Opening the Career Counseling Black Box: Behavioral Mechanisms of Empathy and Working Alliance

Abstract

Many individuals who struggle with career decisions seek professional guidance from career counselors. However, little is known about how to ensure that career guidance sessions are of high quality. Vital factors for a successful counseling process include a positive working alliance and empathy of the counselor. This study investigates how counselors’ behaviors, measured with a behavioral coding instrument, related to positive working alliances and empathy. We introduce a video-based instrument that measures microbehaviors and skills during career counseling sessions. Psychometric analyses were conducted on the basis of 53 video recordings. Double-coded sessions (N = 18) showed excellent interrater reliabilities for the coded microbehaviors. By using a multitrait-multimethod approach, we established construct validity with another coding instrument, whereas criterion validity was established with respect to counselors’ empathy and working alliance ratings. We discuss theoretical implications of using video-based methods in career sessions to advance process research and present practical applications in terms of assessing quality assurance in career interventions.

Keywords

career counseling empathy working alliance video methods observational analyses

Due to globalization, as well as to the increasing changes in work and forms of employment, the demands on graduates and young adults have increased enormously. Individuals entering the workforce often seek guidance for career decisions from professionals¹ (Gati & Levin, 2014). Professionals offering career guidance are found under many different names such as career counseling, career consultancy, guidance interviews, helping professionals, transition advisors, and career or life coaches (Chung & Gfroerer, 2003; Grant, Passmore, Cavanagh, & Parker, 2010; McMahon, 2004). They help individuals to make important career decisions, develop career-related goals (e.g., Grant et al., 2010), and improve overall career planning (Brown et al., 2003). More importantly, professional career advisors are estimated to make up to 0.8% of the entire labor force (Grant et al., 2010; Organisation for Economic Cooperation and Development [OECD], 2003; Whiston, 2003).

Whereas the backgrounds of these professionals are manifold—many having no career-specific training (OECD, 2003)—the question arises as to what these professionals actually do to help their clients to make important life decisions? To answer this, we need fine-grained taxonomies that capture counseling behaviors in situ, which will allow us to study counseling sessions with precision (Heppner & Heppner, 2003). Behavioral taxonomies can be analyzed based on videotaped counseling sessions (e.g., Multon, Ellis-Kalton, Heppner, & Gysbers, 2003), which offer methodological advances to survey-based approaches because the latter can be biased through enhanced self-perceptions or memory inaccuracies (Miller & Mount, 2001; Miller, Yahne, Moyers, Martinez, & Pirritano, 2004). To illustrate, Miller and Mount (2001) showed that counselors’ self-reports of counseling skills shared very little overlap with objective observations of counseling behaviors in practice. That is, counselors tended to largely overestimate their own skills when answering survey-based measurements.

With little knowledge about what happens within career counseling sessions, research must open the “black box” of the career counseling process (Heppner & Heppner, 2003, Whiston, 2003). Video-based approaches allow one to repeatedly inspect specific behaviors and interaction sequences (e.g., Klonek, Wunderlich, Spurk, & Kauffeld, 2016; Mittendorff, den Brok, & Beijaard, 2010). The current study aims to advance our understanding of career guidance sessions by introducing a video-based observational instrument.

First, we introduce a video-based instrument specifically adapted to the context of career counseling. This instrument is based on the Advanced Interaction Analysis for Consulting (act4consulting, short “ACT”; Hoppe, 2013) and captures practitioners’ microbehaviors and skills within the career guidance process. The behavioral categories (i.e., the counselor’s procedural and socioemotional skills) are theoretically based on control and self-determination theory (e.g., Gregory, Beck, & Carr, 2011; Markland, Ryan, Tobin, & Rollnick, 2005) and refer to a nondirective and respectful inquiry style of the counselor (Van Quaquebeke & Felps, 2016). Second, we provide detailed reliability analyses for the microbehaviors that the instrument captures. Third, we demonstrate construct and criterion validity based on technical (e.g., questions, paraphrasing, summaries) and relational outcome (i.e., empathy and working alliance) criteria. We focus on relationship quality criteria of the counselor–client interaction (i.e., empathy, Moyers & Miller, 2013; working alliance, Baron & Morin, 2009; De Haan, Duckworth, Birch, & Jones, 2013) because these are essential criteria for counseling success across a variety of settings such as psychotherapy, coaching, and career guidance (e.g., McKenna & Davis, 2009).

Reliability of Measures Obtained From Video-Based Instruments

As observational studies are labor-intensive research projects, the reliability of measures that are obtained with video-based instruments is of vital importance (Klonek, Quera, & Kauffeld, 2015; Multon et al., 2003). Accuracy of behavioral observations is usually assessed by using several observers and comparing their obtained scores with statistical measures of interrater reliability. Unfortunately, reliability of observational instruments is often not reported in career counseling studies (e.g., Multon et al., 2003). Another problem is that many researchers only report overall reliability measures (e.g., Gessnitzer & Kauffeld, 2015; Mittendorf et al., 2010), which lack information on category-specific reliability. This approach can be criticized for several reasons: First, an overall reliability index does not guarantee high reliability for specific behavioral categories that are captured by a coding instrument (Hewes, 1985). Second, this approach fails to inform other researchers about category-specific reliability (cf. Klonek, Quera, Burba, & Kauffeld, 2016). For example, Moyers, Martin, Catley, Harris, and Ahluwalia (2003) reported category-specific reliability estimates for all counselors’ microbehaviors of their observational instrument and showed that 50% of variables provided poor reliability. This underscores that high reliabilities in observational research should not be taken for granted. Low reliabilities can be circumvented by using a coding manual, appropriate observer trainings (Bakeman & Quera, 2011) and the integration of software support (Klonek et al., 2015). The current study applied these aforementioned quality measures (i.e., coding manual, observer training, and software support). Furthermore, we provide information on code-specific reliabilities.

Hypothesis 1 (H1): Microbehaviors and counseling skills obtained from a video-based career counseling instrument will provide reliable scores.

Convergent Validity of Video-Based Assessed Career Counselors’ Microbehaviors

Beside reliability, validity is crucial for a new coding system. Convergent validity can be estimated by comparing results obtained from a second instrument that captures similar measures. We selected this second coding instrument based on two criteria: First, there had to be scientific research using the second observational instrument in a career counseling context (Klonek et al. 2016; Rochat & Rossier, 2016). Second, there had to be data available on the second instrument that indicated strong psychometric properties. Using these criteria, we selected the Motivational Interviewing Treatment Integrity (MITI). First, the MITI has been used in career counseling research contexts: For example, in one study, it was used to evaluate whether career counselors meet quality benchmarks in Motivational Interviewing (MI), which is a client-centered intervention in which the counselor’s empathy is a key component (Klonek et al., 2016). The MITI allows us to measure both client-centered behaviors (open questions and reflective listening, which are important indicators of counselor empathy, Moyers, Martin, Manuel, Miller, & Ernst, 2010), neutral behaviors (giving information, closed questions) and behaviors that are specific to MI interventions (i.e., MI adherent behaviors, e.g., affirmations vs. MI nonadherent behaviors, e.g., confrontations). Second, due to the strong empirical research tradition in MI, there is a substantial amount of publications on the MITI, reporting excellent psychometric properties (e.g., Klonek et al., 2015; Moyers, Martin, Manuel, Hendrickson, & Miller, 2005). Finally, the MITI also overlaps conceptually with existing career counseling instruments such as the “revised hill counselor verbal response modes category system” (Hill, 1985; Multon et al., 2003). Due to these arguments and the ease of accessibility of the MITI (available at no charge at casaa.unm.edu), we used it to assess convergent validity. Overall, we hypothesize that:

Hypothesis 2 (H2): Objectively coded microbehaviors (i.e., close questions, open questions, paraphrasing, summarizing, appreciation, interruption of the client) will be positively associated with conceptually similar categories of the MITI (i.e., closed question, open question, simple reflection, complex reflection, MI adherent behavior, MI nonadherent behavior).

Criterion Validity of Video-Based Assessed Counseling Skills

With respect to criterion validity, we focus on two outcome variables: empathy and working alliance. Whereas empathy is considered to be a crucial success measure in therapeutic counseling contexts (Moyers & Miller, 2013), working alliance is an essential variable in the context of career counseling (Whiston, Rossier, & Barón, 2016).

In counseling research, counselor empathy can be conceptualized as interpersonal efforts to take the clients’ perspectives by trying to accurately understand their cognitive and emotional experiences (Moyers & Miller, 2013). A variety of therapy studies have indicated that empathy of counselors is an essential predictor of client-related therapy outcomes (e.g., Pantalon, Chawarski, Falcioni, Pakes, & Schottenfeld, 2004; Ritter et al., 2002). For example, Pantalon, Chawarski, Falcioni, Pakes, and Schottenfeld (2004) showed that rated counselor empathy, based on videotaped sessions, prospectively explained objective measures of their respective clients’ health. Building on this evidence, Moyers and Miller (2013) concluded that low counselor empathy might even be toxic for clients. Whereas psychotherapy research contributed substantially to a better understanding of how counselor skills affect patient outcomes, research in career counseling has less to offer concerning these issues (Whiston & Oliver, 2005). Based on the evidence from therapy research, we assume that counselor empathy in career interventions provides an equally important ingredient for helping clients to deal with career-related problems. Taken together, counselors’ behavioral skills should be related to how empathetically they approach their client and how well they deal with sensitive content that could occur in the session (Hirschi, 2017). In general, counselor skills associated with empathy comprise the expression of understanding and caring (e.g., De Haan, 2008, Rogers, 1973). Typical behaviors associated with empathy are, for example, paraphrasing or addressing the client’s feelings (e.g., Rogers, 1973, Korman, Bavelas, & De Jong, 2013). These counselor skills can be termed as functional socioemotional skills. We expect that functional socioemotional counselor skills measured with objective video recordings will contribute toward counselor empathy. Thus, we hypothesize:

Hypothesis 3 (H3): There will be a positive relationship between functional socioemotional skills and empathy.

Counselors may also display insensitive behaviors and thereby harm the relational empathic bond with the client. Instead of active listening, paraphrasing, and supporting the client, the counselor may be distracted, express impatience, or even displaying lacking interest. We consider such adverse behaviors as dysfunctional. Thus, we hypothesize:

Hypothesis 4 (H4): There will be a negative relationship between dysfunctional socioemotional skills and empathy.

The Importance of Working Alliance in Career Counseling

The working alliance between counselor and client has received increased attention in career counseling research (Whiston, Rossier, & Baron, 2016). Whiston et al. (2016) reviewed 17 working alliance studies and showed that it positively related with various client outcome measures (e.g., career maturity, job and life satisfaction, or vocational self-efficacy). Despite knowledge about the various consequences of a good working alliance, we know significantly less about the counseling behaviors that positively predict a working alliance (cf. Whiston et al., 2016).

A counseling working alliance characterizes an affective bond with a strong focus on tasks and goals (McKenna & Davis, 2009). The affective bond is strongly associated with constructs such as trust (O’Broin & Palmer, 2010), interpersonal comfort, and mutual sympathy (Baron & Morin, 2009). Certain behaviors are generally considered beneficial to the development of an affective bond (e.g., demonstrating empathy, encouragement, or appreciation; e.g., Gregory & Levy, 2011; O’Broin & Palmer, 2010). The affective bond seems to provide the necessary framework in which the client feels secure enough to open up and to being self-reflective (Greif, 2007; Ianiro, Lehmann-Willenbrock, & Kauffeld, 2014). However, the concept of working alliance goes beyond the affective bond between counselor and client: It also relates to the degree of mutual agreement as to the goals to be attained, as well as to the ways and means of attaining them (Baron & Morin, 2009). Thus, the construct working alliance also refers to the quality of the counseling procedure. There are several microbehaviors that can be related to a high-quality counseling procedure. Research linking control theory and counseling has stressed the importance of a counselors’ goal orientation, the ability to provide structure within the process, and support for the client’s self-reflection (e.g., Grant, 2005; Gregory et al., 2011). To achieve this, counselors need to display procedural skills, such as good time management, suggesting the next procedural step, and staying focused within the session (e.g., Grant, 2007). Another important counseling behavior is the use of open questions because they stimulate self-awareness and require personal responsibility of the client (Passmore & Fillery-Travis, 2011).

Taken together, we term these behaviors associated with the quality of the counseling procedure as functional procedural skills. We expect that functional procedural skills positively contribute to a working alliance. Thus, we hypothesize:

Hypothesis 5 (H5): There will be a positive relationship between functional procedural skills and working alliance.

For many counselors, it is often difficult to use open, instead of closed, questions within the session (Miller & Mount, 2001) and to prevent themselves from quickly suggesting solutions for the client (Klonek et al., 2015). However, in nondirective counseling, it is particularly important for counselors to be able to step back and let the client find his or her own solutions (e.g., Markland et al., 2005). We expect that counselor’s behaviors, such as proposing a problem or solution instead of encouraging the client to find the solution for himself or herself, are dysfunctional in the counseling process and thereby negatively affect the working alliance.

Hypothesis 6 (H6): There will be a negative relationship between dysfunctional procedural skills and working alliance.

Whereas empathy reflects the affective-relational bond between counselors and clients, working alliance is most strongly grounded in goal- and task-related behavior. A coding instrument used for career counseling needs to address both affective goal- and task-related behavior, as well as relational aspects, such as empathy. As our instrument allows for assessing a counselor’s behavior related to goals, and a client’s progress, we expect that—in the context of counseling—it exceeds explanatory power of existing coding instruments that were developed for a specific intervention, and which have a more narrow focus (such as the MITI). Thus, we compared the explanatory variance of our instrument with the MITI and hypothesize:

Hypothesis 7 (H7): Counselor skills (measured with ACT) explain working alliance (in terms of goals and tasks) above and beyond client-centered and motivational counselor skills (measured with the MITI).

Method

Sample

The current analyses are based on a data set of 53 career counseling sessions (with unique counselor–client dyads). Data were collected at two German universities from 2008 to 2013. The counselors were young professionals with a bachelor’s degree in psychology and extensive training in career counseling (approx. 200 hr) over a period of 1 year that covered theoretical (learning about counseling, literature study) and practical (peer coaching) components. The counselors were, on average, 25 years old (SD = 3.64), and the majority was female (92.5%, N = 49), which corresponds to a growing female gender distribution of graduates in psychology (Ostertag & McNamara, 1991). The clients’ sample consisted of young professionals or students with a bachelor’s degree before finishing their master’s degree. The mean client age was 25 years (SD = 3.11), and they showed a balanced gender distribution (56.6% female clients, N = 30).

Procedure

The clients were recruited via advertisement or via the website of the two institutes. They could participate if they had enrolled in one of the two universities that offered the career counseling program. After giving written consent to be videotaped, counselors and clients (who did not know each other before) were randomly assigned to each other. The entire career guidance program covered five well-structured sessions that each took about 90 min.

Video Observation of Counselor Behaviors and Skills

We analyzed the first session because it builds the baseline for establishing a working relationship with the client. Recorded counselor behaviors were coded using the ACT² instrument (Gessnitzer & Kauffeld, 2015, Hoppe, 2013), which is mutually exclusive and exhaustive in that every observational unit is assigned to only one code (Bakeman & Quera, 2011). Videotaped sessions were subdivided into units, each representing the smallest meaningful communicative statement that could be understood by the other member of the interaction. We were interested in microbehaviors of counselors that are considered important for the development of the affective bond, such as active listening, or demonstrating empathy or encouragement (e.g., Duckworth & De Haan, 2009; O’Broin & Palmer, 2010). Further, we coded microbehaviors associated with a counselor’s goal and task orientation, which support clients’ self-regulation (Gregory et al., 2011). Counseling approaches assuming clients’ innate tendency for personal growth, and their own expertise over their issue, emphasize the importance of a counselor’s nondirective and respectful inquiring attitude toward the client (e.g., Markland et al., 2005; Van Quaquebeke & Felps, 2016). In terms of respectful inquiry, three components are important such as asking questions, question openness, and attentive listening (Van Quaquebeke & Felps, 2016). Markland, Ryan, Tobin, and Rollnick (2005) further described counselors’ behavior that provides structure (e.g., helping clients to develop appropriate goals or giving them feedback) and support a client’s autonomy (e.g., exploring options with the client and avoiding coercion). Following this line of reasoning, we focused on counselors’ microbehaviors that showcase a degree of structure, respectful inquiry and directivity. We selected functional and dysfunctional socioemotional and procedural microbehaviors associated with bonding, structure, and respectful inquiry and directivity.

Examples for functional socioemotional skills include establishing contact (e.g., “Hi, was it easy to find your way here?”), bonding (e.g., “In case of emergency, you can call me”), and facilitating behaviors, such as signs of listening (e.g., “Hmm, yes”), addressing client’s feelings (e.g., “You felt excited in this moment”), and appreciation (e.g., “Well done, you are very ambitious!”). Dysfunctional socioemotional skills are, for example, interruptions of the client or disruptive behavior (e.g., “What you told me is not important to me!”).

Functional procedural skills cover behaviors that help to structure sessions such as goal orientation (e.g., “Today, we will talk about your career planning”), time management (e.g., “We still have 10 min left for…”), or open questions (e.g., “What would help you to enhance your person–job fit?”). Dysfunctional procedural skills include losing the train of thoughts and running off topic (e.g., “My cousin also works for this company.…Oh, I need to call him later”). Moreover, directive behaviors, such as suggesting or explaining solutions for the client, are considered dysfunctional (e.g., “An internship in this company looks like a solution to your problem”).

Observer Training

Before coding the analysis sample, three observers received 200 hr of training. All observers coded a training sample and were only allowed to code the analysis sample when their interreliability exceeded κ ≥ .70, that is, after having a substantial agreement (Landis & Koch, 1977). Biweekly and supervised observer meetings gave coders the possibility of addressing questions, deciding on rules, and solving disagreements. All observers kept a diary to note specifics, which were discussed within these regular meetings in order to reach coding accuracy.

Behavioral Measures for Construct Validity

Behavioral measures

Convergent validity was assessed using the MITI (German Version MITI-d; Brueck et al., 2009; Klonek et al., 2015), which is a behavioral coding instrument to assess behavioral skills in the psychological approach of MI (Miller & Rollnick, 2013). Although MI coding instruments are mainly used for MI-based interventions, they include conceptually overlapping microbehaviors with ACT (e.g., open/closed questions, summarizing/reflecting are codes that exist in both MITI and ACT).

A new pair of observers used a software-supported MITI-d Version (Klonek et al., 2015) to code the seven behavioral categories (open/closed questions, simple/complex reflections, MI adherent/nonadherent behavior). Before coding, they also received a coder training for the MITI-d, that is, they read theoretical literature, studied the coding manual (Brueck, Frick, & Loessl, 2006), and participated in gradual learning tasks (i.e., we used coded transcripts provided in Klonek et al., 2015). None of the training material was used in subsequent analysis. Both observers had the possibility to ask the coding coordinator (T.W.) to discuss coding discrepancies. To assess reliability of the MITI-d, we calculated the intraclass correlations (ICCs, Shrout & Fleiss, 1979) for 53 double-coded samples. All seven MITI categories (ICCs > .84) achieved excellent levels of interrater reliability (Cicchetti, 1994).

Outcome Measures for Criterion Validity

Empathy of practitioner

We used a Global Rating Scale from the MITI-d (i.e., gestalt ratings) to assess counselors’ empathy. This global rating is based on practitioner’s behavior of a 20-min sequence and aims to capture an overall impression (i.e., “gestalt rating”) using a 7-point scale to evaluate the extent of counselor empathy (1 = low empathy to 7 = high empathy). High empathy ratings were given when practitioners showed “an active interest in making sure they [understood] what the client [was] saying,” whereas low empathy ratings were given when there was “little interest in the point of view and in the experiences of the client” (Brueck et al., 2006, p. 2). Interrater reliability for 53 empathy ratings was excellent (ICC = .84). To further reduce the measurement error of our rating measure, we used the average empathy ratings from both observers as an outcome variable in subsequent analyses.

Working Alliance Inventory (WAI)

We used the short WAI-S version from Tracey and Kokotovic (1989). Items for the short version (WAI-S) and its three subscales—Bonding, Goal Orientation, and Task Orientation—were obtained from a factor analysis of the WAI (cf. Tracey & Kokotovic, 1989). Furthermore, this short version has been used in previous studies (e.g., Gessnitzer & Kauffeld, 2015; Multon et al., 2003). The WAI-short scales each comprise 4 items and were originally developed as a self-report measure. As we applied an independent observers’ perspective to infer working alliance directly from the videos (cf. Meinecke, Klonek, & Kauffeld, 2016), items were adapted accordingly (e.g., “Counselor/coach and client agree on the things the client will need to do to help improve his/her situation”). Items were scored from both observers using a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree).

Cronbach’s α for both observers was α = .71/.81, and interrater reliability had an ICC of .83. As both observers rated all 53 sessions, we used the observer-mean WAI ratings as a dependent variable in subsequent analyses to further reduce measurement error. (Note that observers who rated empathy and working alliance were not the same observers who coded the counselors’ microbehaviors.)

Overview of Statistical Plan

Reliability analysis

Reliability for behavioral codes was estimated by calculating ICCs for 18 double-coded sessions (Fleiss & Shrout, 1978). The ICC is a statistical index commonly used to estimate reliability because it adjusts for chance agreement and systematic differences between observers (Fleiss & Shrout, 1978); it is also a more conservative estimate than the Pearson correlation. We computed absolute (criterion-referenced) agreement between two observers using the following formula (Bakeman & Quera, 2011), which is stricter than the relative (norm-referenced) agreement: $ICC = [M S_{b} - M S_{e}] / [M S_{b} + M S_{e} - (2 / n) (M S_{o} - M S_{e})]$ , where n is the number of observations, or summary values, obtained by each observer; MS_b is the between-measures mean square; MS_o is the observer, or repeated measures, mean square; and MS_e is the error mean square. According to Cicchetti (1994), we classified ICCs by cutoff criteria: below 0.40 = poor, 0.40–0.58 = fair, 0.60–0.74 = good, and 0.75–1.00 = excellent.

Validity analysis

Convergent validity was assessed by comparing 53 sessions that were coded with both ACT and the MITI-d by independent sets of coders for each instrument. First, we selected conceptually overlapping codes. Second, we derived session summary counts for each behavioral code for both instruments for all 53 sessions. Finally, we calculated correlations of behavioral code measures between both instruments.

For convergent validity (H2), we expected high correlations for conceptually similar codes from ACT and MITI, such as ACT-closed questions with MITI-closed questions, ACT-open questions with MITI-open questions, ACT-paraphrase with MITI-simple reflections, ACT-summary with MITI-complex reflections, ACT-interrupt with MITI–MI nonadherent behavior, and ACT-appreciation with MITI–MI adherent behavior. In terms of discriminant validity, we expected close-to-zero and nonsignificant correlations for codes that were conceptually different.

For criterion validity, we expected that socioemotional skills explain variance in empathy (H3 and H4). Furthermore, we expected procedural behavioral skills to explain variance in working alliance, particularly with regard to goal and task orientation (H5 and H6). Finally, we expected that ACT counseling skills explain variance above and beyond behavioral MITI measures (H7).

Results

Reliability

Sample size for reliability analysis was chosen a priori following two guidelines: Bakeman, Deckner, and Quera (2005) recommend a sample of 15–20% from the corpus to check reliability³. Furthermore, for the calculation of ICCs, a minimal sample of >5 is recommend—with 10 or more sessions resulting in more robust results (Bakeman & Quera, 2011). To fall above these recommended minimal criteria for reliability analyses, we randomly selected a subsample of N = 18 (i.e., 33%) sessions.

H1 stated that microbehaviors and counseling skills obtained from a video-based career counseling instrument will provide reliable scores. Table 1 shows reliabilities for all behaviors as well as for the respective superordinate counseling skills. On average, microbehavioral codes showed M _ICC = .77 (SD _ICC = .22), that is, code-specific reliability mostly showed excellent agreement (Cicchetti, 1994). Measures of functional and dysfunctional socioemotional skills all showed excellent interrater reliability (i.e., ICCs > .74), according to Ciccetti’s criteria (1994)—with the exception of the codes humor (ICC = .73), addressing client’s feelings (ICC = .71), disagreement/criticism (ICC = .65), and summaries (ICC = .35). Measures of functional and dysfunctional procedural counseling skills showed good to excellent reliabilities (ICCs > .63)—with the exception of clarifying one’s own contributions (ICC = .48) and goal orientation (ICC = .06).

Table 1.

ACT Coding Scheme for Career Counselor Skills, Underlying Microbehaviors, and Reliabilities.

Skills and Microbehaviors	ICC	Classification^a	M (SD)
Socioemotional skills (+)	0.75	Excellent
Contact	0.93	Excellent	0.77 (0.9)
Paraphrasing	0.94	Excellent	17.11 (11.85)
Summaries of client statements	0.35	Poor	2.71 (2.51)
Humor	0.73	Good	3.83 (4.71)
Bonding	0.81	Excellent	6.78 (6.6)
Addressing client’s feelings	0.71	Good	2.84 (3.07)
Appreciation	0.77	Good	7.99 (6.52)
Socioemotional skills (−)	0.90	Excellent
Interruption of the client	0.94	Excellent	3.09 (3.74)
Disruptive behavior	1.00	Excellent	0.09 (0.25)
Self-promotive behavior (i.e., self-praise)	1.00	Excellent	0.13 (0.72)
Disagreement/criticism	0.65	Good	0.88 (1.12)
Procedural skills (+)	0.66	Good
Open question	0.98	Excellent	36.2 (16.46)
Goal orientation	0.06	Poor	1.35 (1.95)
Clarifying own contributions	0.48	Fair	10.55 (10.81)
Procedural suggestion	0.93	Excellent	43.62 (19.17)
Prioritizing	0.63	Good	0.94 (1.37)
Time management	0.78	Excellent	0.87 (1.29)
Visualization with media	0.76	Excellent	8.68 (8.86)
Procedural skills (−)	0.84	Excellent
Closed question	0.94	Excellent	35.5 (14.19)
Pointing out a problem/suggesting or explaining a solution	0.85	Excellent	2.52 (3.31)
Losing train of thought and running off topic	0.72	Good	2.39 (4.9)

Note. Code frequencies are standardized to counts per hour. ACT = Advanced Interaction Analysis for Consulting; ICCs = intraclass correlations.

^aCategorization based on Cicchetti (1994) for ICCs, (+) = functional, (−) = dysfunctional.

Following procedures from Moyers et al. (2003), we collapsed behavior codes that theoretically belong to a superordinate category (i.e., counseling skill) and calculated interrater reliability estimates for these categories as well. Reliabilities for these four superordinate counselor skill measures were in the good to excellent range (ICCs = .66–.90; cf. Table 1), thus supporting H1.

Construct Validity

For convergent validity (H2), we compared ACT codes with conceptually similar and different codes from the MITI. In H2, we stated that objectively coded microbehaviors (i.e., close questions, open questions, paraphrasing, summarizing, appreciation, interruption of the client) will be positively associated with conceptually similar categories of the MITI (i.e., closed question, open question, simple reflection, complex reflection, MI adherent behavior, MI nonadherent behavior).

Table 2 shows correlations for MITI codes with conceptually similar ACT codes. The table also contains a monotrait heteromethod block from the multitrait, multimethod (MTMM) matrix (Campbell & Fiske, 1959). Conceptually overlapping codes (from both instruments) are ordered in the same way, such that the diagonal allows estimating convergent validity, whereas the off-diagonal cells indicate discriminant validity results.

Table 2.

Construct (Convergent and Discriminant) Validity for Counselors’ Microbehaviors Measured With Two Conceptually Related Coding Instruments.

Codes	MITI- Closed Question	MITI-Open Question	MITI-Simple Reflection	MITI-Complex Reflection	MITI–MI Nonadherent Behavior	MITI–MI Adherent Behavior
ACT-closed question	.55**	.12	.03	.02	.09	.01
ACT-open question	.03	.56**	−.02	.19	−.19	−.16
ACT-paraphrase	−.07	.13	.25^†	.16	.17	−.03
ACT-summaries	−.11	.12	−.03	.18	−.19	−.25
ACT-interrupt	.02	−.10	.24	.06	.54**	.14
ACT-appreciation	.05	.00	.16	.25	.05	.33*

Note. N = 53. MITI = Motivational Interviewing Treatment Integrity; MI = Motivational Interviewing; ACT = Advanced Interaction Analysis for Consulting.

^† p < .10. *p < .05. **p < .01 (two-tailed).

The MTMM diagonal of Table 2 shows significant correlations for the following ACT and MITI variables: ACT-closed questions with MITI-closed questions (r = .55, p < .01), ACT-open questions with MITI-open questions (r = .56, p < .01), ACT-paraphrase with MITI-simple reflection (r = .25, p = .07), ACT-interrupt with MITI–MI nonadherent (r = .54, p < .01), and ACT-appreciation with MITI–MI adherent (r = .33, p < .05). Only ACT-summaries showed a nonsignificant correlation with MITI-complex reflection (r = .18, p = .19), which might be explained due to low reliabilities for the ACT-code (summaries; reliability determining an upper bound for validity estimates, Bakeman & Quera, 2011).

Overall, these correlations support the convergence of central ACT codes with MITI-codes. Conversely, off-diagonal cells in Table 2 showed nonsignificant correlations that are, on average, close to zero (M_r = .02), thus supporting discriminate validity. By providing empirical results for both convergence and discrimination, these results support H2 (i.e., ACT codes showing convergent and discriminant validity with MITI-codes).

Criterion Validity

To test criterion validity, we performed multiple linear regression analyses for two outcome criteria: counselor empathy and working alliance.

Counselor Empathy

Predictors for empathy were functional and dysfunctional socioemotional skills (F = 2.2, R ² = .08). In H3, we stated that there will be a positive relationship between functional socioemotional skills and empathy. Results indicated that functional socioemotional skills positively predict counselor’s empathy (β = .30, p < .05), thus supporting H3. In H4, we stated that there will be a negative relationship between dysfunctional socioemotional skills and empathy. Contrary to expectations, dysfunctional socioemotional skills were not significantly related to empathy (β = −.17, p > .05). Thus, H4 was not supported.

Working Alliance

Predictor variables for working alliance were functional and dysfunctional procedural skills (F = 3.5, R ² = .12). In H5, we stated that there will be a positive relationship between functional procedural skills and working alliance. Functional procedural skills did not significantly predict working alliance (β = .22, p > .05), thus not supporting H5. In H6, we stated there will be a negative relationship between dysfunctional procedural skills and working alliance. Dysfunctional procedural skills negatively predicted working alliance (β = −.37, p < .05), thus supporting H6.

To test H7, we performed hierarchical multiple regression analyses for the working alliance scales goals and tasks (Table 3). Predictors were entered in three steps: In Model 1, we entered the session duration as a control variable. In Model 2, we entered the behavioral MITI measures. In the Model 3 (which tested H7), we entered procedural counseling skills measured with ACT.

Table 3.

Multiple Hierarchical Regression Using MITI and ACT Measures for Working Alliance Subscale Goals and Tasks.

Predictors	Working Alliance (Goals)						Working Alliance (Tasks)
Predictors	F	R ²	ΔR ²	β	SE_B	β	F	R ²	ΔR ²	β	SE _B	β
Model 1
Session duration	0.27	.01	.01	−.00	.31	−.07	.00	.00	.00	.00	.00	−.01
Model 2	7.98**	.33	.32**				10.82**	.40	.40**
Session duration				−.01	.00	−.33*				−.01	.00	−.28*
MITI(+)				.04	.01	.44**				.05	.01	.54**
MITI(−)				−.06	.02	−.41**				−.06	.02	−.39**
Model 3	6.09**	.39	.07^†				8.41**	.47	.07*
Session duration (covariate)				−.01	.00	−.31^†				−.00	.00	−.14
MITI(+)				.04	.01	.46**				.06	.01	.60**
MITI(−)				−.03	.02	−.23				−.03	.02	−.20
ACT procedural(+)				.00	.00	.17				−1.59	.00	−.00
ACT procedural(−)				−.01	.01	−.33*				−.01	.01	−.35*

Note. N = 53. MITI = Motivational Interviewing Treatment Integrity; MITI(+) = MITI functional skills (open question, simple and complex reflection, MI adherent behavior); MITI(−) = MITI dysfunctional skills (closed question, MI nonadherent behavior); Advanced Interaction Analysis for Consulting (ACT) procedural(+) = ACT functional procedural skills; ACT procedural(−) = ACT dysfunctional procedural skills.

^† p < .10. *p < .05. **p < .01 (two-tailed).

In H7, we stated that counselor skills (measured with ACT) explain working alliance (in terms of goals and tasks) above and beyond client-centered and motivational counselor skills (measured with MITI). With respect to working alliance (i.e., goals), Models 2 and 3 significantly explained variance in this outcome variable (Model 2: F = 7.98, df = 3, p < .01, R ² = .33; Model 3: F = 6.09, df = 5, p < .01; R ² = .39). Further, inclusion of procedural counseling skills (measured with ACT) explained incremental variance in working alliance goals above and beyond the MITI observational measures (ΔR ² = .07, p < .10), thus supporting H7.

With respect to working alliance (i.e., tasks), Models 2 and 3 significantly explained variance in this outcome variable (Model 2: F = 10.82, df = 3, p < .01, R ² = .40; Model 2: F = 8.41, df = 5, p < .01; R ² = .47). Counseling skills (measured with ACT) explained incremental variance in working alliance (tasks) above and beyond the MITI measures (ΔR ² = .07, p < .01) and further supported H7.

Discussion

The current research study investigated the psychometric properties of a video-based instrument for career counselors. By using behavioral analyses of counselors in action, our aim was to better understand the microbehaviors and counseling skills that contribute to empathy and working alliance in career guidance sessions. Overall, two main findings accrued from this study.

First, we provided detailed, category-specific reliability analyses for both microbehaviors and counseling skills, showing that the video-based instrument can be used to reliably code microbehaviors of career counselors.

Second, results from construct validity analyses showed that counselor microbehaviors of the coding instrument not only conceptually, but also empirically, overlap with microbehaviors that are assessed by an established instrument (i.e., the MITI). In addition to construct validity, our study supported criterion validity of measures obtained from the instrument, whereas functional socioemotional skills positively predicted counselor’s empathy, dysfunctional procedural skills significantly and negatively predicted working alliance, as rated by independent observers. Moreover, our results suggest that the measures from the instrument explained incremental variance in working alliance above and beyond established MITI measures.

Theoretical Implications

Unraveling the career counseling process is a major and current challenge with regard to counseling professionalization (cf. OECD, 2003, Grant et al., 2010). This is where our findings, based on behavioral observations, may make an important contribution. First, our results show that relevant counselor behaviors can be observed systematically and reliably over the course of the interaction with their clients. In line with several authors (e.g., Heppner & Heppner, 2003; Miller et al., 2004), our study suggests that genuine career guidance process measures are well suited to capture microbehaviors of career counselors during a counseling session. Thus, researchers who want to better understand the behavioral ingredients that actively contribute to good career interventions can use the video instrument for these purposes. For example, we have no systematic knowledge of whether professionals trained in specific theoretical approaches (e.g., systems theory framework, Patton & Mahon, 2006, or theory of career construction, Rehfuss, Del Corso, Galvin, & Wykes, 2011) display characteristic behavioral profiles within their sessions. However, this knowledge would allow us to assess training effectiveness in different theoretical schools of thought and allow us to develop benchmarks or empirical comparisons of how different career theories manifest into specific behavioral counseling profiles. Furthermore, the software integration of the behavioral instrument automatically records on- and offsets for all microbehaviors. By means of integrating temporal processes into counseling research, researchers can better pinpoint temporal sequences that are crucial for counselor–client interactions.

Second, our study also contributes to a better understanding of behavioral skills that are related to counselor empathy and working alliance. Previous therapy research has suggested that empathy is a crucial ingredient within a counseling relation that positively impacts client health and adherence to treatment (Moyers & Miller, 2013). In line with several authors (e.g., De Haan, 2008; De Haan, Culpin & Curd, 2009; Rogers, 1973), the current study has shown that functional socioemotional skills in particular, such as using summaries, bonding with the client, or appreciation, positively and significantly predict counselor empathy. Similar to findings from psychotherapy (Sultanoff, 2003), our results suggest that humorous expressions are also related to empathic behavior.

Contrary to expectations, dysfunctional socioemotional skills were not significantly related to empathy. Interrupting the client or disruptive counselor behavior did not contribute to lower ratings on empathy or working alliance. It is possible that such behaviors—rather than being considered as real disruptions—are considered as supportive or essential for successful processes, if well dosed and applied in a sensitive way. Counselors are responsible for structuring and controlling the process and therefore may sometimes have to interrupt clients from time to time (Ianiro, Schermuly, & Kauffeld, 2013). However, behaviors, such as disagreement or criticism, were negatively and significantly related to working alliance (in terms of agreeing about tasks and goals). These findings support the assumption that empathy versus working alliance cover different and complementing aspects of the counselor–client relationship.

Another potential explanation for the lacking association between dysfunctional socioemotional skills, empathy, and working alliance may be that socioemotional skills had a very low mean and small standard deviation, whereas working alliance and empathy scores were above average in our sample of counselors. These statistics suggest a restricted variance in socioemotional skills and further, that counselors in our study were generally above average (in terms of working alliance and empathy). A reduced range of a variable can lessen the strength of an association with another variable, and this interpretation may offer an alternative explanation why we did not find support for H4 and H5.

Interestingly, findings on dysfunctional procedural behaviors appear to be more insightful compared to functional procedural skills. As expected, closed questions and directive behaviors were negatively related to working alliance. The latter results might indicate that effective counselors need to help their clients in a client-centered fashion, so that clients (rather than the counselors) discover the problems and voice their own solutions (cf. Markland et al., 2005; Van Quaquebeke & Felps, 2016). By contrast, counselors who engage in problem-solving themselves, and suggest solutions, might undermine their clients’ active reflection and understanding of what drives their career goals, as well as the obstacles that they are facing. Counselors who heavily engage in problem-solving instead of helping their client to work out their problems might miss out on opportunities for their client to have both innovative moments and self-exploration of the pros and cons of a specific career dilemma (Rochart & Rossier, 2016).

Taken together, these findings suggest differential effects for specific counselor behaviors. To evaluate which counselor behavior is beneficial for counseling success, a microanalysis of specific behaviors appears to be a promising approach. Moreover, the results also suggest that a distinction between functional and dysfunctional skills may be too simplistic. It is conceivable that some of the chosen microbehaviors are neutral, or only have a negative impact if they are applied too often, or expressed in an extreme way (e.g., interrupting the client). Still other behaviors may only have a negative impact if missing (e.g., open questions). The impact of some microbehaviors may only be fully understood when the corresponding nonverbal expressions are considered as well. Overall, we are only beginning to understand the interactional and behavioral dynamics in career interventions. The current study provides a tool that magnifies the behavioral aspect of effective career counseling.

Finally, the current study provides an observational instrument that can inform research about behavioral processes in career guidance interventions. This is an important contribution, especially because scholars have repeatedly criticized the overreliance on self-report survey measures because these can be flawed by social desirability, memory distortions, or limited attention spans (Miller & Mount, 2001). There is strong evidence that self-reported behavior is often little, or not at all, related to how participants behave if we observed them (Miller & Mount, 2001).

Practical Implications

The current study provides important implications for career counseling practice. First, our study informs career practitioners that displaying socioemotional behaviors (e.g., paraphrasing client statements, showing appreciation, and building contact) will help them to convey empathy toward their clients. Furthermore, practitioners learn from our study that interrupting or disagreeing with clients does not necessarily reduce their level of empathy, but that it is more important to engage in positive socioemotional behaviors in this context. By contrast, practitioners who want to focus on building a good working alliance with their clients are advised to refrain from displaying disruptive behaviors or frequently interrupting their clients (i.e., dysfunctional procedural skills). For the working alliance, displaying functional procedural behaviors (e.g., prioritizing or making procedural suggestions) seems to be less important. Taken together, our study implies that practitioners who want to both build a good working relationship and express empathy toward their clients should both display functional socioemotional skills and refrain from displaying dysfunctional procedural skills.

Second, our study directs practitioners to use video-based analyses to better understand their own career counseling approach. By coding their own sessions, they have a tool to systematically understand their microbehavioral repertoire within a session. Practitioners with interests in co- or supervision can use the instrument as a means of providing high-quality feedback. As coded recordings can be easily played back by means of the software, supervisors can give feedback based on visible behavioral sequences, which facilitates discussions with counselors-in-training. As a result, practitioners have a tool to reflect and improve their skills.

Third, agencies providing career counseling training can use video-based analyses for evaluation purposes. Although counseling approaches in psychotherapy have been using observational instruments for nearly two decades to evaluate how trainings affect practitioners’ counseling behaviors (De Roten, Zimmermann, Ortega, & Despland, 2013; Miller & Mount, 2001), these assessments are still relatively uncommon in career research (Hirschi, 2017). We presented a tool that training agencies can use to evaluate specific skills in career counseling, and knowledge of how these effect the relationship quality. As self-reported skills do not share much overlap with objectively assessed behavioral skill measures (Miller et al., 2004), the video-based instrument can provide a more objective quality assurance of career counseling skills (McMahon, 2004).

Limitations and Future Research

We acknowledge several limitations of this study. First, our study has used a relatively homogenic sample of young counselors with a limited amount of counseling experience. This implies a limited generalizability of our findings, and we can only cautiously infer how far our results also generalize to more age-diverse samples. Therefore, we encourage future studies to use video-based analyses in samples that show more variance in terms of age, gender, and experience. As the profession of career counselors constitutes a large variety of professionals, we also expect that behavioral profiles of career counselors with different education and training should differ. Nonetheless, and to the authors’ knowledge, the current study thus far provides the largest reported study that has used behavioral observation of career counselors—therewith exceeding the study from Multon, Ellis-Kalton, Heppner, and Gysbers (2003), who observed behavioral responses of 19 career counselors-in-training.

Second, some behavioral categories showed low reliabilities. Even though our study is not the only study that has reported problems with low reliabilities classifying counselor behaviors (e.g., Moyers et al., 2003, reported poor reliabilities for 50% of counselor behaviors), the aggregate counseling behaviors all provided good to excellent reliability. Furthermore, our study provides a first orientation in category-specific reliability and therewith indicates which behaviors are potentially difficult to assess for observers. Future studies can use our results to focus on behavioral categories and different nuances of nonverbal behaviors that potentially need stronger emphasis during observer training.

Third, in terms of behavioral focus, our study only investigated behaviors of the career counselor. However, there is evidence that suggests that behaviors of the counselor directly interact with clients’ motivational response within the session (cf. Klonek et al., 2016). Future studies should use behavioral analyses of career counselors to investigate the interactional dynamics between counselor and clients by means of advanced statistical methods such as sequential or analysis of behavioral patterns (Bakeman & Quera, 2011; Meinecke et al., 2016).

Fourth, our study does not provide information on how counselor skills affect client development outcomes (e.g., improved career planning, achievement of career goals, or higher self-efficacy in career decisions). We encourage future studies to use the new tool presented here and link counselor behaviors to effectiveness measures of career interventions. We would hypothesize that functional counselor skills positively relate to outcomes such as clients’ career planning behaviors. This research could further extend the applicability of this tool and inform career professionals.

Conclusion

The current study introduced a video-based instrument and examined the microbehaviors and counseling skills of career practitioners. By analyzing recorded career guidance sessions, we have shown that video-based scores are reliable and valid. Counselor procedural skills contributed to empathy and working alliance. A better understanding of microbehaviors in career guidance can contribute both to training and practice, as well to our theoretical understanding of effective career guidance.

Footnotes

Authors’ Note

Florian Klonek and Theresa Will, both authors, are first authors with similar parts of contribution.

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.

ORCID iD

Theresa Will

Notes

References

Bakeman

Deckner

D. F.

Quera

(2005). Analysis of behavioral streams. In Teti

D. M.

(Ed.), Handbook of research methods in developmental science (pp. 394–420). Oxford, England: Oxford.

Bakeman

Quera

(2011). Sequential analysis and observational methods for the behavioral sciences. Cambridge, England: Cambridge University Press.

Baron

Morin

(2009). The coach–coachee relationship in executive coaching: A field study. Human Resource Development Quarterly, 20, 85–106. doi:10.1002/hrdq.20009

Brown

S. D.

Ryan Krane

N. E.

Brecheisen

Castelino

Budisin

Miller

Edens

(2003). Critical ingredients of career choice interventions: More analyses and new hypotheses. Journal of Vocational Behavior, 62, 411–428. doi:10.1016/S0001-8791(02)00052-0

Brueck

R. K.

Frick

K. M.

Loessl

(2006). The motivational interviewing treatment integrity code, German version (MITI-d). Unpublished manual.

Brueck

R. K.

Frick

Loessl

Kriston

Schondelmair

… Berner

(2009). Psychometric properties of the German version of the motivational interviewing treatment integrity code. Journal of Substance Abuse Treatment, 36, 44–48. doi:10.1016/j.jsat.2008.04.004

Campbell

D. T.

Fiske

D. W.

(1959). Convergent and discriminant validation by the multitrait–multimethod matrix. Psychological Bulletin, 56, 81–104. doi:10.1037/h0046016

Chung

Y. B.

Gfroerer

M. C. A.

(2003). Career coaching: Practice, training, professional, and ethical issues. The Career Development Quarterly, 52, 141–152. doi:10.1002/j.2161-0045.2003.tb00634.x

Cicchetti

D. V.

(1994). Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological Assessment, 6, 284–290. doi:10.1037/1040-3590.6.4.284

10.

De Haan

(2008). I doubt therefore I coach: Critical moments in coaching practice. Consulting Psychology Journal: Practice and Research, 60, 91–105.

11.

De Haan

Culpin

Curd

(2009). Executive coaching in practice: What determines helpfulness for clients of coaching? Personnel Review, 40, 244.

12.

De Haan

Duckworth

Birch

Jones

(2013). Executive coaching outcome research: The contribution of common factors such as relationship, personality match, and self-efficacy. Consulting Psychology Journal: Practice and Research, 65, 40–57.

13.

De Roten

Zimmermann

Ortega

Despland

J. N.

(2013). Meta-analysis of the effects of MI training on clinicians’ behavior. Journal of Substance Abuse Treatment, 45, 155–162. doi:10.1016/j.jsat.2013.02.006

14.

Duckworth

De Haan

(2009). What clients say about our coaching. Training Journal, 64–67.

15.

Fleiss

J. L.

Shrout

P. E.

(1978). Approximate interval estimation for a certain intraclass correlation coefficient. Psychometrika, 43, 259–262. doi:10.1007/BF02293867

16.

Gati

Levin

(2014). Counseling for career decision-making difficulties: Measures and methods. The Career Development Quarterly, 62, 98–113. doi:10.1002/j.2161-0045.2014.00073.x

17.

Gessnitzer

Kauffeld

(2015). The working alliance in coaching: Why behavior is the key to success. The Journal of Applied Behavioral Science, 51, 177–197.

18.

Grant

A. M.

(2005). What is evidence-based executive, workplace, and life coaching? In Cavanagh

Grant

A. M.

Kemp

(Eds.), Evidence-based coaching: Theory, research and practice from the behavioral sciences (Vol. 1., p. 4., Chapter range: 1–12). Brisbane: Australian Academic Press.

19.

Grant

A. M.

(2007). Enhancing coaching skills and emotional intelligence through training. Industrial and Commercial Training, 39, 257–266.

20.

Grant

A. M.

Passmore

Cavanagh

M. J.

Parker

H. M.

(2010). The state of play in coaching today: A comprehensive review of the field. International Review of Industrial and Organizational Psychology, 25, 125–167. doi:10.1002/9780470661628.ch4

21.

Gregory

J. B.

Beck

J. W.

Carr

A. E.

(2011). Goals, feedback, and self-regulation: Control theory as a natural framework for executive coaching. Consulting Psychology Journal: Practice and Research, 63, 26–38.

22.

Gregory

J. B.

Levy

P. E.

(2011). It’s not me, it’s you: A multilevel examination of variables that impact employee coaching relationships. Consulting Psychology Journal: Practice and Research, 63, 67–88.

23.

Greif

(2007). Advances in research on coaching outcomes. International Coaching Psychology Review, 2, 222–249.

24.

Heppner

M. J.

Heppner

P. P.

(2003). Identifying process variables in career counseling: A research agenda. Journal of Vocational Behavior, 62, 429–452. doi:10.1016/S0001-8791(02)00053-2

25.

Hewes

D. E.

(1985). Systematic biases in coded social interaction data. Human Communication Research, 11, 554–574.

26.

Hill

C. E

. (1985). Manual for counselor verbal response modes category system (Rev. ed.) (Unpublished manuscript). University of Maryland, College Park.

27.

Hirschi

(2017). Career counseling: An overview and recent trends. In Kauffeld

Spurk

(Eds.), Handbook of career management and career planning. Heidelberg, Germany: Springer.

28.

Hoppe

(2013). What happens in consulting? A microanalytical investigation of the interaction between consultant and client (Doctoral dissertation). TU Braunschweig, Braunschweig, Germany.

29.

Ianiro

P. M.

Lehmann-Willenbrock

Kauffeld

(2014). Coaches and clients in action: A sequential analysis of interpersonal coach and client behavior. Journal of Business and Psychology, 30, 435–456.

30.

Ianiro

P. M.

Schermuly

C. C.

Kauffeld

(2013). Why interpersonal dominance and affiliation matter: An interaction analysis of the coach–client relationship. Coaching: An International Journal of Theory, Research & Practice, 6, 25–46. doi:10.1080/17521882.2012.740489

31.

Klonek

F. E.

Quera

Burba

Kauffeld

(2016). Group interactions and time: Using sequential analysis to study group dynamics in project meetings. Group Dynamics: Theory, Research, and Practice, 20, 209–222. doi:10.1037/gdn0000052

32.

Klonek

F. E.

Quera

Kauffeld

(2015). Coding interactions in motivational interviewing with computer-software: What are the advantages for process researchers? Computers in Human Behavior, 44, 284–292. doi:10.1016/j.chb.2014.10.034

33.

Klonek

F. E.

Wunderlich

Spurk

Kauffeld

(2016). Career counseling meets motivational interviewing: A sequential analysis of dynamic counselor–client interactions. Journal of Vocational Behavior, 94, 28–38. doi:10.1016/j.jvb.2016.01.008

34.

Korman

Bavelas

J. B.

De Jong

(2013). Microanalysis of formulations in solution-focused brief therapy, cognitive behavioral therapy, and motivational interviewing. Journal of Systemic Therapies, 32, 31–45.

35.

Landis

J. R.

Koch

G. G.

(1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159–174. doi:10.2307/2529310

36.

Markland

Ryan

R.M.

Tobin

V.J.

Rollnick

(2005). Motivational interviewing and self–determination theory. Journal of Social and Clinical Psychology, 24, 811–831. doi:10.1521/jscp.2005.24.6.811

37.

McKenna

D. D.

Davis

S. L.

(2009). Hidden in plain sight: The active ingredients of executive coaching. Industrial and Organizational Psychology, 2, 244–260. doi:10.1111/j.1754-9434.2009.01153.x

38.

Meinecke

A. L.

Klonek

F. E.

Kauffeld

(2016). Using observational research methods to study voice and silence in organizations. German Journal of Human Resource Management: Zeitschrift für Personalforschung, 30, 195–224. doi:10.1177/2397002216649862

39.

Miller

W. R.

Mount

K. A.

(2001). A small study of training in motivational interviewing: Does one workshop change clinician and client behavior? Behavioural and Cognitive Psychotherapy, 29, 457–471. doi:10.1017/S1352465801004064

40.

Miller

W. R.

Rollnick

(2013). Motivational interviewing: Helping people change (3rd ed.). New York, NY: Guilford Press.

41.

Miller

W. R.

Yahne

C. E.

Moyers

T. B.

Martinez

Pirritano

(2004). A randomized trial of methods to help clinicians learn motivational interviewing. Journal of Consulting and Clinical Psychology, 72, 1050–1062. doi:10.1037/0022-006X.72.6.1050

42.

Mittendorff

den Brok

Beijaard

(2010). Career conversations in vocational schools. British Journal of Guidance and Counselling, 38, 143–165. doi:10.1080/03069881003601007

43.

Moyers

T. B.

Martin

Manuel

J. K.

Hendrickson

S. M.

Miller

W. R.

(2005). Assessing competence in the use of motivational interviewing. Journal of Substance Abuse Treatment, 28, 19–26. doi:10.1016/j.jsat.2004.11.001

44.

Moyers

T. B.

Miller

W. R.

(2013). Is low therapist empathy toxic? Psychology of Addictive Behaviors, 27, 878–884. doi:10.1037/a0030274

45.

Moyers

Martin

Catley

Harris

K. J.

Ahluwalia

J. S.

(2003). Assessing the integrity of motivational interviewing interventions: Reliability of the motivational interviewing skills code. Behavioural and Cognitive Psychotherapy, 31, 177–184. doi:10.1017/S1352465803002054

46.

Moyers

T.B.

Martin

Manuel

J.K.

Miller

W.R.

Ernst

D.B.

(2010). Revised Global Scales: Motivational Interviewing Treatment Integrity 3.1.1 (MITI 3.1.1). Retrieved from http://casaa.unm.edu/download/MITI3_1.pd

47.

Multon

K. D.

Ellis-Kalton

C. A.

Heppner

M. J.

Gysbers

N. C.

(2003). The relationship between counselor verbal response modes and the working alliance in career counseling. The Career Development Quarterly, 51, 259–273. doi:10.1002/j.2161-0045.2003.tb00606.x

48.

O’Broin

Palmer

(2010). Exploring key aspects in the formation of coaching relationships: Initial indicators from the perspective of the coachee and the coach. Coaching: An International Journal of Theory, Research and Practice, 3, 124–143.

49.

Organisation for Economic Cooperation and Development. (2003). Career guidance and public policy: Bridging the gap. Retrieved January 10, 2004, from http://www.oecd.org

50.

Ostertag

P. A.

McNamara

J. R

. (1991). “Feminization” of psychology. Psychology of Women Quarterly, 15, 349–369. doi:10.1111/j.1471-6402.1991.tb00413.x

51.

Pantalon

M. V.

Chawarski

M. C.

Falcioni

Pakes

Schottenfeld

R. S.

(2004). Linking process and outcome in the community reinforcement approach for treating cocaine dependence: A preliminary report. American Journal of Drug and Alcohol Abuse, 30, 353–367. doi:10.1081/ADA-120037382

52.

Passmore

Fillery-Travis

(2011). A critical review of executive coaching research: A decade of progress and what’s to come. Coaching: An International Journal of Theory, Research and Practice, 4, 70–88.

53.

Patton

W. A.

McMahon

M. L.

(2006). The systems theory framework of career development and counseling: Connecting theory and practice. International Journal for the Advancement of Counselling, 28, 153–166.

54.

Rehfuss

M. C.

Corso

J. D.

Galvin

Wykes

(2011). Impact of the career style interview on individuals with career concerns. Journal of Career Assessment, 19, 405–419. doi:10.1177/1069072711409711

55.

Ritter

Bowden

Murray

Ross

Greeley

Pead

(2002). The influence of the therapeutic relationship in treatment for alcohol dependency. Drug and Alcohol Review, 21, 261–268. doi:10.1080/0959523021000002723

56.

Rochart

Rossier

(2016). Integrating motivational interviewing in career counseling: A case study. Journal of Vocational Behavior, 93, 150–162. doi:10.1016/j.jvb.2016.02.003

57.

Rogers

(1973). The characteristics of a helping relationship. In Bennis

W. G.

Berlew

D. E.

Schein

E. H.

Steele

F. I.

(Eds.), Interpersonal Dynamics, 3rd ed. (pp. 223–36). Homewood, IL: Dorsey.

58.

Shrout

P. E.

Fleiss

J. L.

(1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420–428.

59.

Sultanoff

S. M.

(2003). Integrating humor into psychotherapy. In Schaefer

C. E.

(Ed.), Play therapy with adults (pp. 107–143). Hoboken, NJ: Wiley. VanFleet, R.

60.

Tracey

T. J.

Kokotovic

A. M.

(1989). Factor structure of the working alliance inventory. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 1, 207–210. doi:10.1037/1040-3590.1.3.207

61.

Van Quaquebeke

Felps

(2016). Respectful inquiry: A motivational account of leading through asking questions and listening. Academy of Management Review, 43, 5–27. doi:10.5465/amr.2014.0537

62.

Whiston

S. C.

(2003). Career counseling: 90 years old yet still healthy and vital. The Career Development Quarterly, 52, 35–42. doi:10.1002/j.2161-0045.2003.tb00625.x

63.

Whiston

S. C.

Oliver

L. W.

(2005). Career counseling process and outcome. In Walsh

W. B.

Savickas

M. L.

(Eds.), Handbook of vocational psychology: Theory, research, and practice (pp. 155–194). Mahwah, NJ: Taylor & Francis.

64.

Whiston

S. C.

Rossier

Barón

P. M. H.

(2016). The working alliance in career counseling: A systematic overview. Journal of Career Assessment, 24, 591–604. doi:10.1177/1069072715615849