stohl, 2009), we concluded that the agent–client interactions with this amount of contact were adequate for the study of messages.
Sample
To obtain the sample of women, in 2011, agents from the 16 Michigan counties within an hour and a half drive from the research office were recruited. These counties had 68.5% of the 2011 state population and included rural, suburban, and urban environments. Seventy-three of 77 (94.8%) agents invited to take part agreed. Each participating agent met with a principal investigator to identify eligible clients (i.e., started supervision approximately 3 months prior, supervised by the same agent for 3 months, convicted of a felony, indication of substance misuse). The focus was on substance-involved women, because they constitute the largest subgroup of justice-involved women (Carson & Anderson, 2016; Fearn et al., 2016).
Multiple methods were used to recruit women. With client consent, agents introduced them to an interviewer at the reporting center to hear about the study, and the interviewer carried out the explanation of human subjects’ rights and obtained written consent. With client consent, agents provided phone contact information to research staff who called women to arrange a time to meet and learn about the study. Finally, agents provided women with a project contact card or flyer, so interested women could arrange a time to hear about the study. Of 846 eligible women, 402 took part in the study. Reasons for not taking part were that research staff were not always on site, and some eligible women neither responded to flyers, nor gave agents permission to share contact information. Nonparticipants did not differ significantly from participants in official records of substance use, violations of community supervision conditions, arrests, or convictions; however, slightly more nonparticipants were incarcerated a year after supervision began. The sample for the present analysis was reduced to 388 due to audiotape malfunctions or missing data needed for the present analysis.
Of the 388 individuals, 355 women reported receiving messages from their agents about employment. The majority were White (63.66%, n = 226), with the remaining women identifying as Black (35.77%, n = 127) or Native American (0.56%, n = 2). Participants ages ranged from 18 to 60 years and the average age was approximately 34 years (M = 33.75, SD = 10.47). Regarding work, 57.46% (n = 204) reported they were unemployed but able to work, 18.31% (n = 65) reported full-time work, and 21.7% (n = 78) reported part-time work. The remaining 8.16% (n = 29) said they were currently unable to work due to circumstances such as child care responsibilities, a disability, or attending school. Asked whether employment was a problem for them since supervision began, 38.60% (n = 137) said “yes,” but most women answered “no” (60.85%; n = 216). As shown by the qualitative analysis results below, some of the discrepancy between the proportion of women not working but able to work and the smaller proportion who consider employment a problem is due to women (and some agents) prioritizing other needs.
Procedure
During an in-person interview, clients were asked whether since starting supervision, they and their agents discussed “finding a job, having a job you like, or keeping a job.” Follow-up questions were, “What did the agent do or say?” “Did she or he refer you to any programs or services?” “Did the agent help you get in or get the services?” “Did this make things better for you? How?” and “Did this make things worse for you? How?” The questions about the effect of the discussion referred to the entire employment-related conversation with the supervision agent. Interviewers clarified their questions if the client misunderstood or misinterpreted them.
Coding
Transcripts of the interviews were unitized (i.e., broken into smaller segments for analysis) so that instances of supportive communication about areas of need were segmented from unrelated discussion (Guetzkow, 1950). The same process was used for areas of need relevant to employment challenges (e.g., education, substance use, mental health). A unit was defined as a statement reported as an agent’s comment during a discussion about employment. 1
The preexisting coding protocol, the Social Support Behavior Code (SSBC), was created to identify the five types of supportive communication and subcategories of each type identified by the optimal matching model of social support (Cutrona & Suhr, 1992). For instance, emotional support includes subcategories such as encouragement and listening. Subcategories were used to code the data initially and were then combined into the higher level support categories for analyses. 2 Data were then coded to link effects to specific supportive messages. Each supportive message was coded as linked to a specific effect, or not linked to a specific effect. 3 In some cases, more than one type of supportive communication was connected to a single effect (e.g., both emotional and informational support led to positive effects). 4
Themes relevant to gender responsivity were inductively derived from the data. Some responses were coded into multiple themes. The passages were first coded based on the presence of a range of challenges that are often related to employment problems (e.g., mental health, child and family care). Then passages were coded as either gender sensitive or insensitive. 5
Results
Supportive Communication
To answer the first research question, we investigated the types of supportive communication regarding employment that women received from agents. The most frequent type was informational support, most often in the form of referrals to employment services. For instance, many clients (n = 28), received referrals to a state agency that assists in building resumes and applying for jobs. The second most recurrent informational support subcategory included agents’ suggestions or advice regarding employment. Julie repeated advice she received from her agent: “Look presentable . . . try to explain the felonies before they run your name. You know, don’t lie on the application.” 6 Finally, agents offered informational support through referrals to particular jobs or employment services. Christine described an agent who “goes online and searches for [jobs] herself” to make referrals, and Angela said that her agent “made calls to different places for me,” and then referred her to appropriate employers.
Emotional support was the second most frequently reported category of supportive communication. It includes relationship building, confidentiality, sympathy, understanding, and empathy (Cutrona, 1990). Study participants most often received emotional support in the form of encouragement. According to clients, agents sometimes used phrases intended to motivate them, such as “keep up the good work” (Tiffany), “something will come along” (Hailey), and “don’t give up hope” (Sarah). Natalie said her agent encouraged her to “fulfill my dreams of becoming a social worker . . . just because I’m currently on probation doesn’t mean it’s going to hinder me from becoming a social worker once I graduate with my masters.”
Women also described esteem support. Esteem support includes several subcategories: compliments, validation, and relief of blame (Cutrona, 1990). Of these subcategories, study participants said they most often received validation, which occurs when an agent expresses agreement with a client’s assessment of a situation. Kirsten told her agent that she would rather maintain her current job than attend a job fair to look for new employment: “I knew she was OK with whatever my choice was, and she was going to let me make choices instead of forcing me to do something that I couldn’t really do.” Several clients also discussed receiving esteem support through compliments regarding their success in gaining employment. For instance, according to Lauren, “[Agent] said, ‘Congratulations,’ she was happy that I finally found something.” Other agents complimented women for their employment but also recognizing the obstacles and limited work opportunities for people with felony records. Judy said, “She just told me that she was glad to see that I have stable employment and let me know that that’s like actually kind of rare, and that she was impressed with that.”
Few women received supportive communication in the form of tangible assistance. Most of the few instances were agents’ actions aimed at reducing women’s stress. Clients described occasions when agents called potential employers, communicated with current employers, and helped with resumes and applications to assist in the job search. For example, Sue recalled, I guess they start off with, “I’m going to give you the [job] leads, and you see how far you can go with it.” . . . then I take them to wherever step. But if I come back and say, “Probably won’t take me. Can you call?” I think she can help. I see her doing that.
In sum, clients reported receiving many types of supportive messages about employment from their supervision agents. Communication theory suggests that the most common type found in this sample, informational support, can be problematic. This is because clients may perceive suggestions, advice, or referrals to services as indications that their agents find them incompetent, which offers a threat to their positive identity or their sense of autonomy (e.g., Goldsmith, 2004). Instances of emotional support, esteem support, and tangible assistance were less prevalent, and notably no instances of network support were reported. Past research has shown the importance of esteem support for job seekers, as it improves feelings of competency (Holmstrom et al., 2015). Network support is also especially crucial, as larger social networks are strongly associated with positive job search outcomes (Lin, Cook, & Burt, 2001). Based on the optimal matching model of support, these results suggest increased attention be given to other forms of supportive communication outside of informational support.
The second research question involved the connection of different types of employment-related messages to women’s descriptions of that communication as positive, negative, or mixed in its effects. In this sample, messages that contained informational support had effects that were coded as primarily positive. Maxine, who described employment service referrals, described positive effects. She said the referrals helped her “because they [are] places that I normally wouldn’t think to check in or I’m not about to see, and sometimes they get [knowledge of] places directly to them that we don’t have access to.” This message shows the use of informational support consistent with the optimal matching model, as such support could better inform Maxine of her available options.
However, some instances of informational support had effects on their clients that were coded as negative or mixed. Occasionally, this was because the referrals women received were under-resourced for serving the needs of women. Sally recalled a referral to a particular program, “Like I said they really just work to the men, they really don’t do the woman.” Louisa expressed a similar sentiment, saying, “Well, we had the Women’s Resource Center; she referred me to that but . . . they cut that funding.” These findings are consistent with the optimal matching model’s tenet that type of support should be matched to that actors’ level of control. Referrals to employment services are helpful only when the agent can ensure availability; otherwise, such referrals have negative effects. This is an especially important consideration for women, since in some settings programming for women is quite limited.
Emotional support, the second most prevalent type of message, was associated with positive effects exclusively. Most positive effects stemmed from encouragement. Serena explained that her agent’s encouragement led her to see the agent as dedicated to her: “It makes things better ‘cause again it’s the lines of communication are open. She makes me feel comfortable with the fact that she actually cares, like she actually wants me to succeed.” Nina mirrored this sentiment, explaining that encouragement from her agent “gives me some small amount of confidence . . . to know that she cares.” These comments show how a message at the micro-level of analysis can contribute to the agent-level construct of caring in a dual-role relationship. These comments are also consistent with feminist criminological theory and feminist counseling psychology literature that identifies relationships as so central in women’s lives that they must be integrated into interventions (Covington, 1998; Jordan, 2013).
The third most frequent form of supportive communication, esteem support, was exclusively associated with positive effects as well. Clients frequently described esteem support as validation. Louisa indicated that her agent’s validation was especially meaningful in her job search. “[S]he just tells me encouraging things to keep me striving. Just say, ‘Keep on striving, keep doing what you’re doing, but have a plan, have a goal, and know what you want’ is what she tells me.” Maura remarked that her agent’s validation improved her sense of self. “Her having confidence in me, her making me feel like I could do it. Because I don’t have confidence in myself.” Previous communication research found that esteem support affects job search behavior through an improved sense of confidence and competency (Holmstrom et al., 2015), which illustrates the importance of these messages for producing positive outcomes.
Tangible assistance, the least prevalent type of supportive communication, had only positive effects on clients. All but one instance of tangible assistance involved an agent’s accomplishing a direct task. Kelsey recalled the positive effect her agent’s assistance had in obtaining identification: I didn’t have enough ID in order to try and get my Michigan ID. And I told her I needed something with my address, because I’m trying to get my Michigan ID to go to do my applications. She sent me some papers so I would be able to get my ID. I appreciate that. She was very helpful.
By assisting her with a particular task, Kelsey’s agent empowered her to make positive behavioral changes.
The findings from this section emphasize the positive and negative effects of agent’s messages and the relevance of the optimal matching model of support to explaining the connection of message type to effects. Most messages did have positive effects, which is consistent with past researchers’ conclusions regarding the importance of building self-esteem and addressing the emotional elements of reentry (Wyse, 2013). The negative instances of informational support were also explained by the gender responsivity framework. Women were unable to utilize programs that were designed for male clients, and programs for justice-involved women were sometimes under-resourced (Carlton & Segrave, 2016). The next sections explore the connection between supportive and gender-responsive messages in more detail.
Gender Responsivity through Attention to Women’s Multiple Needs
The third research question was whether agents’ messages about employment addressed a holistic range of needs. Women received messages not only about employment but also about increasing human capital. Specifically, study participants recalled messages that encouraged them to pursue educational opportunities. Rebecca explained, “I told her that I didn’t have my GED and stuff and I was real embarrassed about that. She said, ‘Don’t worry about that, you can still get that.’ You know, she gave me resources for that.” Similarly, Dora said her agent actually prioritized attending school over working: Yeah, I just let her know that I’m still looking [for work]. She said she’s not really worried about it because right now I’ve got income coming in. So I’m full-time school, so she wants me to concentrate on that.
Several women described messages from their agents that prioritized meeting needs other than education over finding a job. Agents were often lenient regarding work requirements because of their clients’ needs. Leigh said her agent was supportive of her staying home with her children in spite of the conditions of her supervision: She just understands that I’m a stay-at-home-mom . . . yeah. And she’s not—cause part of the probation, I know it’s, you’re supposed to be actively seeking, you know, employment. I mean and that’s like she just kind of knows the situation . . . Cause I think they really want you to work so you can be productive and pay your fines and do those kind of things. To me it would put me backwards because daycare is ridiculously expensive. I wouldn’t be able to get ahead.
Agents also left clients with the understanding that mental health and substance abuse treatment had priority over employment. Elizabeth noted that her agent considered her mental health needs in conversations regarding employment: She said that at this time she’s not gonna worry about the job search . . . . So she can see how much I’m going to therapy and stuff, so she said she wasn’t gonna really worry about the job search right now.
Roberta described how her agent prioritized substance abuse recovery over employment: She gave me the lists of some [employment] services and everything but I didn’t go to the services because it was kind of conflicting with the days that I go to treatment . . . so she just said, you know, keep looking for a job and everything but still go to treatment, that’s the most important thing.
In these examples, the women depicted agents as aware of and emphasizing needs other than employment, and they described agreement on the priority for meeting other needs over finding employment.
Some women described instances of insensitivity to their multiple needs. Two indicated inattention to their human capital deficits. Ann indicated that she needed more education to find employment, but the agent “never offered anything” to meet that need. She went on to express her concerns, “Okay, I’m not being served properly in education, help find education, show me any programs that can help me because I don’t have any money.” Betty recounted her frustration with a lack of support for her educational needs.
It’s more along the lines of [agent] saying, you know, “you got to find a job, you got to do this, you got to do this.” And it’s like, with no transportation, no reliable transportation, no education, no . . . having the felony on me, it’s more like a triple wham of “we’ll set her application aside.”
Both Ann and Betty viewed their agents’ inattention to their human capital deficits as hindering their ability to find work.
The final research question is about the connection between women’s perceptions of the combined effects of supportive communication and responsivity to multiple needs in the agents’ messages. Some women received messages that were both supportive and gender responsive, and these messages were associated with positive effects. Ellen recalled that in addition to providing emotional support to find work and help filling out an application, her agent encouraged her to act on her desire to return to school.
She tried to encourage me, that if I had wanted to go back to school that I should do it, you know, and that would be good for me she thought, you know. I said, “Well I think I’m too old.” She said, “You’re never too old.”
Adding tangible support to obtain further education, the agent took her and a group of other women on probation to a local university to learn about the application and registration process. Ellen described the benefit of the visit as “the insight of how it works and the in depth of the registration and the application.”
In another instance, Maura received esteem support (i.e., validation) in a message that supported her efforts to take part in a program that addressed multiple needs simultaneously: She’s talked to me about that [finding or keeping a job] and I told her about what I—you know, me going to therapy, therapists, going to a program, this place to help me get a job or with my learning disability or going to school . . . I came and told her about what I got referred to [by the program], to ask her what does she think, and she said that it was a great idea.
Maura described the positive effects of the communication: So yeah, she stood beside me. Her having confidence in me, her making me feel like I could do it. Because I don’t have confidence in myself, you know, when I look in the mirror I don’t see something good. As [my agent] makes me feel like, you’ve got it, you can do it. So she does positive for me. So yeah, I feel positive with [my agent].
Beyond having an effect on job search behavior, Maura’s perception that the agent supported her efforts increased her self-esteem. The combination of gender responsivity to multiple needs, appropriate prioritization of goals, and esteem support, resulted in positive effects.
As previously noted, informational support was the only type of supportive communication associated with negative and mixed effects. Not only the optimal matching model but also gender responsivity explained the connection of some types of messages to positive versus negative effects. For example, Kelley’s agent gave her informational support so she could both care for relatives and obtain employment: She just asked me if I had a job, and I told that I’m currently unemployed because I take care of my mother and my grandmother, and she also told me that I could, if I get paid by the state for taking care of my mother and my grandmother that that would be considered as a job.
In this case, the agent and client agreed on the client’s immediate needs and responsibilities, and the agent supported the women’s strivings to have multiple needs met simultaneously. Similarly, Melissa said her agent provided support by getting her into a women’s halfway house that provided employment services. Asked if this made things better, worse, or had no effect, Melissa replied, “One hundred percent better, because I would not be straight right now if I had not gone to [the halfway house]. I would’ve been absconded and I’d have been homeless and sleeping with bed bugs.” This agent’s referral impacted Melissa’s living situation, employment search, and adherence to parole requirements, thereby providing a holistic intervention, consistent with a feminist criminological approach.
Instances in which agent’s informational support ignored the multiple needs of their clients were associated with negative effects. Toya explained that her agent stated she “basically didn’t need one [a job]” because she needed “to go and fix my alcohol problem first,” and referred her to a recovery program without providing job search resources. She conveyed her frustration: But she don’t understand when I don’t work, I drink. When I work, I don’t drink . . . It’s like just right now she wants me to go get help, so now I’m unemployed, so I can’t pay my bills and I can’t take care of my children and she sees that as an okay situation, which it’s not.
Toya felt that her supervising agent’s advice failed to recognize that not only did she have a substance abuse problem, but working limited her drinking, and she needed to be employed so she could pay her bills and support her children. This passage suggests the relevance of a holistic assessment of women’s needs, matching of communication to those needs, and agent–client agreement about the order of meeting needs.
Uma’s interactions with her supervision agent also highlight the connection between instances of agent informational support, insensitivity to multiple needs, and negative effects. She explained the inconsistency of her lived reality with the agent’s suggestion and advice: And every time I tell her I’m looking for a job and then she just says the most unrealistic things. . . . She goes, “Finding a job is like having a full-time job, so you’re at work for 8-10 hours a day, you need to work 8-10 hours to find a job.” So you expect me with no car to just walk up and down every day? With a newborn baby?
This excerpt again shows the importance of agreement on priorities and on the sequencing of efforts to address multiple needs—in this case, education, caring for children, transportation, and work—through a shared decision-making process, which is a principle of motivational interviewing (Miller & Rollnick, 1991). Because this principle is ignored, Uma is faced with recommended action that she views as unrealistic given circumstances that are out of her and the agent’s control, specifically having no car and a newborn child. Consistent with the optimal matching model and the gender responsivity framework, Uma described the negative effect of this message. Uma noted that she dismissed the advice by saying to the agent, “Sure, okay, I will. I’ll get it done. Don’t worry.” She followed this with a comment to the interviewer, “It’s just the most annoying thing in the world.” The analysis of gender-responsive and gender-insensitive messages suggests a potential channel of influence between responsiveness to multiple needs, supportive communication, and the perceived valence of messages.
Discussion
Guided by the optimal matching model from the field of communication and feminist criminological theory that highlights the importance of addressing women’s multiple needs, the present research investigated specific messages that women on community supervision recalled from their conversations with supervision agents about employment. This research is unique in criminal justice because it carried out a message-level analysis. However, the research has implications for other models that are used to understand the criminal justice system. In particular, the nature of messages from supervision agents to clients appears to support individual-level agent–client relationships that are the focus of research on dual-role relationships. Especially when it is gender responsive, supportive communication at the message level is associated with women’s perceptions of caring relationships with agents.
The optimal matching model of social support specifies that nurturing support is necessary for addressing overwhelming and severe stressors (Cutrona & Suhr, 1992). The limited occurrence of tangible assistance by agents in the area of employment may occur because agents are aware of the obstacles clients face, and the agents may be limited in their ability to provide clients with direct assistance in gaining employment. The optimal matching model of social support also made it possible for us to understand that informational support in the form of advice does not have positive effects when neither the advice-giver nor the recipient has much control over the stressor (i.e., lack of work). In such cases, there are more appropriate forms of support. Emotional support can reduce negative feelings, and instrumental support such as referrals or tangible assistance can help the recipient address the problem (Cutrona & Suhr, 1992). In contrast, advice in the absence of the consideration of the client perspective may produce negative effects, because the client perceives it as impersonal and inappropriate.
Messages that took into account a variety of problems commonly shared by women on probation and parole had positive effects, whereas those that were relevant to employment, but failed to consider other needs had negative effects. Both gender responsivity and the use of the type of communication most helpful given the malleability of the stressor may account for dual-role relationships that clients view as caring. Considerable research in the discipline of communication confirms that the receipt of social support through communication expedites job searches, which in turn increase employment (e.g., Holmstrom et al., 2015; Lim et al., 2016; McDonald, 2011). The present research suggests that for women, messages that are also seen as gender responsive are most supportive, and produce positive effects.
Limitations and Future Research
There are both limitations and future directions for research that should be addressed. The instances of supportive communication and their effects were nested within the study participants, who were nested within the supervision agents’ caseloads. However, because of the very few instances of communication that provided emotional, esteem, and tangible support, it was not possible to use quantitative multilevel analysis to address the nonindependence of the data. An experimental design in which the type of support is manipulated would provide additional evidence of the effects of alternative support types.
The finding that no women reported network support may be due to the focus of questions on agent–client discussions to the exclusion of discussions about referral targets. Referrals to employment services and programs were coded as informational support, per the SSBC. These referrals may have indirectly provided network support, resulting in an underestimate of network support resulting from referrals to community resources.
In addition, this study suffered from some inconsistencies in the way that interviewers posed questions about supportive communication effects. Some interviewers did not ask questions regarding the valence of each instance of supportive communication. This resulted in lower numbers of reported valances. Also, the high proportion of reports of informational support may be overinflated because interviews contained specific questions asking women whether they were referred to any services to assist them in obtaining employment.
The assessment of gender responsivity in messages was limited to women’s unprompted comments about attention to multiple needs in the course of describing what the supervision agent said about employment. Interviewers did not explicitly ask women whether discussions about employment simultaneously focused on or ignored other needs, though sections of the interview asked about other needs, and some women talked about messages about those needs in relation to discussions of employment. Future research should incorporate direct questions about whether agents talked about prioritizing multiple needs in discussions of employment.
Finally, the Michigan context is unique, and the findings may not generalize to men and to the constellation of needs and life demands that may be uniquely important to them. Comparative research on settings where probation and parole are not expected to be gender responsive and replication of the research on men would provide useful information.
Implications
This study provides support for integrating theories from the field of communication and from feminist criminology to understand criminal justice practices and their effects on women. The current study and prior research (Holmstrom et al., 2017; Morash, Kashy, Smith, & Cobbina, 2015; Smith et al., 2016) revealed that an interdisciplinary approach can extend theory and knowledge for multiple disciplines simultaneously. The SSBC (Cutrona & Suhr, 1992) that was developed by communication scholars is a helpful tool for examining supervision agent–client messages and may prove to be useful in other criminal justice contexts.
In addition, this work provides evidence to support the importance of gender-responsive messages from supervision agents. Gender responsivity in tandem with social support is associated with positive effects on clients. Women in this study recounted positive experiences with agents when they viewed them as sensitive to multiple needs. This finding is consistent with past work on gender responsivity in corrections, and it contributes to the limited prior research (e.g., Sydney, 2005) on gender responsivity in community supervision.
The first practical consideration based on these findings is that the cooperation and collaboration of multiple agencies, including potential employers, with supervision departments would be helpful to provide agents with resources to help their clients gain employment. Agents need resources to assist their clients in finding employment, especially if employment is considered a requirement of supervision. Given the potential for negative effects such as frustration and reactance when agents provide advice without adequate resources, or are insensitive to the range of clients’ interrelated needs, practice may need to be altered to be more consistent with the circumstances and resources available to clients. As illustrated by the data, expectations and requirements that women find employment could be modified to better align with women’s opportunities and needs. Shifting expectations and assigning priority to urgent issues such as substance abuse and mental health problems at the beginning of supervision, and allowing additional time to obtain employment, may promote more positive outcomes when women do start seeking employment and may lead to a better agent–client relationship.
This research also suggests that many supervision agents understand the limited employment opportunities for women in their caseloads, so they provided employment-related emotional support, and in some cases prioritized meeting other needs over employment. Given the positive effects of agent’s emotional and esteem support, and positive effects associated with gender-responsive communication, agents should be trained to recognize and provide these types of support. Although this study did not examine the co-occurrence of types of support, it may be that advice is better received when preceded by or paired with emotional and esteem support (Feng, 2009, 2014). It may also be that tangible assistance and particular types of informational support such as referrals are required for advice to produce positive outcomes. Future research should examine the order and co-occurrence of support and the subcategories of support in relation to effects to best inform practitioners of the ways to assist their clients who are struggling to find stable employment, along with the ways agents can integrate gender responsivity and social support to produce the best possible outcomes for their female clients.
Finally, the focus at the message level provided insight into the range in communication that women receive from supervision agents and their reactions to it. These types of examples may be useful not only to scholars but also to practitioners and training specialists who seek examples of a variety of ways to provide effective and ineffective support to women.
Footnotes
Authors’ Note:
This article is based on work supported by the National Science Foundation under Grant 1126162 and by a Strategic Partnership grant from the Michigan State University Foundation.