In the focal article, Melson-Silimon, Harris, Shoenfelt, Miller, and Carter (Reference Melson-Silimon, Harris, Shoenfelt, Miller and Carter2019) discuss the implications for general personality and personality disorder research in the context of organizational testing and the Americans with Disabilities Act (ADA). The authors make recommendations regarding the use of personality tests in these contexts as well as how practitioners should consider which tests to choose. We agree that a better understanding of the distinction between general and clinical models of personality assessment is imperative in the context of employment decisions. Additionally, we agree that these models share structural similarities that are important for organizational researchers and practitioners to understand. However, we contend that it is also important for these and other stakeholder groups (e.g., clinicians, employees) to recognize the conceptual distinctions between the models, as well as the appropriate roles of those involved in assessment.
Since the inception of the categorical model of personality disorders that was introduced in the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM; American Psychiatric Association [APA], 1980), critics have argued that the model sacrificed validity in the service of increasing reliability. These arguments were supported by problems of excessive comorbidity, substantial diagnostic heterogeneity, inadequate symptom coverage, diagnostic thresholds that were considered arbitrary, and the framework’s basis in a personality typology model that is inconsistent with years of personality science research (Morey, Benson, Busch, & Skodol, Reference Morey, Benson, Busch and Skodol2015). With the inclusion of the Alternative Model for Personality Disorders (AMPD) in Section III of the DSM-5 (APA, 2013), the classification and assessment of personality disorders in the DSM is moving toward a stronger empirical foundation, with personality disorders defined, in part, as maladaptive variants of normal range personality traits. Although we agree with the authors’ conclusion that more conversation is needed across psychology disciplines and the EEOC in light of these changes, we are concerned that the focal article appears to suggest that DSM revisions have resulted in measures of normal range personality traits now becoming de facto clinical measures that can be used to infer the presence of personality disorders, thus making them susceptible to ADA challenges. Measures of general personality have long shown relevance to clinical practice (Costa & McCrae, Reference Costa and McCrae1992), though by nature, their utility in assessing abnormal personality characteristics and psychopathology is limited. Three areas of discussion appear relevant in this case:
1. Measures developed to assess general personality traits do not adequately assess maladaptive traits.
2. The presence of maladaptive traits is not sufficient to infer the presence of a personality disorder.
3. Extrapolations about measures of general personality traits from case law involving the MMPI are not compelling.
Measures of general personality and the assessment of maladaptive traits
The Big Five trait constructs are suggested to represent the “building blocks” of personality structure, and maladaptive traits underpinning abnormal personality have shown to reflect extreme variants of these same traits (Markon, Krueger, & Watson, Reference Markon, Krueger and Watson2005). This dimensional view of psychopathology casts great doubt about the existence of any “line in the sand” between normal and abnormal personality. As noted 15 years ago, “In the last 20 years, the categorical approach has been increasingly questioned as evidence has accumulated that the so-called categorical disorders … seem to merge imperceptibly both into one another and into normality … with no demonstrable natural boundaries” (First, Reference First, Tasman, Kay and Lieberman2003, p. 661). So although it is indeed true that understanding personality disorders requires an understanding of general personality, we concur with other researchers that maladaptive personality measures are necessary to assess maladaptive traits (Dilchert, Ones, & Krueger, Reference Dilchert, Ones and Krueger2014).
Most Big Five instruments do not possess the content or sensitivity necessary for the assessment of maladaptive traits, as they were developed for the study of general personality and therefore designed to provide information within the moderate range of the larger continuum (Dilchert, Ones, & Krueger, Reference Dilchert, Ones and Krueger2014). For example, although numerous studies have demonstrated that measures of Big Five traits are associated with symptoms and traits of corresponding personality disorders, several studies have not supported this association (see Mullins-Sweatt & Widiger, Reference Mullins-Sweatt, Widiger, Krueger and Tackett2006, for a review). In the case of the NEO-PI-R (Costa & McCrae, Reference Costa and McCrae1992), this is likely due to the fact that items are more likely to describe behaviors considered adaptive or desirable, rather than maladaptive. This was demonstrated empirically by Haigler and Widiger (Reference Haigler and Widiger2001), who showed that relevant Big Five traits exhibited minimal associations with personality disorders considered to reflect maladaptive variants of these traits, but when NEO-PI-R items were altered to suggest the behavior was maladaptive (e.g., by adding words/phrases such as “too much” or “excessively”), correlations strengthened considerably. In sum, to make inferences about personality disorders based on measures of personality, item content assessing maladaptive personality traits is critical.
Personality disorders are more than just maladaptive traits
The focal article states that “[n]onclinical assessments of normal personality traits thus far have been permitted under the ADA because of the belief that they do not provide medical information” (Melson-Silimon et al., Reference Melson-Silimon, Harris, Shoenfelt, Miller and Carter2019, p. 120), but suggests that this position may soon be untenable because of evolving views on the dimensional nature of personality and psychopathology. We further suggest that making inferences about the possible presence of a personality disorder based on extreme scores on a measure of general personality is not defensible because maladaptive traits are just one part of a personality disorder diagnosis. Although the AMPD does stipulate that a personality disorder diagnosis requires the presence of at least one maladaptive personality trait, the diagnostic scheme also requires moderate or greater impairment in personality functioning, defined as disturbances in self and interpersonal functioning (APA, 2013). In fact, the general impairment in personality functioning is considered a core component of a personality disorder, with the maladaptive traits reflecting individual differences in the phenotypic manifestation of the personality disorder. Both are necessary, and the presence of maladaptive traits does not necessarily entail functional impairment. In other words, personality disorders display exaggerations of personality styles and therefore reflect characteristics existing in all individuals (Oldham, Reference Oldham1995; Oldham & Morris, Reference Oldham and Morris2012; Thompson, Payne, Horner, & Morey, Reference Thompson, Payne, Horner and Morey2012); yet the degree of dysfunction is crucial in determining clinical assessments.
Emerging research in the field of psychopathology is demonstrating that much is left unexplained when it comes to inferring functional impairment based on symptomology (or traits). Recent meta-analyses have suggested that anxiety disorder symptomology accounts for just 12% of the variance in functional impairment (McKnight, Monfort, Kashdan, Blalock, & Calton, Reference McKnight, Monfort, Kashdan, Blalock and Calton2016) and depressive symptomology accounts for just 25% of the variance in impairment (McKnight & Kashdan, Reference McKnight and Kashdan2009). An example of this disconnect can be found in a case study reported by Costa and McCrae (2005), in which a woman who scored quite low on agreeableness experienced no evidence of occupational impairment in her job as an attorney, as her aggressive interpersonal style was useful in adversarial contexts. Even in the case of neuroticism, a vulnerability factor for the development of psychopathology and a variable that saturates most measures of psychopathology, support for the benefits of high neuroticism in busy work environments has been found (Smillie, Yeo, Furnham, & Jackson, Reference Smillie, Yeo, Furnham and Jackson2006). Taken together, although information about elevated or extreme personality traits is surely inappropriately used to screen out individuals who may be at risk for psychopathology, this is not the same as saying the test actually provides medical information.
Previous case law
The focal article provided an interesting and informative review of the case law pertaining to personality assessment in employee selection. Conclusions based on this review appear to be based on the following logic: clinical assessments designed to provide information about abnormal personality and psychopathology (such as the MMPI) have been determined to be medical tests; therefore, given a more explicit recognition by the DSM that personality disorders reflect extreme variants of Big Five traits, general measures of personality are at risk of similar categorization as a medical test. Although the previous two sections also speak to this issue, there are two additional points to add. First, the purpose of the MMPI is to aid in psychiatric assessment and diagnosis and to distinguish between normal and abnormal groups. This is in direct contrast to measures of general personality, which although they may be quite useful in clinical practice, primarily attempt to comprehensively assess normal range personality characteristics. Second, the MMPI was developed using a criterion-keying strategy, resulting in item content that may lack face validity but is empirically related to certain psychiatric content areas. In contrast, measures of general personality such as the NEO-PI-R are relatively unambiguous with regard to their underlying meaning. This is a critical distinction when considering that job applicants prefer selection methods with high face validity (Ni & Hauenstein, Reference Ni and Hauenstein1998; Rynes & Connerley, Reference Rynes and Connerley1993). Overall, we wonder if speculation about normal range personality instruments based on case law primarily involving a clinical measure of psychopathology that includes invasive and non-face-valid items is an “apples to oranges” comparison.
Utility of general personality instruments in the workplace
In their influential meta-analysis, Barrick and Mount (Reference Barrick and Mount1991) identified that general personality traits (i.e., the five-factor model) are differentially related to unique measures of job performance as well as training proficiency. Furthermore, specific traits are predictive of performance for certain occupations whereas others are valid across job groups. Correspondingly, the work by these authors and their colleagues has found consistent evidence for the use of general personality in organizational contexts (Barrick & Mount, Reference Barrick and Mount1991; Barrick, Mount, & Judge, Reference Barrick, Mount and Judge2001; Barrick, Stewart, Newbert, & Mount, Reference Barrick, Stewart, Neubert and Mount1998).
Conclusions and recommendations
We appreciate the attention the focal article brings to the common structure of normal and abnormal models of personality and the possible implications for personality testing in the workplace. At this time, our interpretation of the research is that if “normal” range personality measures are being used to screen for psychopathology at the time of employee selection, then they are inappropriately being used as medical tests. As the focal article contends, both general and clinical personality models belong on a common continuum. Therefore, they reflect characteristics that all individuals possess to varying degrees (Oldham, Reference Oldham1995; Oldham & Morris, Reference Oldham and Morris2012). However, researchers or practitioners using “normal” personality assessments to identify maladaptive characteristics for diagnostic purposes would be akin to individuals using knowledge of health symptoms to self-diagnose without speaking to a medical doctor.
According to Section 9.02(a) of the APA ethics codes, “Psychologists administer, adapt, score, interpret, or use assessment techniques, interviews, tests, or instruments in a manner and for purposes that are appropriate in light of the research on or evidence of the usefulness and proper application of the techniques.” Based on the evidence provided in this commentary, we suggest that general personality measures used by industrial-organizational psychologists may be appropriate and valid for certain workplace questions, but these same tests are not valid or appropriate as a diagnostic screening tool following this guideline:
General personality measures = valid/appropriate for certain workplace questions; not valid/appropriate for screening for personality disorders or other forms of psychopathology outside of clinical settings where other medical information is available, and the assessment is completed by a mental health professional
Related to the issue of personality disorders in the workplace specifically, we strongly encourage research into how appropriate accommodations can be designed for these heterogeneous, but common, conditions. As the focal article states, job analysis is a necessary component of personality assessment in employee selection. However, this argument is typically in the context of justification for adverse criteria decisions rather than understanding how the work environment can be designed to best meet the needs of employees. In contrast, we recommend that researchers and practitioners explore ways in which organizations can support and/or accommodate those with diagnosed personality disorders and high levels of personality dysfunction. In other words, rather than using personality assessments for diagnostic or employment decision-making purposes, identifying how those who have been assessed by a medical professional can be integrated into the workplace would be an important area for research to focus on. Given that approximately 15% of the U.S. population is purported to possess one or more personality disorders (Grant et al., Reference Grant, Hasin, Stinson, Dawson, Chou, Ruan and Pickering2004), and these traits reflect characteristics all individuals vary on to some degree, it is critical to understand how employers can facilitate productive working environments for all individuals. Additionally, employment typically provides access to health care as well as a pathway to stability and well-being in one’s life. Therefore, we argue that employers should seek ways to include rather than exclude these individuals into the workplace.
As Stone and Colella (Reference Stone and Colella1996) explain, the reactions of employees and employers to individuals with various types of disabilities are not uniform. Therefore, providing accommodations for individuals with maladaptive personality characteristics may be challenging. However, the Job Accommodation Network (JAN) provides a host of resources regarding accommodations for varying types of medical conditions, including suggestions regarding workplace supports, job or environmental restructuring, and policy recommendations. In a series of publications addressing accommodation and compliance issues, JAN (2018) notes that limitations among individuals with personality disorders will vary and therefore so will the degree of accommodations required.
We encourage continued research into the assessment of maladaptive traits in workplace settings, including research examining the associations between general personality and maladaptive trait measures. As all employees vary along personality dimensions, it is imperative to better understand the ways in which all aspects of personality relate to workplace functioning.
You have
Access
In the focal article, Melson-Silimon, Harris, Shoenfelt, Miller, and Carter (Reference Melson-Silimon, Harris, Shoenfelt, Miller and Carter2019) discuss the implications for general personality and personality disorder research in the context of organizational testing and the Americans with Disabilities Act (ADA). The authors make recommendations regarding the use of personality tests in these contexts as well as how practitioners should consider which tests to choose. We agree that a better understanding of the distinction between general and clinical models of personality assessment is imperative in the context of employment decisions. Additionally, we agree that these models share structural similarities that are important for organizational researchers and practitioners to understand. However, we contend that it is also important for these and other stakeholder groups (e.g., clinicians, employees) to recognize the conceptual distinctions between the models, as well as the appropriate roles of those involved in assessment.
Since the inception of the categorical model of personality disorders that was introduced in the third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM; American Psychiatric Association [APA], 1980), critics have argued that the model sacrificed validity in the service of increasing reliability. These arguments were supported by problems of excessive comorbidity, substantial diagnostic heterogeneity, inadequate symptom coverage, diagnostic thresholds that were considered arbitrary, and the framework’s basis in a personality typology model that is inconsistent with years of personality science research (Morey, Benson, Busch, & Skodol, Reference Morey, Benson, Busch and Skodol2015). With the inclusion of the Alternative Model for Personality Disorders (AMPD) in Section III of the DSM-5 (APA, 2013), the classification and assessment of personality disorders in the DSM is moving toward a stronger empirical foundation, with personality disorders defined, in part, as maladaptive variants of normal range personality traits. Although we agree with the authors’ conclusion that more conversation is needed across psychology disciplines and the EEOC in light of these changes, we are concerned that the focal article appears to suggest that DSM revisions have resulted in measures of normal range personality traits now becoming de facto clinical measures that can be used to infer the presence of personality disorders, thus making them susceptible to ADA challenges. Measures of general personality have long shown relevance to clinical practice (Costa & McCrae, Reference Costa and McCrae1992), though by nature, their utility in assessing abnormal personality characteristics and psychopathology is limited. Three areas of discussion appear relevant in this case:
1. Measures developed to assess general personality traits do not adequately assess maladaptive traits.
2. The presence of maladaptive traits is not sufficient to infer the presence of a personality disorder.
3. Extrapolations about measures of general personality traits from case law involving the MMPI are not compelling.
Measures of general personality and the assessment of maladaptive traits
The Big Five trait constructs are suggested to represent the “building blocks” of personality structure, and maladaptive traits underpinning abnormal personality have shown to reflect extreme variants of these same traits (Markon, Krueger, & Watson, Reference Markon, Krueger and Watson2005). This dimensional view of psychopathology casts great doubt about the existence of any “line in the sand” between normal and abnormal personality. As noted 15 years ago, “In the last 20 years, the categorical approach has been increasingly questioned as evidence has accumulated that the so-called categorical disorders … seem to merge imperceptibly both into one another and into normality … with no demonstrable natural boundaries” (First, Reference First, Tasman, Kay and Lieberman2003, p. 661). So although it is indeed true that understanding personality disorders requires an understanding of general personality, we concur with other researchers that maladaptive personality measures are necessary to assess maladaptive traits (Dilchert, Ones, & Krueger, Reference Dilchert, Ones and Krueger2014).
Most Big Five instruments do not possess the content or sensitivity necessary for the assessment of maladaptive traits, as they were developed for the study of general personality and therefore designed to provide information within the moderate range of the larger continuum (Dilchert, Ones, & Krueger, Reference Dilchert, Ones and Krueger2014). For example, although numerous studies have demonstrated that measures of Big Five traits are associated with symptoms and traits of corresponding personality disorders, several studies have not supported this association (see Mullins-Sweatt & Widiger, Reference Mullins-Sweatt, Widiger, Krueger and Tackett2006, for a review). In the case of the NEO-PI-R (Costa & McCrae, Reference Costa and McCrae1992), this is likely due to the fact that items are more likely to describe behaviors considered adaptive or desirable, rather than maladaptive. This was demonstrated empirically by Haigler and Widiger (Reference Haigler and Widiger2001), who showed that relevant Big Five traits exhibited minimal associations with personality disorders considered to reflect maladaptive variants of these traits, but when NEO-PI-R items were altered to suggest the behavior was maladaptive (e.g., by adding words/phrases such as “too much” or “excessively”), correlations strengthened considerably. In sum, to make inferences about personality disorders based on measures of personality, item content assessing maladaptive personality traits is critical.
Personality disorders are more than just maladaptive traits
The focal article states that “[n]onclinical assessments of normal personality traits thus far have been permitted under the ADA because of the belief that they do not provide medical information” (Melson-Silimon et al., Reference Melson-Silimon, Harris, Shoenfelt, Miller and Carter2019, p. 120), but suggests that this position may soon be untenable because of evolving views on the dimensional nature of personality and psychopathology. We further suggest that making inferences about the possible presence of a personality disorder based on extreme scores on a measure of general personality is not defensible because maladaptive traits are just one part of a personality disorder diagnosis. Although the AMPD does stipulate that a personality disorder diagnosis requires the presence of at least one maladaptive personality trait, the diagnostic scheme also requires moderate or greater impairment in personality functioning, defined as disturbances in self and interpersonal functioning (APA, 2013). In fact, the general impairment in personality functioning is considered a core component of a personality disorder, with the maladaptive traits reflecting individual differences in the phenotypic manifestation of the personality disorder. Both are necessary, and the presence of maladaptive traits does not necessarily entail functional impairment. In other words, personality disorders display exaggerations of personality styles and therefore reflect characteristics existing in all individuals (Oldham, Reference Oldham1995; Oldham & Morris, Reference Oldham and Morris2012; Thompson, Payne, Horner, & Morey, Reference Thompson, Payne, Horner and Morey2012); yet the degree of dysfunction is crucial in determining clinical assessments.
Emerging research in the field of psychopathology is demonstrating that much is left unexplained when it comes to inferring functional impairment based on symptomology (or traits). Recent meta-analyses have suggested that anxiety disorder symptomology accounts for just 12% of the variance in functional impairment (McKnight, Monfort, Kashdan, Blalock, & Calton, Reference McKnight, Monfort, Kashdan, Blalock and Calton2016) and depressive symptomology accounts for just 25% of the variance in impairment (McKnight & Kashdan, Reference McKnight and Kashdan2009). An example of this disconnect can be found in a case study reported by Costa and McCrae (2005), in which a woman who scored quite low on agreeableness experienced no evidence of occupational impairment in her job as an attorney, as her aggressive interpersonal style was useful in adversarial contexts. Even in the case of neuroticism, a vulnerability factor for the development of psychopathology and a variable that saturates most measures of psychopathology, support for the benefits of high neuroticism in busy work environments has been found (Smillie, Yeo, Furnham, & Jackson, Reference Smillie, Yeo, Furnham and Jackson2006). Taken together, although information about elevated or extreme personality traits is surely inappropriately used to screen out individuals who may be at risk for psychopathology, this is not the same as saying the test actually provides medical information.
Previous case law
The focal article provided an interesting and informative review of the case law pertaining to personality assessment in employee selection. Conclusions based on this review appear to be based on the following logic: clinical assessments designed to provide information about abnormal personality and psychopathology (such as the MMPI) have been determined to be medical tests; therefore, given a more explicit recognition by the DSM that personality disorders reflect extreme variants of Big Five traits, general measures of personality are at risk of similar categorization as a medical test. Although the previous two sections also speak to this issue, there are two additional points to add. First, the purpose of the MMPI is to aid in psychiatric assessment and diagnosis and to distinguish between normal and abnormal groups. This is in direct contrast to measures of general personality, which although they may be quite useful in clinical practice, primarily attempt to comprehensively assess normal range personality characteristics. Second, the MMPI was developed using a criterion-keying strategy, resulting in item content that may lack face validity but is empirically related to certain psychiatric content areas. In contrast, measures of general personality such as the NEO-PI-R are relatively unambiguous with regard to their underlying meaning. This is a critical distinction when considering that job applicants prefer selection methods with high face validity (Ni & Hauenstein, Reference Ni and Hauenstein1998; Rynes & Connerley, Reference Rynes and Connerley1993). Overall, we wonder if speculation about normal range personality instruments based on case law primarily involving a clinical measure of psychopathology that includes invasive and non-face-valid items is an “apples to oranges” comparison.
Utility of general personality instruments in the workplace
In their influential meta-analysis, Barrick and Mount (Reference Barrick and Mount1991) identified that general personality traits (i.e., the five-factor model) are differentially related to unique measures of job performance as well as training proficiency. Furthermore, specific traits are predictive of performance for certain occupations whereas others are valid across job groups. Correspondingly, the work by these authors and their colleagues has found consistent evidence for the use of general personality in organizational contexts (Barrick & Mount, Reference Barrick and Mount1991; Barrick, Mount, & Judge, Reference Barrick, Mount and Judge2001; Barrick, Stewart, Newbert, & Mount, Reference Barrick, Stewart, Neubert and Mount1998).
Conclusions and recommendations
We appreciate the attention the focal article brings to the common structure of normal and abnormal models of personality and the possible implications for personality testing in the workplace. At this time, our interpretation of the research is that if “normal” range personality measures are being used to screen for psychopathology at the time of employee selection, then they are inappropriately being used as medical tests. As the focal article contends, both general and clinical personality models belong on a common continuum. Therefore, they reflect characteristics that all individuals possess to varying degrees (Oldham, Reference Oldham1995; Oldham & Morris, Reference Oldham and Morris2012). However, researchers or practitioners using “normal” personality assessments to identify maladaptive characteristics for diagnostic purposes would be akin to individuals using knowledge of health symptoms to self-diagnose without speaking to a medical doctor.
According to Section 9.02(a) of the APA ethics codes, “Psychologists administer, adapt, score, interpret, or use assessment techniques, interviews, tests, or instruments in a manner and for purposes that are appropriate in light of the research on or evidence of the usefulness and proper application of the techniques.” Based on the evidence provided in this commentary, we suggest that general personality measures used by industrial-organizational psychologists may be appropriate and valid for certain workplace questions, but these same tests are not valid or appropriate as a diagnostic screening tool following this guideline:
General personality measures = valid/appropriate for certain workplace questions; not valid/appropriate for screening for personality disorders or other forms of psychopathology outside of clinical settings where other medical information is available, and the assessment is completed by a mental health professional
Related to the issue of personality disorders in the workplace specifically, we strongly encourage research into how appropriate accommodations can be designed for these heterogeneous, but common, conditions. As the focal article states, job analysis is a necessary component of personality assessment in employee selection. However, this argument is typically in the context of justification for adverse criteria decisions rather than understanding how the work environment can be designed to best meet the needs of employees. In contrast, we recommend that researchers and practitioners explore ways in which organizations can support and/or accommodate those with diagnosed personality disorders and high levels of personality dysfunction. In other words, rather than using personality assessments for diagnostic or employment decision-making purposes, identifying how those who have been assessed by a medical professional can be integrated into the workplace would be an important area for research to focus on. Given that approximately 15% of the U.S. population is purported to possess one or more personality disorders (Grant et al., Reference Grant, Hasin, Stinson, Dawson, Chou, Ruan and Pickering2004), and these traits reflect characteristics all individuals vary on to some degree, it is critical to understand how employers can facilitate productive working environments for all individuals. Additionally, employment typically provides access to health care as well as a pathway to stability and well-being in one’s life. Therefore, we argue that employers should seek ways to include rather than exclude these individuals into the workplace.
As Stone and Colella (Reference Stone and Colella1996) explain, the reactions of employees and employers to individuals with various types of disabilities are not uniform. Therefore, providing accommodations for individuals with maladaptive personality characteristics may be challenging. However, the Job Accommodation Network (JAN) provides a host of resources regarding accommodations for varying types of medical conditions, including suggestions regarding workplace supports, job or environmental restructuring, and policy recommendations. In a series of publications addressing accommodation and compliance issues, JAN (2018) notes that limitations among individuals with personality disorders will vary and therefore so will the degree of accommodations required.
We encourage continued research into the assessment of maladaptive traits in workplace settings, including research examining the associations between general personality and maladaptive trait measures. As all employees vary along personality dimensions, it is imperative to better understand the ways in which all aspects of personality relate to workplace functioning.