The effects of assessment feedback on rapport-building and self-enhancement processes. (Research).

An experiment was conducted to test the effects of assessment feedback on rapport and self-enhancement. After adult participants (N = 83) completed the Millon Index of Personality Stylers, the experimental group was given personalized assessment feedback; the control group received only general information about the inventory. After the session, all rapport-related scores (positive evaluations of examiner and session) and most of the self-enhancement-related scores (accurate mirroring, self-esteem, self-competence, and self-understanding) were significantly higher in the group that received assessment feedback. These results suggest that both processes are mechanisms by which the provision of assessment feedback produces positive change. Implications for mental health counselors are drawn.

**********

Assessments are used as starting points for several psychotherapeutic approaches, including cognitive-behavioral (e.g., Silverman & Kurtines, 1996), motivational (e.g., Miller & Rollnick, 1991), and personality-based approaches (e.g., Butcher & Keller, 1988). Traditionally, pre-treatment assessments have been regarded as having primarily diagnostic utility, conducted in order to gather accurate information about the client, conceptualize and diagnose the client's problems, and develop a treatment plan (Tallent, 1992). Recently, however, a shift in views of the purposes of assessments has occurred, with some scholars arguing that assessments can have significant therapeutic, in addition to diagnostic, utility when clients are provided with feedback about assessment results (Finn & Tonsager, 1992, 1997; Newman & Greenway, 1997).

Drawing upon clinical experience, many writers have noted various client benefits following a feedback session (for a review, see Finn, 1996). Indeed, current ethical guidelines state that clients should be given feedback about the assessments they have been given (e.g., Principle 4 D.7, American Mental Health Counseling Association, 2000). However, little research has been done to evaluate the impact of assessment or the provision of assessment feedback upon clients (Finn & Tonsager, 1997). As the popularity of treatment models using assessment and the provision of assessment feedback as initial motivating strategies have grown (Dunn, Deroo, & Rivara, 2001), so has the need for mental health professionals to understand the specific effects of assessment feedback, particularly as it impacts processes thought to increase clients' engagement in treatment and their motivation to change. The purpose of this study was to investigate two such processes: rapport-building and self-enhancement.

The few studies that have been conducted on the effects of feedback offer evidence of its potential. For example, one established brief therapeutic modality, Motivational Enhancement Therapy (MET: Miller & Rollnick, 1991; Miller, 1997), integrates the use of feedback with substance abuse clients in order to enhance motivation to change behavior. MET proponents provide feedback to clients on their Substance Abuse Subtle Screening Inventory (SASSI; Miller) profiles in order to increase awareness and reduce the defensive denial that often characterizes substance-abusing clients. Studies of the effects of SASSI feedback indicate that it positively affects both client motivation for treatment and treatment outcome (DiClemente, Bellino, & Neavins, 1999). However, studies such as these do not illuminate why or how feedback can have these effects.

In the first systematic investigation of a therapeutic model of assessment, Finn and Tonsager (1992) began isolating the mechanisms by which feedback has an impact on therapeutic outcomes. Specifically, they endeavored to investigate whether either assessment feedback itself or attention from the examiner had an impact by having half of an undergraduate sample recruited from a university-based outpatient clinic waitlist complete the MMPI-2 (Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989) and receive feedback about assessment results (attention + assessment/feedback). Control group participants received attention from the examiner equal in time but did not complete the MMPI-2 or receive feedback (attention, no assessment/feedback). Participants who received feedback on the assessment reported, across a 2-week period, significant decreases in symptomatology, significant increases in self-esteem, and greater hopefulness about the future than control group participants. However, because only the experimental group completed the MMPI-2, the impact of feedback was confounded in this study with the effects of taking (or not taking) the assessment.

Newman and Greenway (1997) replicated and extended this first study, investigating basic questions left unanswered regarding the potential beneficial effects of assessment feedback, their duration, and the sources of these benefits. Both a control group and an assessment group in a university counseling center completed the MMPI-2. Those in the experimental condition received feedback before completing measures of self-esteem and symptomatology, whereas those in the control condition completed the outcome measures without the benefit of feedback (but subsequently received it). Participants in the feedback condition reported significant increases in self-esteem and decreased symptomatology compared to the control group at a 2-week follow-up. This finding replicated Finn and Tonsager's (1992) results but experimentally controlled for the effects of completing the MMPI-2. Newman and Greenway also confirmed that the benefits to the assessment feedback group were specifically due to their receiving feedback, not to completing the MMPI-2 or receiving examiner attention. Despite this design innovation, what is responsible for the effects is still unknown, and Finn and Tonsager's question, "Which aspects of the feedback session were responsible for the changes?" (p. 279) remains unanswered.

In the clinical literature, two processes can be identified that likely underlie the impact of assessment feedback on various therapeutic outcomes. First, assessment feedback has been described as providing an opportunity for rapport building early in the therapist-client relationship. Consistent with this rapport-building hypothesis, some have suggested that sharing specific and relevant feedback reduces client defensiveness and instills client confidence in the helping process, thus contributing to a collaborative working relationship (e.g., Rozensky, Sweet, & Tovian, 1997). Others have also noted the relationship-building function of feedback, particularly when it is delivered in an accurate and empathic manner (Baucom & Epstein, 1990; Finn, 1996). Thus, it is reasonable to hypothesize that assessment feedback enhances and accelerates the rapport-building process.

Second, assessment: feedback has been described as an intervention that enhances self-related processes such as self-understanding, self-verification, positive self-regard, and self-awareness (e.g., Arkowitz, 1992; Finn, 1996). Consistent with the self-enhancement hypothesis, Escovar (1997) and Finn and Tonsager (1997) argued that concrete and accurate assessment information addresses basic human motivations of self-verification and insight, self-esteem, self-discovery, and self-efficacy. Many clinicians believe that the early enhancement of client self-awareness facilitates the collaborative identification of treatment goals and the motivation to work toward them (e.g., Baucom & Epstein, 1990; Millon, 1999). Thus, it is reasonable to hypothesize that assessment feedback enhances therapeutic outcomes through self-enhancement processes (e.g., accelerating client self-awareness and increasing self-efficacy and self-esteem). Alternatively, and most likely, feedback has both rapport-building and self-enhancing effects.

THE PRESENT STUDY

The primary aim of the present study was to extend what is known about the impact of assessment feedback by identifying processes that may promote therapeutic outcomes. To do so, we designed an experimental study in which participants were randomly assigned to an experimental or control condition. Participants in the experimental condition received feedback on specific assessment results, whereas participants in the control condition received no feedback about their assessment. The study included measures intended to evaluate the impact of condition (feedback vs. no feedback) on the two therapeutic processes that have been hypothesized to act as mechanisms of therapeutic change (i.e., rapport-building and self-enhancement).

Second, we sought to add to the empirical knowledge base on the therapeutic value of assessment by including a manipulation check assessing the impact of the experimental manipulation on participants' perceptions of examiner credibility. Some research has indicated that examinees' perception of examiner credibility can impact their reaction to feedback, with examinees being more likely to accept feedback from examiners to whom they attribute greater credibility (Furnham & Schofield, 1987). Therefore, we considered it important to the aims of the study to rule out the possibility that participants in the experimental and control groups perceived the examiner differently.

Finally, we sought to also answer Finn & Tonsager's call (1997) to extend the generalizability of previous research on the utility of feedback. Although the basic mechanisms hypothesized to underlie the assessment process could be reasoned to be relatively universal, previous research has primarily included White college students seeking treatment services, and empirical research on assessment effects in other populations has been scant (although important; Snyder & Clair, 1977). The degree to which findings about feedback effects generalize to diverse populations is virtually unknown. Using a primarily non-White, non-clinical student population in an experimental setting where standardized feedback was appropriate, we could begin to assess the basic mechanisms that might underlie feedback effects for a broader range of individuals, in a broader range of assessment circumstances.

To summarize, the study allowed us to test the following hypotheses. Hypothesis 1 addressed the impact of feedback on rapport-related processes. We expected that participants in the experimental (feedback) condition would report significantly better rapport with their examiner (a more positive relationship, less negative affect, and greater satisfaction with the assessment session) than participants in the control (no feedback) condition. Hypothesis 2 addressed the impact of feedback on self-enhancement processes. We expected that participants in the experimental (feedback) condition would report a significantly greater sense of self-verification (feeling understood), higher trait and state self-esteem and self-regard, greater self-efficacy, and new self-understanding than participants in the control (no feedback) condition.

METHOD

Participants

Participants for this study were a convenience sample of students (N = 83) from a large, urban, public university in southeast Florida. This university enrolls a primarily commuter student population, and students have few opportunities to interact with each other outside of class, diminishing the possibilities for contamination between experimental groups that could compromise the independent nature of the conditions. The sample consisted of 12 (15%) male and 71 (85%) female volunteers recruited from undergraduate psychology courses (ages 18-54, M = 22.6, SD = 6.71). The proportion of male and female students in this sample was similar to proportions of students reported in previous assessment feedback studies (Finn & Tonsager, 1992; Newman & Greenway, 1997). Ten participants (12%) identified themselves as non-Hispanic White, 8 (10%) as non-Hispanic Black, 57 (69%) as Hispanic, and 8 (10%) as Asian or Pacific Islander. Students received course credit for participation.

Measures

At the beginning of the assessment session, all participants completed a demographic information form and the stimulus assessment (Millon Index of Personality Styles MIPS; Millon, Weiss, Millon, & Davis, 1994). The post-assessment battery comprised the manipulation check measure, two rapport-related measures, two awareness-enhancement measures, and two therapeutic effects measures.

Stimulus. The Millon Index of Personality Styles (Millon, Weiss et al., 1994) was used as the stimulus object to test for the effects of assessment feedback. The MIPS is an established standardized test that assesses normal variants of the Millon Clinical Multiaxial Inventory (MCMI-III; Millon, Millon, & Davis, 1994) personality scales. This scale is designed for use with non-clinical individuals, and thus was deemed the most appropriate inventory for use in this study. It is a self-report questionnaire consisting of 180 true-false items. The MIPS has four scales; in this study, feedback was given only on the Interpersonal Behaviors scales. In order to provide specific feedback to a participant, detailed and therapeutically appropriate descriptions of each of the 10 Interpersonal Behavior personality prototypes were written (and are available from the first author; see also Appendix). The descriptions were based on interpretations by Choca and Van Denburg (1997), Millon (1999), Oldham and Morris (1990), Sperry (1995), and Weiss (1997). The length of the written material for each scale was standardized to be approximately 300 words (M = 296.5).

The Counselor Perception Measure (CPM). The CPM (McLennan, 1990), a measure of counselor credibility, served as a manipulation check. Experimental manipulations in previous studies (Finn & Tonsager, 1992; Newman & Greenway, 1997) controlled for the impact of both assessment and examiner attention. We included the CPM as an additional check on the impact of (a) examiner attention + specific feedback versus (b) examiner attention that included no specific feedback on participants' perceptions of examiner credibility. Participants were provided with 12 positive, descriptive adjectives such as expert, confident, and understanding and were asked to rate the examiner (rather than the counselor, as in the original CPM) on those adjectives along a 1 to 7 scale (from not at all to very much like the examiner). In the validation study for the instrument, factor analyses supported the construct validity of the CPM; it also had high internal consistency and demonstrated concurrent validity in field settings. In the present study, the reliability coefficient was [alpha] = .88.

The Assessment Questionnaire-2 (AQ-2). Two subscales (Positive Relationship, Negative Feelings) of the AQ-2 (Finn & Tonsager, 1992; 1997) were used to test the hypothesis that assessment feedback accelerates the rapport-building process. Participants are asked to rate the questions, on a Likert-type scale, from 1 (strongly disagree) to 5 (strongly agree). The 12-item Positive Relationship scale indicates participants' positive feelings between themselves and the examiner (e.g., "it was easy to trust the examiner"); the 9-item Negative Feelings scale indicates whether participants' felt hurt, judged, or exposed by the experience (e.g., "participation in this experiment was emotionally draining"). The wording of several items on the AQ-2 was changed slightly (e.g., the term therapist was changed to examiner). Another subscale (Positive Accurate Mirroring) of the AQ-2 was used to test the hypothesis that assessment feedback accelerates rapport building. The 11-item Positive Accurate Mirroring scale indicates whether participants' perceptions about themselves were reaffirmed and they felt proud, secure, or important after the testing experience (e.g., "this experiment captured the 'real' me"). The last subscale (New Self-Awareness/Understanding) of the AQ-2 was used to further understand how assessment feedback enhances a sense of self-enhancement, particularly a sense of self-discovery. The 13-item New Self-Awareness/Understanding scale reflects participants' view that they learned information about themselves that they regarded as potentially useful (e.g., "the examiner introduced me to new aspects of myself").

The subscales of the AQ-2 were derived through rational, factor-analytic, and item-analytic techniques. In validation studies of the scale, the AQ-2 subscales were shown to have good psychometric properties, including construct validity and internal consistency (Finn & Tonsager, 1992; 1997). In the present study, the Positive Relationship, Negative Feelings, Positive Accurate Mirroring, and New Self-Awareness/Understanding subscales yielded reliability coefficients of [alpha] = .92, [alpha] = .88 [alpha] = .91 and [alpha] = .94, respectively.

Self-esteem measures. The Rosenberg Self-Esteem Scale (RSE; Rosenberg, 1979) was used in this study (as in previous studies) to test the hypothesis that assessment feedback has a positive impact on global self-esteem. The RSE has received more psychometric analysis and empirical validation than any other self-esteem measure and is widely regarded as a reliable and valid measure of global self-worth (Robins, Hendin, & Trzesniewski, 2001). Participants rate 10 items (e.g., "I feel that I have a number of good qualities") on a 4-point Likert scale (from strongly agree to strongly disagree) as a measure of global self-esteem. Because a high score is a negative response (i.e., poor self-esteem) for this scale, items were reverse-scored before means were computed. In this study, the reliability coefficient for the RSE was [alpha] = .85.

The Self-Liking/Self Competence Scale (SLCS-R; Tafarodi & Swann, 1995, 2001) was used as an additional measure for assessing the impact of assessment feedback on self-enhancement, with the goal of pinpointing specific (in addition to global) self-esteem effects. Participants are asked to rate themselves using a Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). SLCS-R measures self-esteem as a two-dimensional construct, with subscales assessing Self-Liking (i.e., a sense of self-worth) and Self-Competence (i.e., a sense of personal efficacy). A typical item from the 8-item Self-Liking subscale is "I am very comfortable with myself." A typical item from the 8-item Self-Competence subscale is "I am highly effective at the things I do." In initial validation studies of the SLCS, structural analyses supported the conceptualization of Self-Liking and Self-Competence as distinct but related dimensions of global self esteem. Scale authors reported Cronbach alpha coefficients of .92 for the Self-Liking subscale and .89 for Self-Competence. In the present study, the reliability coefficient for Self-Liking and Self-Competence subscales were [alpha] = .88 and [alpha] = .79, respectively.

The State Self-Esteem Scale (SSES; Heatherton & Polivy, 1991) was used as a third measure for testing the hypothesis that assessment feedback enhances examinee's self-esteem. The SSES was included because it is a state (rather than trait) measure of self-esteem designed to assess variation in positive feelings about domains of the self that are sensitive to transient (and experimental) situational effects. The SSES of consists of 20 items measuring three esteem domains: Performance (e.g., "I feel confident about my abilities"), Social (e.g., "I feel concerned about the impression I am making"), and Appearance Self-Esteem (e.g., "I am pleased about my appearance right now"). In a series of validation studies, the scale demonstrated high internal consistency ([alpha] = .92), a factor structure that confirmed the three intended domains; concurrent and discriminant validity in laboratory, classroom, and clinical settings; and sensitivity to transient effects. In this study, the alpha reliability coefficients for the Performance, Social, and Appearance subscales were .73, .80, and .79, respectively.

Feedback Assessment Questionnaire (FAQ). The FAQ was specifically constructed for this study to assess participants' satisfaction with processes related to rapport building during the assessment session. Participants rate seven face-valid statements (e.g., "I feel that participating in this experiment was a positive and valuable experience for me as a person") on a 5-point Likert scale ranging from strongly disagree to strongly agree. The reliability coefficient for the FAQ in this study was [alpha] = .90. This questionnaire is available from the first author.

Procedure

When participants reported for their assessment session, they were randomly assigned to either the Experimental Condition (feedback) or the Control Condition (no feedback). After assignment, participants were asked to complete the demographic information form and the MIPS. Following completion of the MIPS, the participant received a 15-minute feedback session with the content (described below) appropriate to their assigned condition. After the 15-minute post-assessment session with the examiner, all participants completed the post-assessment battery.

Experimental Condition (Feedback). The Experimental (feedback) Condition was designed to test the effects of assessment feedback. During the post-assessment discussion session with the examiner, the participants received feedback on their specific assessment results on the MIPS Interpersonal Behavior Scales. Feedback was thus operationalized as the presentation of profile elevations, scale scores, and interpretations to participants. Each participant in the experimental condition received a profile of their MIPS scores with written feedback for the two scales with the highest elevations (for an example, see Appendix). In addition, these participants listened to a taped descriptive interpretation of their most and second most elevated scales. Auditory (i.e., tape recorded) and visual information (profile and text of tape recording) was provided simultaneously, primarily to ensure all participants' understanding of the feedback, and secondarily to simulate procedures often used in assessment feedback sessions where clients receive a profile of their results.

Control Condition (No Feedback). The Control Condition (no feedback) was designed to control for the effects of completing the assessment and subsequent examiner attention. Participants in the control condition, like those in the experimental condition, met individually with the examiner for a 15-minute post-assessment discussion session but they received no feedback on their specific assessment results. Instead, tape-recorded and written general descriptions of the MIPS as a testing instrument, equivalent in length to two scale descriptions (approximately 300 words) were provided. No personal profile information was provided.

For both conditions, participants' questions were handled using the same procedures. Specifically, the examiner addressed participants' questions about feedback in the experimental condition or questions about test content in the control condition by indicating the applicable written information and making minor explanatory elaborations. Thus, participants in each condition received examiner attention of comparable amount and form, but they differed in the type of information they received. Whereas the experimental group received information about their own assessment results, the control group received general information about the assessment itself but no specific feedback about their test results.

RESULTS

Preliminary Analyses

A Chi-square test was performed for both gender and ethnicity to analyze whether the two feedback groups differed significantly in their distribution on these variables. An alpha level of .05 was used for this and all subsequent analyses. No statistically significant differences emerged: therefore, the two groups were regarded as equivalent in these respects.

Correlation of the Instruments. As shown in Table 1, scales in the two domains of interest (i.e., rapport-related measures and self-enhancement measures) tended to be significantly correlated within, but not between, these respective domains. Additionally, scales measuring more than one dimension of a construct (the AQ-2, SSES, SLCS-R) had strongly correlated subscales, attesting to the scales' construct validity as used in this study. One exception was the Negative Assessment subscale (AQ-2), which was significantly negatively correlated with Counselor Perception (r = -.30, p < .01), Feedback Satisfaction (r = -.29, p < .01), and Self-Competence (r = -.22, p < .05), but unrelated to other measures used in the study. These patterns of correlations generally support the discriminant and convergent validity of the measures used in this study to observe the effects of feedback assessment.

Manipulation Check. A one-way analysis of variance (ANOVA) was conducted as the manipulation check, ascertaining whether participants in both conditions evaluated the examiner's credibility similarly. As shown in Table 2, the groups were not significantly different. Therefore, we concluded that the differences found in the subsequent analyses were attributable to the experimental variable (i.e., feedback) and were not confounded by a differential "source prestige" effect that could have been created by the differing types of information provided by the examiner (Furnham & Schofield, 1987).

Hypotheses

To test the hypotheses, two multivariate analyses of variance (MANOVAs) were conducted with Condition (experimental vs. control) as the independent variable and the measures that comprised the post-assessment battery as the dependent variables (see Table 2). MANOVA (using Wilks' criterion) was used to diminish the chance of Type 1 error because the scales used as dependent measures to assess both hypotheses tended to be significantly correlated.

The first MANOVA assessing the impact of feedback on rapport-related processes (i.e., the AQ-2 and FAQ scores) examined mean group differences in the Rapport-Related Measures. The multivariate F was significant using Wilks' criteria, F (4, 78) = 38.33, p < .001; estimated eta squared, [[eta].sup.2] = .66. Univariate tests indicated that participants in the feedback condition versus those in the control condition reported significantly higher scores for Positive Relationship with the Examiner, F (1, 81) = 11.06, p = .01; significantly lower scores for the Negative Feelings About the Assessment, F (1, 81) = 8.70, p = .004; and significantly higher scores on having a valuable and satisfying experience, F (1, 81) = 11.67, p < .001. As shown in Table 2, the effect sizes were 12%, 10%, and 63%, respectively.

The second MANOVA was conducted to test Hypothesis 2, that participants in the feedback condition would report significantly higher scores on self-enhancement measures than participants in the control no-feedback condition. Using Wilks' criteria, this MANOVA was also significant F (8, 74) = 9.20, p < .001; [[eta].sup.2] = .50. Univariate tests indicated that scale 1 of the AQ-2 (Positive Accurate Mirroring) was significant in the hypothesized direction F (1, 81) = 47.54, p < .001). Similarly, the New Self-Awareness/Understanding scale was also significant in the expected direction; F (1, 81) = 46.61,p < .001, indicating that those receiving feedback were more likely to feel that they learned potentially useful information about themselves. Those receiving feedback were much more likely to feel understood by the examiner. Participants in the experimental (feedback) condition reported significantly higher trait self-esteem than participants in the no-feedback condition on the Rosenberg scale F (1, 81) = 5.41,p = .05), the SLCS-R Self-Liking subscale F (1, 81) = 9.97, p = .01), and the SLCS-R Self-Competence subscale F (1, 81) = 5.19, p = .04. Means indicated that those receiving feedback reported higher self-regard at the conclusion of the feedback session than those who did not. Contrary to our expectations, univariate tests revealed no condition effects for the subscales of the State Self-Esteem Scale (Performance, Social, or Appearance self-esteem).

DISCUSSION

Our results supported the recent notion that in addition to having diagnostic utility, providing assessment feedback results may have therapeutic utility. Although previous research on assessment feedback has established its potential to impact particular outcomes (such as symptom reduction), very few studies have systematically attempted to identify the processes or mechanisms underlying the observed outcomes. In this study, we found that personalized feedback was associated with processes related to initial establishment of rapport and to the enhancement of intra-individual states such as self-awareness and self-regard. Thus, our results add to a growing body of literature indicating that receiving specific, personalized feedback following assessment can be beneficial in several respects (e.g., DiClemente et al., 1999).

In this study, the results pertaining to rapport-building as a hypothesized underlying process were clear and consistent. Participants receiving feedback following an assessment felt a stronger rapport with their examiner and had fewer negative feelings about the assessment session than participants receiving no feedback. As expected, they also rated the information they received following their assessment as significantly more valuable than participants receiving only general information.

The results pertaining to self-enhancement as a hypothesized underlying process were somewhat less consistent. Participants in the feedback condition did report a significantly greater sense of self-verification and self-efficacy/self-discovery than participants in the control condition. Consistent with previous research, they also scored higher on self-esteem and self-liking. The effects of feedback on the State Self-Esteem Scale, on the other hand, were not significant. We had included the scale to provide a comprehensive picture of feedback effects on self-esteem because it was sensitive to experimental manipulation in previous research. However, careful inspection of the content of the items of the three SSES subscales (Performance, Social, and Appearance self-esteem) suggested that the SSES represents aspects of extrinsically derived self-regard (versus intrinsically-derived feelings of self-liking and self-competence) that are less sensitive to psychological assessment feedback. Nevertheless, overall, our results indicated that receiving specific feedback enhances many aspects of the recipient's sense of self.

The present study also contributes to our knowledge about the source of the effects of assessment feedback. Because some research has indicated that examiner credibility may have an impact on clients' reactions, we included a manipulation check to assess for this effect. Previous studies of assessment feedback found (as did we) that the positive effects are not the result of taking the assessment itself and also not the result of examiner attention, but these studies did not rule out the possibility that the effects were confounded by a differential impact of the experimental manipulation upon examiner credibility. We specifically assessed this possibility and found that the level of feedback the examiner provided did not affect participants' perception of examiner credibility. This finding also supports our conclusion that the differences between the conditions obtained for the rapport-related and self-enhancement variables were also the result of the different levels of information provided. In other words, this study extended previous research in ruling out several possible causes of the positive impact associated with assessment feedback. Apparently, it is neither a result of the assessment itself nor a result of examiner attention, nor is it the result of examiner credibility associated with the type of feedback provided. The positive impact associated with assessment feedback, after ruling out these other possible causes, appears to be due to the informational value of the feedback itself.

In addition to the primary aim of the study--seeking to identify possible mechanisms of the positive change associated with assessment feedback--this study also provided the opportunity to address Finn and Tonsager's (1997) call to extend the generalizability of previous research. Our results suggest that the mechanisms operating in non-clinical populations who may seek assessments and feedback for non-clinical purposes are similar to the mechanisms thought to underlie therapeutic effects of feedback in clinical populations. Additionally, our findings with a primarily Hispanic sample are similar in type and direction to those of previous studies with primarily White samples and so further support the generalizability of the therapeutic model of assessment offered by Finn and Tonsager. Thus, there is accruing evidence that further development of a general theoretical model of assessment processes and their impact on therapeutic (or positive) outcomes is merited.

Limitations and Future Directions

This study's results should be considered in the context of several limitations of the methodology. Generalizing the findings to other populations should be done cautiously, because the participants who led us to our conclusions were a relatively young, mostly female, and largely Hispanic non-client student population. In addition, we analyzed the relationship of several variables within a relatively small sample size. Future assessment feedback research, with both client and non-client populations, should include a wider range of adults (with respect to age, racial and ethnic diversity, and cognitive ability) more representative of those seeking services in settings where assessment feedback may be sought and given. Then possible individual and group differences in feedback effects can be assessed. For example, we did not, due to lack of statistical power, analyze specific MIPS personality profiles as they related to our rapport and self-enhancement variables. In this study, non-clinical participants received feedback on normal variants of personality profiles. Previous research has not found participants' severity and type of psychopathology to be related to the improvements found in client symptomatology and self-esteem following feedback (Finn & Tonsager, 1992: Newman & Greenway, 1997). However, it is possible and theoretically predictable that participants with various personality constellations might respond differently to assessment feedback in ways that would be reflected in the rapport or self-enhancement variables. It remains for future research to explore this possibility.

Because we assessed self-esteem variables only in the post-assessment battery, it is possible that the two feedback groups may have differed in their initial levels of self-liking, self-competence, and self-esteem. Given the other robust differences we found that are more clearly attributable to the different feedback conditions, we think this unlikely, but taking steps to ensure the equivalence of experimental groups would strengthen future studies.

Finally, although this study and previous studies found preliminary empirical support for the therapeutic model of assessment, additional studies are needed to further clarify both the effects of feedback and the mechanisms by which these effects operate. For example, it is interesting to note that in our study, the two largest differences between the experimental and control groups were on the AQ-2 subscales of Positive Accurate Mirroring and New Self-Awareness/Understanding. It may be that these two mechanisms are causally linked (i.e., new self-awareness and self-understanding are enhanced as a result of positive accurate mirroring), although this relationship could not be assessed in this study. Additionally, while specific, personalized feedback was expected to result in greater self-awareness than general feedback, the question of why receiving feedback, which included both more and less favorable aspects for each personality type (see Appendix), resulted in higher self-regard and self-competence is intriguing and deserves further investigation. Future studies should illuminate relations between assessment feedback, which evidently may be perceived by clients as accurate positive mirroring, and other process and outcome variables.

Implications for Mental Health Counselors

An implication of our study is that when practitioners are faced with a choice of whether or not to conduct standardized assessments at the outset of the treatment, they should not overlook the possible benefits of assessment feedback to clients. Our results, which are similar to those of other studies done with client populations and using other assessments, illustrate that administering assessments and subsequently providing informative, personalized feedback can enhance treatment. Specifically, providing individuals with information about themselves can have a positive effect on their rapport with the examiner and their sense of self-regard, self-competence, self-understanding, and positive evaluations of the experience. Positive feelings toward the assessment experience and the examiner can help to build an alliance and foster clients' engagement in treatment. Because assessment feedback can result in positive feelings about the self, it may also allow individuals who are initially reluctant consumers of supportive services to become more positively disposed toward therapeutic intervention.

Additionally, mental health counselors should consider the benefits of the informational aspects of feedback for clients. Integrating our findings with previous literature, we believe that when mental health counselors provide assessment information to clients, a focus is created for both; and this sharing of personalized information assists in establishing client expectations for active, collaborative participation. Thus, the administration of pertinent standardized assessments and the provision of feedback to clients are powerful procedures that mental health counselors could justifiably and routinely use, given the current pressures in the profession to accomplish much within a limited number of sessions.

However, as noted previously, our findings as well as those in previous studies using client populations were obtained with a stable population that was relatively young, presumably well-adjusted, and cognitively capable. Community client populations also include individuals who are unstable or who have compromised information-processing capacities. Ultimately the mental health counselor must decide on a case-by-case basis the extent to which assessment feedback will be of value and when the risks incurred might outweigh the potential benefits. Using assessment information primarily to guide diagnostic and intervention decisions while giving clients only general information about the purpose and results of their assessments may sometimes be more appropriate, especially when they are likely to find specific individualized feedback overwhelming or upsetting, thus impeding their therapeutic progress. Nevertheless, in making this decision, mental health counselors should also consider the effect upon individuals of taking an assessment and not receiving subsequent specific assessment feedback. Our study indicated that those who do not receive personalized feedback following assessment have more negative feelings about the assessment experience in general and a diminished rapport with the examiner in particular. Lack of exchange of information in the client-assessor relationship can be detrimental to the goal of enhancing clients' motivation to implement positive behavior change (Rollnick, Mason, & Butler, 1999). Assessments must be understood as clinical tools that require appropriate follow-up, or deleterious effects may result.

In conclusion, knowledge about the impact of assessment feedback should be incorporated into our understanding of routine assessment practices. In addition to its diagnostic utility, assessment feedback is also useful for establishing initial rapport and self-awareness that can be the springboard for subsequent client engagement. As others have suggested, the specific information provided and the new self-understanding that is fostered can be used to facilitate collaborative goal setting and enhance client motivation for change (Miller & Rollnick, 1991). Because assessment feedback can result in positive feelings about the self, it may allow individuals who are initially resistant to therapeutic interventions to become more positively disposed to the experience, and may foster positive outcome expectancies for therapy. Feedback can also boost the sense of self-efficacy that clients bring into the collaborative relationship. Thus, administering an objective assessment to clients at the outset, and subsequently providing them with specific assessment feedback, can help get psychotherapy off to a good start.

APPENDIX

The Retiring Personality Style

Persons scoring high on this scale usually feel most comfortable when they are alone. They tend to feel at ease when they are by themselves and do not necessarily seek companionship all the time.

When confronted with a problematic situation or difficult decision, persons scoring high on this scale are likely to 'use their heads.' They do not tend to become too emotional but rather remain rational and well composed. In relationships with others they tend to take the passive role without feeling the need to engage in social events frequently. They are satisfied focusing their attention on practical things. They are likely to choose solitary work or studious engagement over socializing events. A monotonous work schedule or scientifically oriented curriculum does not repel; rather they choose it over something more people-oriented. They appreciate most to work with objects. Interpersonal contact at the workplace is secondary to them and is not necessary for them to be satisfied with their job. They are more interested in concrete things and shy away from exploring emotional and metaphysical things.

Persons scoring high on this scale do not tend to overreact in an emotionally charged situation. They do not let minor disappointments ruin their day. Moreover, they remain calm if things don't work out right. Even when faced with an exciting event, they tend to not become too emotionally aroused.

People scoring high on this scale sometimes have difficulties forming close friendships or relationships. They usually handle things on their own and are reluctant to seek out support. They are more inclined to tackle the problems by themselves and are likely to come up with a strategic and rational problem-solving approach.

These people are relatively private. They are likely to accept things for what they are and do not necessarily feel that they have to constantly 'change the world.' Their mood changes very infrequently. They do not tend to have emotional outbursts. While they are capable of having stable and lasting relationships, they tend to avoid overly dramatic ones. Due to their focus on the concrete, they have some difficulty reading 'between the lines.' Others might at times perceive such persons as uninvolved and indifferent. They usually do not feel compelled to analyze someone else's state of mind.

Table 1. Correlations of Examiner, Assessment Satisfaction,
and Self-Regard Measures

                        Accurate    Self-      Pos.     Neg.
                        mirroring   under-     Eval.    Eval.
                                    standing

Counselor Perception    .22 *       .16        .53 **   -.30 **
  Accurate Mirroring    --          .83 **     .65 **   -.06
  Self-understanding                --         .50 **   -.04
  Positive Evaluation                          --       -.17
  Negative Evaluation                                   --
Feedback Satisfaction
Global Self-Esteem
  Performance
  Social
  Appearance
  Self-Liking

                        Feed-        Self     Perfor-   Social
                        back         Esteem   mance
                        Sattisfac.

Counselor Perception     .24 *        .29 **   .07      -.05
  Accurate Mirroring     .78 **       .19     -.06       .05
  Self-understanding     .78 **       .03     -.13      -.07
  Positive Evaluation    .53 **       .24 *    .04       .07
  Negative Evaluation   -.29 **      -.19     -.01      -.01
Feedback Satisfaction   --            .22 *    .01       .06
Global Self-Esteem                   --        .56 **    .51 **
  Performance                                 --         .64 **
  Social                                                --
  Appearance
  Self-Liking

                        Appear-   Self      Self
                        ance      Liking    Compe-
                                            tence

Counselor Perception     .19       .25 *     .36 **
  Accurate Mirroring     .18       .14       .21
  Self-understanding     .04       .03       .05
  Positive Evaluation    .12       .19       .31 **
  Negative Evaluation   -.14      -.20      -.22 *
Feedback Satisfaction    .26       .27 *     .23 *
Global Self-Esteem       .48 **    .83 **    .58 **
  Performance            .45 **    .63 **    .58 **
  Social                 .44 **    .49 **    .36 **
  Appearance            --         .53 **    .46 **
  Self-Liking                     --         .57 **

Note: * p<.05, ** p<.01. Inter-scale correlations are shown in bold.

Table 2. Group Differences on Examiner, Rapport-Related,
and Self-Related Measures

                               Experimental     Control
Measure                         M      SD       M      SD
Examiner Perception Measure
Counselor Perception           5.82    .94     5.74    .89

Rapport-Related Measures
Positive evaluation (AQ-2)     3.63    .74     3.14    .62
Negative evaluation (AQ-2)     1.30    .41     1.69    .74
Feedback satisfaction (FAQ)    4.21    .44     2.79    .64

Self-Enhancement Measures
Self-Verification
  Accurate mirroring (AQ-2)    3.84    .59     2.78    .78
Self-Esteem
Rosenberg Self Esteem          3.53    .41     3.33    .39
Self-Liking (SLCS-R)          35.33   4.79    31.51   6.08
State Self Esteem (SSES)
  Performance                  4.40    .44     4.22    .54
  Social                       4.20    .67     3.92    .73
  Appearance                   3.99    .66     3.72    .68
Self-Efficacy/Self-Discovery Measures
Self-Understanding (AQ-2)      3.77    .54     2.67    .88
Self-Competence (SLCS-R)      30.63   4.96    28.30   5.19

                                      Univariate
Measure                           F        [[eta].sup.2]
Examiner Perception Measure
Counselor Perception             .19             .002

Rapport-Related Measures
Positive evaluation (AQ-2)     11.06 **          .12
Negative evaluation (AQ-2)      8.70 **          .10
Feedback satisfaction (FAQ)   136.07 ***         .63

Self-Enhancement Measures
Self-Verification
  Accurate mirroring (AQ-2)    47.54 ***         .37
Self-Esteem
Rosenberg Self Esteem           5.41 *           .06
Self-Liking (SLCS-R)            9.97 **          .11
State Self Esteem (SSES)
  Performance                   2.85
  Social                        3.32
  Appearance                    3.35
Self-Efficacy/Self-Discovery Measures
Self-Understanding (AQ-2)      46.61 ***         .37
Self-Competence (SLCS-R)        4.32 *           .05

* p < .05. ** p < .01. *** p < .001.

REFERENCES

American Mental Health Counseling Association. (2000). Code of Ethics of the American Mental Health Counselors Association. Alexandria, VA: Author.

Arkowitz, H. (1992). Integrative theories of therapy. In D. Freedheim (Ed.), History of psychotherapy (pp. 261-303). Washington, DC: American Psychological Association.

Baucom, D., & Epstein, N. (1990). Cognitive-behavioral marital therapy. New York: Brunner/Mazel.

Butcher, J. N., & Keller, L. S. (1988). Objective personality assessment. In G. Goldstein & M. Hersen (Eds.), Handbook of psychological assessment (pp. 213-247). New York: Pergamon.

Butcher, J. N., Dahlstrom, W. G., Graham, J. R., Tellegen, A., & Kaemmer, B. (1989). Minnesota Multiphasic Personality Inventory-2 (MMPI-2): Manual for administration and scoring. Minneapolis, MN: University of Minnesota Press.

Choca, J., & Van Denburg, E. (1997). Interpretive guide to the Millon Clinical Multiaxial Inventory (2nd ed.). Washington, DC: American Psychological Association.

DiClemente, C., Bellino, L., & Neavins, T (1999). Motivation for change and alcoholism treatment. Alcohol Research and Health, 23(2), 86-92.

Dunn, C., Deroo, L., & Rivara, F. P. (2001). The use of brief interventions adapted from motivational interviewing across behavioral domains. Addiction, 96, 1725-1742.

Escovar, L. (1997). The Millon inventories: Sociocultural considerations. In T. Millon (Ed.), The Millon inventories: Clinical and personality assessment (pp. 264-285). New York: Guilford.

Finn, S. (1996). Using the MMPI-2 as a therapeutic intervention. Minneapolis, MN: University of Minnesota Press.

Finn, S. E.. & Tonsager, M. E. (1992). Therapeutic effects of providing MMPI-2 test feedback to college students awaiting therapy. Psychological Assessment, 4, 278-287.

Finn, S., & Tonsager, M. (1997). Information-gathering and therapeutic models of assessment: Complementary paradigms. Psychological Assessment, 9, 374-385.

Furnham, A., & Schofield, S. (1987). Accepting personality test feedback: A review of the Barnum effect. Current Psychological Research and Reviews, 6, 162-178.

Heatherton, T. F., & Polivy, J. (1991). Development and validation of a scale for measuring State self-esteem. Journal of Personality and Social Psychology, 60, 895-910.

McLennan, J. (1990). Clients' perceptions of counsellors: A brief measure for use in counselling research, evaluation, and training. Australian Psychologist, 25, 133-146.

Miller, G. (1997). The Substance Abuse Subtle Screening Inventory (SASSI). Bloomington, IN: SASSI Institute.

Miller, W. R., & Rollnick, S. (1991). Motivational interviewing: Preparing people to change addictive behavior. New York: Guilford Press.

Millon, T. (1999). Personality-guided therapy. New York: Wiley.

Millon, T., Millon, C., & Davis, R. D. (1994). Millon Clinical Multiaxial Inventory Manual III. Minneapolis: National Computer Systems.

Millon, T.. Weiss, L., Millon, C., & Davis, R. D. (1994). MIPS: Millon Index of Personality Styles Manual. San Antonio: Psychological Corp.

Newman, M., & Greenway, P. (1997). Therapeutic effects of providing MMPI-2 test feedback to clients at a university counseling service: A collaborative approach. Psychological Assessment, 9, 122-131.

Oldham, J., & Morris, L. (1990). Personality self-portrait: Why you think, work, love, and act the way you do. New York, NY: Bantam.

Robins, R. W., Hendin, H. M., Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin, 27, 151-161.

Rollnick, S., Mason, P., & Butler, C. (1999). Behavior health change: A guide for practitioners. New York: Churchill Livingstone.

Rosenberg, M. (1979). Conceiving the self. New York, NY: Basic Books.

Rozensky, R., Sweet, J., & Tovian, S. M. (1997). Psychological assessment in medical settings. New York: Plenum Press.

Silverman, W. K., & Kurtines, W. M. (1996). Anxiety and phobic disorder. New York: Plenum Press.

Snyder, C. R., & Clair, M. S. (1977). Does insecurity breed acceptance? Effects of trait and situational insecurity on acceptance of positive and negative diagnostic feedback. Journal of Consulting and Clinical Psychology, 45, 843-850.

Sperry, L. (1995). Handbook of diagnosis and treatment of the DSM-IV personality disorder. New York, NY: Brunner/Mazel.

Tafarodi, R. W., & Swann, W. B. (1995). Self-liking and self-competence as dimensions of global self-esteem: Initial validation of a measure. Journal of Personality Assessment, 65, 322-342.

Tafarodi, R. W., & Swann, W. B. (2001). Two-dimensional self-esteem: Theory and measurement. Personality and Individual Differences, 31, 653-673.

Tallent, N. (1992). The practice of psychological assessment. Englewood Cliffs, NJ: Prentice Hall.

Weiss, L. (1997). The MIPS: Gauging the dimensions of normality. In T. Millon (Ed.), The Millon Inventories (pp. 498-525). New York: The Guilford Press

Andrea Allen is a doctoral student and a therapist in the child Anxiety and Phobia Program at Florida International University. Marilyn Montgomery, Ph.D., is an assistant professor; Jonathan Tubman, Ph.D., Leslie Frazier, Ph.D., and Luis Escovar, Ph.D., are associate professors; all are in the Department of Psychology and the Mental Health Research and Services Program at Florida International University, Miami. E-mail: [e-mail].