Working with Black vs. White patients: an experimental test of therapist decision-making in cognitive behavioral therapy for depression

ABSTRACT

Whether and how therapists’ delivery of cognitive behavioral therapy (CBT) for depression differs by patients’ ethnicity or race remains unclear. In this study, 218 therapists were randomized to clinical vignettes that involved the same text but varied in whether the accompanying image depicted a Black or White patient. Therapists exhibited three key differences in their views of clinical strategies for working with Black as compared to White patients. They viewed cognitive change strategies as less therapeutic and validation strategies as more therapeutic for Black patients. They reported similar differences for the time they would spend on each kind of strategy. When asked to compare the relative importance of cognitive change vs. validation strategies specifically, therapists rated validation as more important for Black than White patients. Among therapists presented with Black patients, positive racial attitudes were associated with viewing cognitive change and validation strategies as more therapeutic. These results suggest therapists tend to believe it is desirable to incorporate cognitive methods more limitedly when working with Black patients. Whether such adaptations enhance or detract from the care of Black patients is an important issue that merits future investigation.

KEYWORDS:

• Race
• cognitive behavioral therapy
• depression
• psychotherapy process

Several experts have advocated for modifications of psychotherapies with a variety of specific cultural adaptations (Hays, 2009; Huey et al., 2014), and manuals themselves often call for therapists to make various efforts to personalize these treatments. Yet, little is known about the extent and nature of any adaptations therapists are making for specific groups when delivering psychotherapy in clinical practice. Both the idea of a need for therapists to provide care with cultural competence and concerns about therapists’ biases suggest it is important to understand how therapists tend to modify their approach in session when working with ethnic and racial minority patients. In this study, we focus on cognitive behavioral therapy (CBT), a well-established treatment for depression (Hollon et al., 2006; Strunk et al., 2017). To understand therapists’ decisions in providing this treatment, we experimentally investigate their views of how to work with a series of Black or White patients for whom all other features are presented identically.

In CBT, a primary goal is to help patients identify and correct unrealistic evaluations that serve to perpetuate their emotional distress (Beck et al., 1979). Therapists use cognitive methods (i.e. techniques that help patients identify, evaluate, and correct their overly negative views) to bring about cognitive change, which in turn appears to bring about decreases in depressive symptoms (Schmidt et al., 2019). In a recent effort from our group, we drew data from three prior studies of CBT for depression (combined n = 229 patients) to examine naturalistic differences in therapists’ delivery of treatment when working with White and Black patients (Ezawa & Strunk, 2021). Using observer ratings of process variables measured during the early sessions of treatment, we found therapists working with Black patients utilized cognitive methods less extensively than when they worked with White patients. These findings provide initial evidence that differences in CBT treatment process occur when therapists provide care to Black versus White patients. However, why Black patients’ sessions tended to be characterized by different therapeutic techniques was not clear. Some psychologists have suggested that culturally competent therapists working with minority patients would be well-advised to avoid questioning the validity of patients’ thoughts and beliefs, and instead validate their patients’ self-reported experiences (Hays, 2009). Were therapists using other techniques, such as validation (i.e. affirming patients’ experiences are understandable and make sense), in place of cognitive methods? Or perhaps, were there some clinical characteristics other than race that might account for these differences? Furthermore, might these differences be related to therapists’ own racial attitudes? Studies that directly address these questions with the relevant focal comparisons are needed.

Vignette studies are an important method for making such focal comparisons. Clinical case vignettes allow researchers to experimentally manipulate one specific variable that cannot be experimentally manipulated in clinical trials or naturalistic psychotherapy research. In doing so, researchers can be confident that any observed changes in the outcome of interest is attributable to change in the manipulated variable. Experimental tests of race using clinical case vignettes have similarities with other experimental study designs (e.g. resume audit studies; Darolia et al., 2016) that have also been used as a tool for assessing quality of care in medical settings (e.g. Peabody et al., 2000, 2004). For example, Sabin et al. (2008) conducted a study on physician attitudes about race and its effect on quality of medical care. In their study, they tested four medical case vignettes (each with two race conditions) and observed differences in treatment recommendations made by the participating physicians. They found that physicians were more likely to recommend home care rather than hospital care for Black patients and more likely to recommend additional days of antibiotic treatment and inpatient care for White patients. The use of clinical case vignettes to investigate how patient race and ethnicity may impact therapists’ decision-making has largely yet to be examined.

Although our study is focused on experimentally investigating therapists’ clinical decision-making in the context of CBT in working with patients from different racial backgrounds, it is important to note that meta-analytic investigations have examined potential ethnic or racial differences in therapeutic outcomes. A recent meta-analysis suggests that psychotherapy outcomes do not differ greatly as a function of patients’ racial or ethnic minority status (Ünlü Ince et al., 2014). Although such comparisons are important, it is the treatments being studied in clinical trials, not ethnic or racial groups, that are compared experimentally. Thus, the comparisons may be confounded by differences among patients apart from race or ethnicity. Moreover, these comparisons of outcome do not assess how therapies are being delivered. Suboptimal processes are possible even when outcomes are comparable. Additionally, given mistrust of researchers, some individuals from ethnic and racial minority groups may be less likely to participate in such trials. Thus, comparisons of outcomes in clinical trials may be limited in their ability to adequately recruit diverse samples and, even if they were not, psychotherapeutic processes merit investigation independent of therapeutic outcomes.

Purpose of this study

To address these issues, we examined therapists’ judgments of the appropriateness and their own intention to use different clinical strategies in CBT for depression when working with patients of differing racial backgrounds. To examine this, we experimentally manipulated patient race (White and Black) in clinical case vignettes describing patients with major depressive disorder. We randomized therapists to be presented with clinical case vignettes of all Black or all White patients. Based on our recent finding that therapists naturalistically utilized cognitive methods less with Black patients as compared to White patients in CBT for depression (Ezawa & Strunk, 2021), we hypothesized that therapists would rate the use of such cognitive change strategies as less important for Black patients. In addition, we expected that therapists would tend to view the most appropriate clinical strategy when working with Black patients as consisting of an approach that emphasizes understanding and validation with more limited use of the core change strategies of CBT. Additionally, we examined the relationship of therapists’ racial attitudes towards Black individuals with their clinical decision-making among those randomized to consider CBT in working with Black patients. We also explored whether the racial salience of the vignettes (i.e. the extent to which vignettes included circumstances in which discrimination may seem more likely) moderated the impact of patient race on therapists’ clinical decision making.

Methods

Participants

A total of 218 therapists (109 randomized to the White patient condition, 109 randomized to the Black patient condition) completed this study and were included in subsequent analyses.¹ Please see Figure 1 for a flow chart of study participants. The average age was 39.24 (SD = 12.45) years. Regarding ethnicity, 9 (4.1%) therapists identified as Hispanic. Regarding race, 184 (84.4%) therapists identified as White, 18 (8.3%) as Asian, 8 (3.7%) as Black, 2 (0.9%) as American Indian or Alaskan Native, and 1 (0%) as Native Hawaiian or other Pacific Islander. Five (2.3%) identified as “other”. Ninety-seven (44.5%) therapists in the study indicated having a Ph.D., 26 (11.9%) indicated having a Psy.D., 85 (39.0%) indicated having a master’s degree, 8 (3.7%) indicated having a bachelor’s degree, and 2 (0.9%) described their highest degree as “other.” They reported the portion of their clinical work that was based on CBT, with an average of 70.53% (SD = 25.76).

Figure 1. Flow chart of study.

Measures

Clinical case vignettes

Each participating therapist was presented with four clinical case vignettes (see the Online Supplement). Each clinical case vignette consisted of information describing the background and presenting problem of a hypothetical depressed patient, an image of the patient, and a series of questions intended to capture the participating therapist’s judgment on the appropriateness of different clinical strategies with the patient. Each participating therapist was randomized to one of two conditions: (1) vignettes with images of Black patients, or (2) vignettes with images of White patients. The clinical case vignettes were inspired by vignettes provided in a study examining registered nurse attitudes and their clinical decision-making in working with patients of varying ethnic/racial backgrounds (Haider et al., 2015). Specifically, the first vignette described a situation in which a female patient was experiencing difficulty in her romantic relationship. The second vignette described a situation in which a male patient was applying for a job. The third vignette described a situation in which a female patient had been followed around a store by a security guard. The fourth vignette described a situation in which a male patient had been pulled over by a police officer. The latter two vignettes were designed to be more racially salient. Each participating therapist was presented with identical surveys (including the same four vignettes, in the same order, and followed by the same set of questions), except for pictures displayed with each vignette.

Image selection for case vignettes

We selected two images for each vignette that were matched on gender, age, and ratings of attractiveness, but varied in race (i.e. Black or White). Images were provided through the University of Chicago’s Chicago Face Database (Ma et al., 2015), which provided high-resolution photographs of the faces of adults with different ethnic/racial backgrounds. The Chicago Face Database also provided norming data for each of the photos, including objective information on physical attributes and subjective ratings such as attractiveness made by independent raters. For the images included in this study, we first randomly selected an image of one Black individual for each clinical case vignette (images of two Black women and two Black men). Second, we selected images of White individuals from the Chicago Face Database who were the closest match in terms of gender, age, and rated attractiveness.

Questions following vignettes

Each vignette was followed by 11 zero- to six-point Likert scale items assessing the appropriateness of cognitive change and validation strategies to be used with the patient described in the vignette. Items were created by the authors of the study who have extensive experience with CBT, in consultation with other CBT researchers. Individual item scores were averaged across the four case vignettes and across the item sets described below.

The first set of items, which will be referred to as “cognitive change-therapeutic,” included two items assessing the extent to which cognitive change strategies were deemed as therapeutic to use with the patient described (higher scores indicated greater belief that these strategies were therapeutic). As an example, one item asked: “To what extent do you think it would be therapeutic to emphasize cognitive change strategies with this patient during this session?” Responses for this item could range from zero (“not at all therapeutic”) to six (“completely therapeutic”). The second set of items, which will be referred to as “validation-therapeutic,” included two items assessing the extent to which validation techniques were deemed as therapeutic (higher scores indicated greater belief that these strategies were therapeutic). The third set of items, which will be referred to as “cognitive change-time,” included three items reflecting the amount of time therapists would spend in the session utilizing cognitive change strategies (higher scores indicated greater amounts of time). The fourth set of items, which will be referred to as “validation-time,” included three items assessing the amount of time therapists would spend in the session utilizing validation strategies (higher scores indicated greater amounts of time). In addition, there was a single item asking therapists to indicate the relative importance of using cognitive change vs. validation strategies. This variable will be referred to as “cognitive change vs. validation.” Exploratory factor analyses supported the item groupings described above. For details on the analyses that informed scoring, please see the Online Supplement. Items are available from the authors of the study upon request.

Racial attitudes

The Symbolic Racism 2000 Scale (Henry & Sears, 2002) is an eight-item scale that assesses contemporary racial attitudes towards Black individuals, without referencing government policies. A total score is used, with higher scores indicating more positive racial attitudes. This scale has demonstrated acceptable internal consistency and predictive validity in relation to White American attitudes toward racial policies (Henry & Sears, 2002).

Social desirability

The Marlowe-Crowne Social Desirability Scale-Short Form (Reynolds, 1982) is a 13-item version of the classic Marlowe-Crowne Social Desirability Scale (Crowne & Marlowe, 1960) that measures the tendency to respond in a socially desirable manner using true or false statements. A total score is used, with higher scores indicating greater tendencies to respond in a socially desirable manner. The short form has shown evidence of greater reliability and validity than the original version (Loo & Thorpe, 2000).

Procedures

Mental health therapists across the United States and Canada were recruited to participate in a study of clinical decision-making in CBT using word-of-mouth and online communication. Interested therapists were provided a link to the online experiment. Once therapists followed the link, they were provided a consent form for the study. After providing consent, therapists were randomized to one of two conditions using the Qualtrics randomization feature with conditions evenly presented. Conditions were identical except for patient images shown with case vignettes. Across both conditions, therapists were asked to complete a demographics and CBT experiences questionnaire. Therapists were then directed to read a series of clinical case vignettes followed by questions assessing the appropriateness of therapy techniques to use with the patient described in each vignette. After the questions about the vignettes, therapists were asked to complete racial attitudes and social desirability scales. At two points in the survey (i.e. after completing the fourth vignette and after completing the racial attitudes scale), therapists were required to pass attention check questions (two basic math problems). The study involved deception in that participants were not told about the randomization or that race was a variable of interest. Accordingly, participants were given the option to have their data withdrawn following the debriefing. See Figure 2 for a figure depicting stages of the experiment. This study was approved by an institutional review board at The Ohio State University.

Figure 2. Stages of study.

Analytic strategy

Primary analyses

We conducted a series of general linear models using SAS 9.4 proc glm to examine the effect of condition on key study variables while controlling for social desirability. This involved a total of five tests, one for each of the following study variables: cognitive change-therapeutic, cognitive change-time, validation-therapeutic, validation-time, and cognitive change vs. validation.

We also explored whether racial attitudes were related to therapists’ responses regarding appropriateness of different clinical strategies when they were presented with images of Black patients. In the subset of the sample who were presented with Black patients, we planned to evaluate a regression model using SAS proc reg in which each of the key study variables (i.e. cognitive change-therapeutic, cognitive change-time, validation-therapeutic, validation-time, cognitive change vs. therapeutic) were entered as predictors of racial attitudes, while controlling for social desirability.

Exploratory analyses

Of the four vignettes we used, the latter two involved circumstances that involved law enforcement and were intended to be viewed as being more likely to involve discrimination (when the patient was Black). We first created a binary variable that indicated if the vignette was racially salient (“1”) or not (“0”). The first two vignettes were coded as “0”, whereas the latter two vignettes were coded as “1”. Then, we tested whether the racial salience of the vignettes would moderate the effect of patient race on key study variables. To examine this possibility, we planned to use repeated measures regression models using SAS proc mixed without specification of random effects to test the interaction between the race of the patient and racial salience of the vignette in predicting each of the key dependent variables. We included social desirability scores as a covariate in each of these models. We modeled the variance-covariance of the repeated measures of the dependent variable using the repeated statement. For these models, we examined several different covariance structures and autoregressive was identified to have the best fit (on the basis of Akaike’s Information Criterion, Schwarz’s Bayesian Criterion, and −2 Res Log Likelihood). Examination of residuals and fit diagnostics did not reveal substantial violations of regression model assumptions regarding linearity, homoscedasticity, independence, and normality of errors (Wicklin, 2018).

Results

Descriptive statistics by condition are reported in Table 1 for the key study variables (i.e. cognitive change-therapeutic, cognitive change-time, validation-therapeutic, validation-time, and cognitive change vs. validation), as well as for social desirability and racial attitudes. Table 2 provides correlations among study variables. The cognitive change variables were positively correlated with each other and the validation variables were also positively correlated with each other. Additionally, cognitive change-therapeutic was negatively corelated with validation-time, whereas cognitive change-time was negatively correlated with validation-therapeutic.

Table 1. Descriptive statistics and differences on study variables between vignette conditions

			Difference
Variable	BlackM(SD)	WhiteM(SD)	t	g	95% CI for g
Cognitive change-therapeutic	3.63(1.03)	4.35(0.97)	−5.30***	−0.72	[−0.99, −0.44]
Cognitive change-time	3.32(1.11)	3.86(0.96)	−3.81***	−0.52	[−0.79, −0.25]
Validation-therapeutic	4.54(0.88)	4.29(1.09)	1.85†	0.25	[−0.02, 0.52]
Validation-time	3.80(0.82)	3.71(0.91)	0.80	0.10	[−0.16, 0.37]
Cognitive change vs. validation	3.42(1.00)	2.91(0.95)	3.82***	0.52	[0.25, 0.79]
Social desirability	4.80(2.98)	4.58(3.03)	0.55	0.07	[−0.19, 0.34]
Racial attitudes	25.66(2.25)	24.75(2.99)	2.49*	0.34	[0.08, 0.61]

^†p < .10, *p < .05, ***p < .001.

Table 2. Correlations among Study 2 variables

	1.	2.	3.	4.	5.
1. Cognitive change-therapeutic	–
2. Cognitive change-time	.59***	–
3. Validation-therapeutic	−.09	−.25***	–
4. Validation-time	−.23***	.13^†	.27***	–
5. Social desirability	.10	.12^†	−.12^†	.07	–
6. Racial attitudes	.00	−.14*	.28***	.05	−.08

^†p < .10, *p < .05, ***p < .001.

Experimental tests of therapist responses to patient race

For the first two tests examining the effect of patient race on therapists’ judgments regarding cognitive change strategies (controlling for social desirability scores), we found significant effects. Specifically, therapists believed cognitive change strategies were more therapeutic (f(1, 214) = 29.09, p < .0001), and indicated that they would spend more time during session working on cognitive change strategies with White patients than with Black patients (f(1, 214) = 16.15, p < .0001).

For the second two tests, we examined the effect of patient race on therapists’ judgment of validation techniques (again controlling for social desirability scores). Patient race had a significant effect on viewing validation as therapeutic, with therapists presented with Black patients rating validation techniques as more therapeutic than therapists presented with White patients (f(1, 214) = 4.06, p = .045). The effect of patient race on validation-time was not significant (f(1, 214) = 0.48, p = .49).

For our fifth test, we examined the effect of patient race on cognitive change vs. validation (while controlling for social desirability scores). This effect was significant (f(1, 214) = 15.46, p < .0001). We also evaluated whether scores within each condition differed from the middle of scale rating (i.e. both strategies being rated as equally important). For Black patients, the mean showed validation strategies were rated as significantly more important than cognitive change strategies, t(108) = 4.38, p < .0001. For White patients, the mean did not differ significantly from the middle of the rating scale, with cognitive change strategies being numerically rated as more important than validation techniques, t(108) = −0.95, p = .34.

Relation of racial attitudes with therapist responses

We examined a regression model in which each of the key study variables were entered as predictors of racial attitudes for therapists who were presented with Black patients, while controlling for social desirability scores. See Table 3 for full model results. In this model, cognitive change-therapeutic and validation-therapeutic were significant predictors of racial attitudes. Specifically, higher ratings of cognitive change and validation as therapeutic were related to more positive racial attitudes. The remaining variables were not significantly related to racial attitudes.²

Table 3. Predictors of racial attitudes among therapists presented with Black patients

Variable	bstandardized	b	SE	t
Cognitive change-therapeutic	0.40	0.85	0.29	2.90**
Cognitive change-time	−0.18	−0.36	0.28	−1.26
Validation-therapeutic	0.35	0.91	0.27	3.38**
Validation-time	−0.04	−0.10	0.27	−0.37
Cognitive change vs. validation	0.19	0.43	0.35	1.22
Social desirability	−0.15	−0.11	0.07	−1.65

**p < .01.

Exploratory: racial salience of vignettes as a moderator of the effect of patient race on therapist’s clinical decision-making

As the latter two vignettes in the survey were intended to be more racially salient than the first two vignettes, we conducted analyses examining whether the relations between study condition (i.e. patient race) and key study variables (i.e. cognitive change-therapeutic, cognitive change-time, validation-therapeutic, validation-time, cognitive change vs. therapeutic) were moderated by the racial salience of the vignettes (i.e. low vs. high). Results of the models are reported in Table 4. The online supplement includes a figure depicting the nature of these interactions. Four of the five interactions we tested for patient race by vignette racial salience were significant. The significant interactions were identified in predicting: cognitive change-therapeutic, cognitive change-time, validation-therapeutic, and cognitive change vs. validation. First, cognitive change was seen as especially less therapeutic for Black patients for the more racially salient vignettes. Second, therapists indicated they would spend much less time on cognitive change with Black patients for the more racially salient vignettes. Third, therapists indicated validation would be particularly more therapeutic for Black patients in the racially salient vignettes. Fourth, therapists rated cognitive change as especially less important than validation for Black patients in the racially salient vignettes.

Table 4. Parameter estimates from mixed models examining interactions between patient race and racial salience of vignettes in predicting key study variables

Outcome	Predictors	b	SE	df	t
Cognitive change-therapeutic
	Intercept	3.80	0.15	212	26.04***
	Social desirability	0.04	0.02	217	1.85^†
	Patient race	0.44	0.15	217	2.94**
	Racial salience	−0.61	0.12	217	−5.03***
	Interaction	0.49	0.17	217	2.91**
Cognitive change-time
	Intercept	3.39	0.15	212	23.14***
	Social desirability	0.05	0.02	217	2.34*
	Patient race	0.28	0.15	217	1.89^†
	Racial salience	−0.57	0.10	217	−5.84***
	Interaction	0.51	0.14	217	3.74***
Validation-therapeutic
	Intercept	4.56	0.14	212	31.66***
	Social desirability	−0.04	0.02	217	−1.93^†
	Patient race	0.00	0.15	217	0.00
	Racial salience	0.29	0.11	217	2.72**
	Interaction	−0.46	0.15	217	−3.04**
Validation-time
	Intercept	3.55	0.13	212	28.17***
	Social desirability	0.02	0.02	217	1.01
	Patient race	−0.02	0.12	217	−0.14
	Racial salience	0.14	0.08	217	1.87^†
	Interaction	−0.07	0.11	217	−0.62
Cognitive change vs. validation
	Intercept	3.21	0.14	212	22.87***
	Social desirability	−0.03	0.02	217	−1.66^†
	Patient race	−0.17	0.15	217	−1.15
	Racial salience	0.67	0.13	217	5.28***
	Interaction	−0.65	0.18	217	−3.62***

The interaction in each model is the interaction of patient race and racial salience.

CC = cognitive change. ^†p < .10, *p < .05, **p < .01, ***p < .001.

Discussion

Using an experimental design, we found that therapists’ clinical decision-making in providing CBT differed when presented with Black versus White patients who were otherwise described identically. Therapists, on average, rated cognitive change strategies as less therapeutic and reported that they would spend less time on these strategies with Black as compared to White patients. Therapists also rated validation techniques as more therapeutic for Black than White patients. We also found that racial attitudes were related to therapists’ ratings of validation and cognitive change as therapeutic for Black patients, such that more positive racial attitudes were related to seeing both strategies as more therapeutic. In contrast, racial attitudes were not found to be significantly related to therapists’ ratings of the time they indicated they would spend on either strategy. Each of these results were found in the context of controlling for therapists’ tendencies to respond in a socially desirable manner.

Even when all other aspects of cases were presented identically, the race of the patient has a substantial causal effect on some important aspects of therapists’ clinical decision-making. Moreover, when asked to compare the importance of cognitive change versus validation, therapists presented with Black patients rated validation strategies as significantly more important, whereas therapists working with White patients did not rate these strategies as differing significantly on importance (though cognitive change strategies were rated higher numerically). These findings support our hypothesis that therapists tend to regard validation strategies as more appropriate to use than cognitive change strategies when treating Black patients. We suspect that therapists were inclined to think that using cognitive change strategies would risk suggesting that the therapist does not believe or fully appreciate the patients’ experience. Cognitive change procedures are intended to be directed at specific thoughts and beliefs, without questioning the validity of one’s emotional experience (Beck et al., 1979). Nonetheless, therapists may have been concerned about sending a harmful message (Hays, 2009). Therapists indicated a general preference for using cognitive change procedures less extensively with Black patients, particularly in the racially salient vignettes. There may be circumstances in which therapists would be well advised to take more time to appreciate the patients’ experience. What kinds of cognitive change procedures therapists could use in these circumstances was not addressed by our study. Future research on what kinds of cognitive change procedures are helpful in different circumstances is needed. For example, can therapists usefully focus on a patients’ inferences about an experience of discrimination (e.g. “There’s no point in trying.”).

It is also important to consider our findings in light of previous work on the process of change in psychotherapy. The limited available evidence suggests that cognitive methods are not less strongly related to outcome for Black as compared with White patients (Ezawa & Strunk, 2021). If cognitive methods are equally therapeutic across Black and White patients, our results raise the possibility that these well-intentioned efforts might serve to compromise the quality of the therapy provided to Black patients. Nonetheless, it may be important to consider the context in which these strategies are used. Therapists in our study showed not only a general tendency to use different strategies with Black than White patients, but a more marked difference in racially salient circumstances. Some studies of psychodynamic treatment have suggested that variability in the use of specific techniques or use of different techniques is related to positive outcomes (Chen et al., 2020; Owen & Hilsenroth, 2014). Although this work was not focused on race or ethnicity or on the specific strategies that we assessed, these studies highlight the need for future research on whether therapist flexibility in using cognitive change procedures may be particularly important either when working with Black patients generally or when working with Black patients who are coping with minority stress or discrimination.

Limitations and future directions

Several limitations are important to note. First, both the clinical case vignettes and key study measures were created for the purpose of this study. By using vignettes, we were able to present case information that was identical in all elements except for the image of the patient. Future research on the psychometric properties of the measures we used (including reliability and validity) is needed. Second, although vignettes of Black and White patients were identical except for the images used, the same description provided could mean something different for Black versus White patients. This was a fact that we deliberately leveraged with the racially salient vignettes. Third, we did not assess the stressfulness of the vignettes. It is possible that the intended differences in the racial salience of the vignettes may have also differed in event stressfulness. Additionally, the vignettes presented with images of Black patients could have influenced differences in racial attitudes (see Smith et al., 2008). These are additional areas for future research. Fourth, study recruitment was limited to American and Canadian therapists, and results may not generalize to therapists from other countries. Future research is needed to examine how therapists’ views may vary across different countries and regions. Despite these limitations, experimental tests of therapist decision making using clinical vignettes can address important questions often left unresolved by other kinds of psychotherapy research. We encourage researchers to continue exploring how such methods can be used to improve our understanding of how patient characteristics and contextual factors impact therapist decision-making.

Conclusions

In summary, we found that therapists indicated they would use validation techniques more extensively and make less use of cognitive methods early on in their work with Black (as opposed to White) patients. Whether these efforts improve outcomes or allow therapists to use cognitive change strategies more successfully is an important question for further investigation. We encourage researchers to continue complementing findings from treatment outcome research with experimental investigations suited to address questions about how therapists deliver treatment, the effects of specific intervention strategies, and how the answers to these questions may vary across different patient characteristics, such as race and ethnicity.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

Iony D. Ezawa was supported in part from an NIMH training grant (T32-MH18921) during completion of this work.

Notes

1. Those who did and did not complete the study did not differ significantly on any of the following variables: age, ethnicity, race, degree, portion of clinical work based on CBT, or assignment to study condition.

2. Although we planned to focus on the subset of therapists presented with African American patients for this analysis, we did also explore whether racial attitude scores were related to the key study variables in the sample of therapists presented with White patients. In this model, none of the key study variables were significantly related to racial attitudes (ps > .10).