Abstract
The research related to the working alliance in career counseling is reviewed in this article. This review indicates that the working alliance does typically increase over the course of career counseling. Furthermore in career counseling, most of the correlations between the working alliance and various outcome measures were significant and hovered around .30, which is consistent with findings related to the correlation between the working alliance and the outcome in psychotherapy. In terms of factors that predict the working alliance’s contribution to career counseling outcome, there is a lack of studies and more research is needed in this area. This article also provides suggestions for practice in career counseling and recommendations for future research.
Meta-analyses indicate that individual career counseling is either the most effective or one of the most efficacious methods of providing career counseling (Brown & Ryan Krane, 2000; Whiston, 2002; Whiston, Sexton, & Lasoff, 1998). Whiston, Sexton, and Lasoff (1998) found individual career counseling to be the most effective career counseling modality and that its effect size was substantial (i.e., d = 1.08). M. J. Heppner and Heppner (2003), however, contend that remarkably little is known about the process and mechanism that produce these large effect sizes in individual career counseling. Heppner and Heppner go on to delineate 10 promising avenues of process-oriented research of which the first one involves examining the influence of the working alliance in career counseling. In a seminal article, Meara and Patton (1994) argued that the working alliance is a critical element of career counseling and contributes to its effectiveness. Nevertheless, in summarizing the research related to the working alliance’s contribution to career counseling effectiveness, Whiston and Rahardja (2008) concluded that the findings were mixed and that one could not conclude that the working alliance directly influenced the effectiveness of career counseling. Whiston and Rahardja (2008), however, was not a systematic review of the research on the working alliance in career counseling and, furthermore, did not include more recent research in this area. This research review is designed to address that gap and systematically review the research related to the working alliance in career counseling.
Although there are numerous models of the client/counselor relationship within psychotherapy, one of the more prominent is Bordin’s conceptualization of the working alliance (Gelso & Carter, 1994). The working alliance as conceptualized by Bordin (1979) is a tripartite model made up of (a) goals, (b) tasks, and (c) bonds. The goals refer to the degree to which there is agreement between the client and the counselor on what is to be done, the tasks denote how the counselor and client go about accomplishing those goals, and the bonds are the emotional connections or relationships between the client and the counselor. This is a definition that is not specific to a particular therapeutic approach, and working alliance can therefore be considered as a “pan-theoretical concept” (Horvath & Luborsky, 1993). Gelso and Carter (1994) suggested that working alliance is fundamental in all forms of counseling and psychotherapy and accounts for a significant amount of variance in outcome. Working alliance can thus be considered as a “common” process variable (e.g., Martin, Garske, & Davis, 2000). Several authors, such as Amundson (1995), Swanson (1995), or Patterson in an interview (Freeman, 1990), have suggested that establishing a positive working alliance or relationship between the client and the counselor is one of the first steps in career counseling and is considered crucial in initiating client change.
There has been, however, limited exploration of the working alliance in career counseling as compared to psychotherapy. Interestingly, Lewis (2001) did not detect any significant difference between the levels of the working alliance in psychotherapy and career counseling. This finding was substantiated with a much larger sample by Perdrix, de Roten, Kolly, and Rossier (2010). In a meta-analysis that analyzed the relationship between the working alliance and the therapeutic outcome in psychotherapy, Horvath and Symonds (1991) found an average effect size of r = .26. More recent meta-analyses of the relationship between the working alliance and the outcome in psychotherapy have found similar results with Horvath and Bedi (2002) finding an effect size of r = .21 and Horvath, Del Re, Fluckiger, and Symonds (2011) finding an effect size of r = .28 (k = 190). However, there has been no systematic review of the relationship between the working alliance and the outcome in career counseling.
The search for pertinent studies for this systematic review was conducted in three phases. First, the databases of both PsycInfo and ERIC were searched over the last 25 years (1990 through 2014). The key words employed were career counseling, occupational guidance, or vocational counseling and either the working or the therapeutic alliance. These terms were selected because they were either germane to the project or a common classification within the databases (e.g., occupational guidance). In order for a study to be selected, it needed to empirically measure the working alliance consistent with Bordin’s (1979) conceptualization of the working alliance. Both published and unpublished (e.g., dissertations) were included in this review. Furthermore, studies published in English and French were also included. It should be noted that we searched for articles in other languages but only identified ones in English and French. The second phase of the search was to use the reference lists from studies identified from the search of databases. The third phase of the search was to review summaries of career counseling outcome research (e.g., Whiston & Rahardja, 2008) for studies that were applicable. These three phases resulted in us identifying 17 empirical studies about the working alliance in career counseling. Reviews, purely theoretical contributions, studies that have not specifically considered the working alliance but only mention it in the discussion, and studies of the importance of the working alliance between career counselors in training and their supervisor were not included.
This review of the working alliance in career counseling is divided into five sections, which emerged as we summarized the data. The first section addresses the question of whether the working alliance increases during the course of career counseling as it typically does in psychotherapy. The second section examines the relationship between the working alliance in career counseling and outcome. This section delves into whether the working alliance has an impact on the effectiveness of career counseling. The third section of the article examines what factors influence the working alliance in career counseling. This section includes both client factors (e.g., personality) and counselor factors (e.g., techniques and activities) that may influence the working alliance. Implications and suggestions for practice are included in the fourth section of the article. In particular, this section will focus on techniques counselors can use to positively influence the alliances they establish with clients. The fifth and last section suggests a research agenda to further study the role and importance of the working alliance in career counseling.
Does the Working Alliance Increase During Career Counseling?
In psychotherapy, there is a consistent finding that the working alliance increases over the course of the counseling, and there has been some speculation about whether the working alliance in career counseling follows a similar trajectory as that found in psychotherapy (M. J. Heppner & Heppner, 2003). In an early study of the working alliance using a sample of 55 adult clients, M. J. Heppner, Multon, Gysbers, Ellis, and Zook (1998) found that the working alliance generally increased over the course of career counseling. Using a holistic approach to career counseling, Multon, Heppner, Gysbers, Zook, and Ellis-Kalton (2001) also examined the trajectory of the working alliance during career counseling and found, once again, that the working alliance increased over the course of the career counseling. The finding that the working alliance increased over the course of career counseling was also found by M. J. Heppner et al. (2004) with clients who received at least five sessions of career counseling. This result was confirmed by Covali, Bernaud, and Di Fabio (2011) who studied the level of two aspects of the working alliance, namely, agreement about the goals and the tasks with older French adolescents. They measured agreement on goals and tasks at the beginning of the second and at the end of the third and fourth session of a five-session intervention. They observed an increase for these two components of the working alliance with a medium to large effect size (d = 0.63). Finally, Elad-Strenger and Littman-Ovadia (2012) examined the trajectory of the working alliance over three sessions and found that both client and counselor ratings of the working alliance increased over the course of counseling. Therefore, there is substantial research that supports that career counselors do indeed establish therapeutic relationships with clients and that the working alliance does increase over the course of the career counseling. An increase in the working alliance, in and of itself, does not mean necessarily that clients change and improve. Therefore, the next section involves an analysis of the impact of the working alliance on career counseling outcome.
Is the Working Alliance Related to Career Outcome?
Before discussing the relationship between the working alliance and the career counseling outcome, it is important to define outcome. While outcome is typically seen as the result of the counseling process, it can be simply seen as the degree to which there are improvements in clients’ presenting problems, symptoms, or functioning after treatment (Crits-Christoph, Gibbons, & Mukherjee, 2013). Table 1 is a summary of the correlations between the working alliance and the career outcome measures found in the eight studies that examined this relationship. In this table, the correlations are between total working alliance measures and the specific outcome measures listed. Unless otherwise noted below, the studies in Table 1 used the client version of the Working Alliance Inventory–Short Form (WAI-S, Horvath & Greenberg, 1989). The WAI-S is a 12-item instrument that is commonly used in psychotherapy research on the alliance (Horvath, Del Re, Fluckiger, & Symonds, 2011).
Correlations Between Measures of Working Alliance and Career Outcomes.
aSignifies partial correlation controlling for pretest. bUsed the Working Alliance Inventory-Client. cUsed another measure.
Using hierarchical linear modeling, Multon et al. (2001) found that the linear growth in the working alliance was related to multiple measures of psychological distress and goal instability, which decreased during the career counseling. This career counseling was provided at a university-based career counseling center and was with predominately Caucasian clients. The researchers found, however, that the amount of variance accounted for by the working alliance was small to medium (i.e., it ranged from 1% to 12% of the variance in outcomes).
M. J. Heppner et al. (2004) found that as the working alliance increased, so did clients’ problem-solving appraisal with adult clients who received at least five sessions of career counseling at a university-based career counseling center. Specially, these researchers found that those clients who made the most improvement in problem-solving appraisal also reported an increase in their working alliance over time. Whereas, the methodology for this study did not allow for causal inferences, the finding that both the problem-solving appraisal and the working alliance simultaneously increased is important since problem-solving appraisal has been linked to coping with career transitions (P. P. Heppner, Cook, Strozier, & Heppner, 1991), career decidedness (Larson, Toulouse, Ngumba, Fitzpatrick, & Heppner, 1994), and certainty with approaching career-related problems (McCracken & Weitzman, 1997).
Carozzoni (2002) conducted a study that focused exclusively on the relationship between the working alliance and the career outcomes. This study was also conducted with a predominately Caucasian sample. While controlling for the pretest, he found significant partial correlation coefficients of .46, −.40, and .48 between the working alliance and vocational identity, career indecision, and career development, respectively. It should be noted that as the working alliance increased, career indecision decreased. Furthermore, the researcher speculated that these correlation coefficients were an underestimation of the magnitude of the relationship because of floor or ceiling effects for the variables of vocational identity, career indecision, and career development. Carozzoni found that the influence of the working alliance occurred early in the counseling process, as the average number of sessions was only 2.16; however, Carozzoni also found that those clients who had more than two sessions had higher working alliance scores than those who attended only one or two sessions.
With older Swiss adolescents (average age 17.9), Masdonati, Massoudi, and Rossier (2009) examined the relationships among the working alliance assessed after the third session and measures of life satisfaction, career decision-making difficulty, and satisfaction with the intervention measured at the conclusion of the counseling. Although these researchers did correlated the three subscale of the WAI (i.e., goal, task, and bond), we will present the partial correlation coefficients only for the total score in order to be consistent with other studies in this area. Once again, these researchers controlled for the pretest scores in the correlations and found partial correlation coefficients of .32 for satisfaction with life, −.25 for career decision-making difficulties, and a substantial coefficient of .67 for satisfaction with the intervention. All of these partial correlation coefficients were significant.
Guillon, Dosnon, and Herrmann (2010) assessed the working alliance, using a French translation of the WAI, and an index of the clients’ perception of the effectiveness after a one-session intervention with a small sample of French high school students (N = 26). The index of effectiveness included both clients’ assessment of usefulness of the intervention and satisfaction with the counselor. They found that this index of effectiveness was strongly correlated with the working alliance (r = .70).
Covali and colleagues (2011) examined the relationships among a measure of the working alliance that only include agreement on goals and tasks with two career outcome measures (i.e., vocational self-efficacy and career maturity) with a sample of 127 French adolescents. Correlation between the working alliance and the two outcomes were negligible at the beginning of the intervention (<.01 and −.06, respectively). However, at the last measurement point, which was after the fourth session, the correlation coefficients were quite high, .61 and .60, respectively. Interestingly, the level of the working alliance at the end of the intervention also correlated with the beginning outcome measures (i.e., .28 for vocational self-efficacy and .30 for career maturity). Thus, clients with higher levels of vocational self-efficacy and career maturity developed better working alliances.
Elad-Strenger and Littman-Ovadia (2012) found in their study comparing counselors’ and clients’ rating of working alliance in a three-session intervention that counselors’ ratings of the working alliance were not predictive of clients’ career exploration. The magnitude, however, of the relationship between the clients’ rating of the working alliance and career exploration was .28 for the first session, .26 for the second, and .28 for the third. In addition, they found that clients’ rating of the working alliance at the first session predicted career exploration but that ratings of the second and third sessions did not predict any additional variance. This means that counselors should strive to establish a working alliance early in the career counseling process.
Similar to career counseling is vocational rehabilitation counseling, and Lustig, Strauser, Rice, and Rucker (2002) examined impact of the working alliance in vocational rehabilitation counseling. Using a measure called the Working Alliance Survey, which included items on the bond and agreement on goals and tasks, these researchers contacted 2,732 clients by telephone 60 days after they received services. These researchers found that there was a significant difference between employed participants rating of the working alliance as compared to unemployed participants with an effect size of d = 0.73. Furthermore as Table 1 reflects, the working alliance was strongly related to individuals’ views of future job prospects for both employed and unemployed participants.
In order to further specify the relationship between the working alliance and outcome, Masdonati, Perdrix, Massoudi, and Rossier (2014) examined the direct, moderator, and mediator effects of the working alliance on outcomes with a sample of Swiss clients. Moderation would indicate that the working alliance influences the strength or the magnitude of the relationship between the preintervention (i.e., predictor) and the postintervention (i.e., criterion) levels of outcome; whereas, mediation would stress the mechanism by which the preintervention (i.e., predictor) influences the working alliance, which in turn influence the postintervention (i.e., criterion). The outcome measures in this study of moderation and mediation were career decision-making difficulties, life satisfaction, and satisfaction with intervention. In regard to correlations, the working alliance was highly correlated with satisfaction with counseling (.59), significantly correlated with career decision-making difficulties (−.28) and significantly correlated with life satisfaction (.20), which are consistent with findings from other studies.
Masdonati et al. (2014) found that the type of working alliance (i.e., goals, tasks, and bond) impacted differently the outcome measures in terms of moderation and mediation. The working alliance had a direct influence on satisfaction with the intervention. In terms of moderation, the higher the levels of agreement on tasks and goals, the larger the decrease in lack of information, which is one of the subcategories of career decision-making difficulties. The total score on career decision making was also moderated by agreement on tasks. Therefore, career counselors should particularly focus on agreement on tasks when they want to decrease career decision-making difficulties. Related to mediation, the total working alliance score partially mediated the total career decision-making difficulties score. Agreement on goals was the dimension of the working alliance that better mediated the decrease in total career decision-making difficulty. Furthermore, inconsistent information was mediated by the total working alliance score, agreement of tasks, and agreement on goals. The subcategory of bond did not mediate any of the outcome measures. Taking these results together, the career counselor should focus on agreement on tasks when the goal is to increase career information; however, if the client has inconsistent information about occupations, then the career counselor should focus on both agreement on tasks and agreement on goals.
In conclusion, there does appear to be a significant relationship between the working alliance in career counseling and the various measures of outcome (see Table 1). Although there are probably not a sufficient number of studies to conduct a meta-analysis, we do see consistencies concerning the magnitude of these correlation coefficients and can make some conclusions. The correlation coefficients range from .10 to .70. We estimate that the average correlation between the working alliance and the outcome in career counseling would be in the .30 range. If one takes the simple average of the correlation coefficients listed in Table 1, it is equal to .36. We believe that the average of .36 is somewhat inflated as some of the weaker studies had higher correlation coefficients, so our estimate is more conservative. Cohen (1988) concluded that a correlation effect size of .30 is considered a medium effect. It should also be noted that an average correlation coefficient of .30 is consistent with the findings of Horvath and colleagues (2011) who found an average effect size of .28 when examining the relationship between the working alliance and outcome in psychotherapy. Although one should be cautious when interpreting the average relationship in career counseling between the working alliance and outcome because of the small number of studies, we cautiously suggest that the working alliance may play a comparable role in career counseling as it plays in psychotherapy.
What Factors Influence the Working Alliance?
In examining the factors that influence the working alliance, it is important to explore both sides of the dyad (i.e., client and counselor factors).
Client Factors
Surprisingly, there has been little examination of client demographic variables (e.g., gender and race/ethnicity), which appears to be a fertile area of research. In one of the first analyses of the working alliance in career counseling, M. J. Heppner and Hendricks (1995) used a single-subject design to compare the working alliances with an indecisive client and an undecided client. An indecisive client is one who has chronic problems with both career and other decisions, whereas the undecided client is one whose career indecision is developmentally appropriate. Due to the chronicity of decision problems with indecisive clients, it has been speculated that career counseling with indecisive clients should be more intensive, relationally oriented, and longer than for clients who are undecided (Salomone, 1982). Therefore, Heppner and Hendricks hypothesized that the working alliance would be more central to the indecisive client, as it was anticipated that this career counseling would be more similar to psychotherapy. However, the undecided client who received more traditional career counseling rated the working alliance as being stronger in this study.
Ray (1998) investigated whether the development of the working alliance was related to the five factors of personality (i.e., neuroticism, extroversion, openness to experience, agreeableness, and conscientiousness). He found that clients high in extroversion tended to have better working alliances as compared to those who scored more toward introversion. He speculated that this finding may be related to extroverts’ preference for interacting with others. Ray also found that agreeableness predicted working alliance total scores. Interestingly, neuroticism, openness to experience, and conscientiousness were found not to significantly influence the total working alliance scores. Covali et al. (2011), however, found that at the beginning of the intervention, the working alliance correlated positively with extroversion (r = .15) and conscientiousness (r = .13), using the Revised NEO Personality Inventory. The correlations between personality traits and the working alliance at the end of Session 4 were higher with a correlation of −.17 with neuroticism, .24 with extroversion, .23 with openness, and .17 with conscientiousness. As will be discussed later, more research needs to be conducted that examines personality and the development of the working alliance in career counseling.
The type of difficulties and the level of distress are also known for having an impact on the facility/difficulty of establishing a positive working alliance in psychotherapy (Castonguay, Constantino, & Grosse Holforth, 2006). As mentioned above, Covali et al. (2011) observed that clients with lower levels of vocational self-efficacy and career maturity also reported lower levels of working alliance. Masdonati and colleagues (2014) observed that the level of career decision-making difficulties and life satisfaction at the beginning of the intervention were, respectively, negatively (r = −.20) and positively correlated (r = .18) with the level of working alliance after the third session of an intervention of about four to five sessions. The first correlation was due to the negative correlation between the subdimension of inconsistent information and working alliance (r = −.26). This means that it will be more difficult to establish a positive working alliance with clients who have higher levels of career decision-making difficulties and those who are more dissatisfied with their lives.
Although level of distress was found by Masdonati and colleagues (2014) to be related to the working alliance, Murray (1997) did not find a difference on ratings of the working alliance between veterans with posttraumatic stress disorder (PTSD) and those without PTSD. These findings, however, should be interpreted cautiously as the study involved a small sample size and inconsistencies when the working alliance was measured.
Counselor Factors
In addition to client factors, the working alliance is affected by counselor-related factors such as counselor demographic variables and intervention strategies. There has been little research related to counselor variables that may influence the working alliance in career counseling. One exception to this dearth of research is M. J. Heppner et al.’s (1998) finding that counselor self-efficacy in regard to providing career counseling was not associated with better working alliances between the client and the counselor.
In an attempt to provide career counselors with detailed instruction on how to create a working alliance in career counseling, Multon, Ellis-Kalton, Heppner, and Gysbers (2003) investigated the relationship between counselor verbal response modes and the working alliance. There is substantial psychotherapy research that indicates that the quality of the working alliance is particularly important in the early phase of therapy (Sexton & Whiston, 1994): Hence, Multon et al. focused on the first three session of the career counseling. These researchers used the Revised Hill Counselor Verbal Response Modes Category System to categorize the counselors’ in-session responses. Interestingly, provision of information, paraphrasing, and closed question accounted for 79% of the counselors’ total responses. These responses, however, were not significantly correlated with the working alliance. The only significant correlation found was that between counselor self-disclosure and the working alliance (r = −.58). Sendrowit (2011) investigated whether two specific types of self-disclosure versus no self-disclosure influenced the working alliance. Specifically, she examined whether coping-mastery self-disclosure (the counselor disclosed having a similar problem and subsequently described how he or she resolved the problem), similarity self-disclosure (the counselor simply disclosed a similar problem and did not model coping strategies), or no self-disclosure affected the working alliance; however, she did not find any significant difference among the three types of counselor self-disclosures.
The content or ingredients of an intervention might also have an impact on the establishment of a working alliance. Masdonati et al. (2009) used four of the five critical ingredients identified by Brown and Ryan Krane (2000) in their exploration of the connection between the working alliance and three measure of outcome. The five critical ingredients of career counseling as found by Brown and Ryan Krane are workbooks and written exercises, individualized attention and feedback, world of work information, attention to building support, and modeling of career decision making. The counselors in the Masdonati et al.’s (2009) study did not include modeling in their career counseling. The counselors were supervised to ensure that the four ingredients were indeed provided to the clients. The researchers found that the working alliance was indeed related to the outcome measures of career decision-making difficulty, life satisfaction, and counseling satisfaction (see Table 1).
Ray (1998) hypothesized that clients who took the Strong Interest Inventory in the second session after an initial interview would have a stronger working alliance than clients who took the Strong Interest Inventory in the first session followed by an interview. He found that those who received the interest inventory in the first session had a lower working alliance rating than the other group after the first session; however, the groups did not vary in terms of the working alliance after both groups participated in the second session. Therefore, it appears that the interview is critical to the working alliance and an important factor in the process of career counseling.
In summarizing the factors that contribute to the working alliance in career counseling, there appears to be little research that could serve as guidance to the practitioner. Although we know that the working alliance tends to increase during the course of career counseling, we know little about what counselor behaviors lead to the development of the working alliance in career counseling or what client factors contribute to its formation.
Implications for Practice
Perhaps the most important finding of this review is that the working alliance may play just as robust factor in career counseling as it does in psychotherapy. Therefore, career counselors would be advised to attend to the alliance between themselves and their clients. We, however, would assert that building an alliance will vary depending on the client and it is, therefore, important to listen closely to the client and react individually. The focus of the working alliance on agreement on goals and tasks provides structure to career counseling, which may need to be frequently revisited. For example, counselors may want to revisit agreement on goals each session, as the goals of the sessions may change as the career counseling advances and evolves. Early goals might involve the identification of values and interests, whereas later goals could involve the integration of information and decision making. Moreover, with a constructivist approach, the client would need to agree on the task of telling stories and the overall goals of identifying a career theme; whereas with a cognitive information processing approach, the client would need to understand the processes of career problem solving and decision making. What does appear important is attending to issues of agreement on specific tasks, an agreement on goals for the session, and goals for the career counseling.
Within the working alliance model, there is also the focus on the bond between the client and the counselor. Although Multon et al. (2003) found very few counselor verbal responses affected the working alliance in career counseling, clinicians can adopt many of the counselor behavior found in psychotherapy research to influence bond. In a recent meta-analysis, Elliott, Bohart, Watson, and Greenberg (2011) found an overall relationship between empathy and outcome to be r = .30. Interestingly, the perspective of the rater made a difference for empathy–outcome correlations and clients’ assessment of empathy best predicted outcome. This is consistent with our finding that the client’s view of the working alliance is the best predictor of outcome. In another meta-analysis of psychotherapy outcomes, Farber and Doolin (2011) found that positive regard and affirmation were significant related to outcome with an effect size of r = .27. This meta-analysis, however, was only based on 18 studies, whereas the meta-analysis on empathy was based on 57 studies. Finally, in terms of congruence and genuineness, Kolden, Klein, Wang, and Austin (2011) found a mean weighted effect size of r = .24, which was based on 16 studies. All three of these effect sizes would be considered moderate to large and substantiates the importance of empathy, positive regard, and congruence.
Another important finding from this review is that counselors should attend to establishing a working alliance early in the career counseling process. For example, Elad-Strenger and Littman-Ovadia (2012) found that the working alliance in the first session significantly predicted career exploration. Moreover, we found that even with brief interventions, the working alliance appeared to play a critical role and was related to outcome (Guillon, Dosnon, & Herrmann, 2010). Although only one study (i.e., Masdonati, Massoudi, & Rossier, 2009) examined the association of Brown and Ryan Krane’s (2000) critical ingredients and the working alliance, we may be able to draw some tentative conclusions from this study. In our estimation, the critical ingredient that may be most salient in terms of influencing the working alliance is individualized attention and feedback. In particular, it may be important to provide assessment feedback in an individualized manner and with attention to the client’s characteristics and worldview.
In terms of client characteristics, clinicians are encouraged to consider level of distress or difficulties and level of neuroticism. Indeed, Covali and colleagues (2011) observed that clients with low levels of vocational self-efficacy and career maturity developed lower level of working alliance. M. J. Heppner and colleagues (2004) also observed that clients with more psychological resources and facing fewer barriers developed higher levels of working alliance. This was confirmed by Masdonati and colleagues (2014) who observed that the level of career decision-making difficulties at the beginning of the intervention correlated negatively with the level of alliance at the middle of the intervention (r = −.20).
In the psychotherapy literature, they often use the term “tear and repair” to represent the process when there is a rupture in the alliance which is later repaired or resolved by the counselor (Safran, Muran, & Shaker, 2014). In career counseling, ruptures in the working alliance can occur when the client does not have the necessary skills and abilities to enter a career of interest. For example, a client may aspire to be a physician but does not have the necessary aptitudes in science. The career counselor will probably need to broach this delicate topic, which will probably have a negative impact on the quality of the working alliance. Interestingly, Kirschner, Hoffman, and Hill (1994) found that challenge was one of the most effective counselor intentions in career counseling. Therefore, challenging the client does not necessarily result in negative outcomes, but it may result in a tear in the working alliance that needs to be repaired by the counselor.
In conclusion, when considering the practice of career counseling, we assert that there is now research which supports Meara and Patton (1994) call for counselors to attend to the working alliance in career counseling. In particular, it is important that counselors focus on all three aspects of the working alliance (i.e., agreement on goals, agreement on tasks, and the bond). Furthermore, counselors should attempt to establish a strong working alliance early in the process. We further suggest that there is research support for attending to the working alliance even in brief career interventions. Counselors should also consider that some clients may be more challenging in terms of building an alliance. Career counselors should also understand that the alliance may rupture or decrease during the career counseling and it is important to repair the alliance.
Future Directions for Research
Although there is research that indicates that the working alliance does increase over the course of career counseling and that this working alliance is related to the effectiveness of career counseling, we know very little about how to facilitate a working alliance in career counseling. We propose that the correlation between the working alliance and outcome is comparable to the correlation between the working alliance and outcome in psychotherapy. This, however, is a preliminary finding as it is based on only eight studies, whereas there have been hundreds of studies that have examined the relationship between the working alliance and the outcome in psychotherapy (Horvath & Bedi, 2002; Horvath & Symonds, 1991; Horvath et al., 2011). The dearth of studies regarding the role of the working alliance in career counseling highlights the need for additional studies in this area. There are an insufficient number of studies to conduct a meta-analysis, which would give a more precise and appropriately weighted indicator of the relationship between the working alliance and the career counseling outcome than the estimate we have provided. Therefore, researchers are encouraged to conduct more correlational studies of the working alliance and measures of career counseling outcome. In conducting these studies, researchers should consider utilizing the WAI-Short and Revised (WAI-SR, Hatcher & Gillaspy, 2006) as compared to the WAI-Client (Horvath & Greenberg, 1989), as it was found by researchers (Perdrix, de Roten, Kolly, & Rossier, 2010) to have a more robust factor structure and similar psychometrics in both career and personal counseling.
Although we suggest more correlational studies, it should be acknowledged that there are limitations in terms of the conclusions that can be drawn from correlational studies. For example, we cannot say that a strong working alliance in career counseling causes better outcomes. Problems with correlational studies of process and outcome are not limited to career counseling for many of the studies in psychotherapy also involve correlational designs (see Crits-Christoph et al., 2013). Crits-Christoph, Gibbons, and Mukherjee (2013) suggest that experimental manipulation of variables is necessary to truly establish causation. Although experimental manipulation of the working alliance would probably be unethical in complex cases where we hypothesize that the working alliance is critical, it may be possible in specific areas of career counseling. For example, in the area of providing occupational information, it may be possible to manipulate the working alliance where one group receives occupational information with a focus on the alliance and another group receives the information from a counselor who simply provides the occupational information and then compare the effectiveness of the two interventions.
It would also add to the knowledgebase if researchers used more sophisticated modeling procedures in their data analyses. These more sophisticated modeling techniques (e.g., structural equation modeling) could also address the problem of nested data. Nested data occur when you have more than one client receiving career counseling from one career counselor, which is often the case in many studies.
Another avenue of research concerns the content of the career counseling and its impact on the relationship between the working alliance and outcome. For example, it might be interesting to compare career counseling that includes the five critical ingredients of Brown and Ryan Krane (2000) with other forms of career counseling. In addition, although the life design approach has emphasized the importance of establishing a good working alliance at the beginning of the career counseling intervention (Savickas et al., 2009), the impact of this working alliance on life design interventions has not been yet examined. Moreover, social constructionism, social cognitive, or cognitive interventions might have differentiated impacts on the quality of the working alliance and, inversely, the working alliance might have a differentiated impact on the effectiveness of these different types of career interventions.
An additional issue in this area concerns when in the career counseling process, the working alliance is measured. In psychotherapy research, the therapeutic or working alliance is typically measured between the third and fifth session (Horvath, 2013). Career counseling is frequently shorter in duration than psychotherapy with Whiston (2002) finding that, on average, individual career counseling lasted only 1.35 sessions. For certain client issues (e.g., needing outlook information on a specific occupation), brief approaches may be the modality of choice. This raises the question of whether brief approaches versus longer, more comprehensive, and individualized approaches require the same level and intensity of the working alliance. Therefore, we suggest research that incorporates clients’ needs in evaluating the differing levels of the working alliance. More specifically, these studies might incorporate clients’ readiness to change indices as discussed by Sampson, McClain, Musch, and Reardon (2013).
In order to provide effective career intervention and contribute to social justice, it is important that these career interventions benefit all clients and especially those who need them the most. For this reason, more research should be conducted concerning the relationship between the client’s background and the working alliance. If clients with the most career decision-making difficulties develop less positive working alliance, specific techniques should be developed to help counselors develop strong working alliances. Moreover, some dysfunctional personality traits (e.g., schizotypal traits) may impact the ability to make a career choice, enter the labor market, and remain employed (Kjos, 1995; Skodol et al., 2002) and may also influence the development of the working alliance.
Finally, it could be interesting to evaluate whether the working alliance has an impact on the proportion of dropouts and on intervention adherence. It is hypothesized that a stronger working alliance would decrease the probability of a client dropping out of career counseling, but this has not been studied. Furthermore, we do not know if the working alliance actually contributes to clients’ abilities to make a career choice or on the actual implementation of that choice.
Conclusions
Studies summarized in this review of the working alliance in career counseling reflect that the working alliance typically increases over the course of career counselling. Moreover, a number of researchers found a significant relationship between the working alliance and the career counseling outcome. This relationship, however, needs to be further studied and analyzed. We suggest that after additional studies have been conducted that a meta-analysis of this relationship be conducted. In terms of practice, the current research indicates that career counselors should focus on establishing a working alliance in their work with clients and that the establishment of such an alliance should begin early in the process. Furthermore, there is a need for additional research that examines both client and counselor factors that contribute to the establishment of an effective working alliance in career counseling and the outcomes of those factors. There is also a need to expanded research in this area, particularly research that might involve experimental designs and more sophisticated methodologies.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.