In 1945, doctoral-level education and training programs in clinical psychology began to become accredited through American Psychological Association (APA) because of the growing demand for psychological services by such agencies as the Department of Veterans Affairs (American Psychological Association, Committee on Accreditation, 2009; Darley & Marquis, 1946; Sears, 1947) and to provide psychologists with necessary training for professional practice. Soon after the origination of these programs came the implementation of postdoctoral training programs. These programs served the purposes of providing additional, necessary, and advanced training and supervision in general and specialized areas such as clinical psychology (Hoch, Ross, & Winder, 1966; Santostefano, 1960; Weiner, 1968).
Leaders in the field generally agreed that, as an experiential component contributing to the competency for professional practice, the 1 year of clinical internship (integrated within the doctoral program’s curricula) was insufficient to produce a competent psychologist (Weiner, 1973). Therefore, postdoctoral training served to supplement doctoral education and was considered necessary for achieving entry-level competency for clinical psychologists, especially for those aspiring to have an independent, professional practice; for those aspiring to teach, supervise, or direct graduate programs; or for those aspiring to conduct clinically related research (Committee on Training in Clinical Psychology, 1947; Hoch et al., 1966). Over time, all states with the exception of Alabama that required doctoral-level training for psychologists passed legislation requiring postdoctoral training for licensure of psychologists (Plante, 1988). Unlike the doctoral programs, which were evaluated and accredited through comprehensive standard criteria based upon the recommendations of the APA’s Committee on Training in Clinical Psychology, postdoctoral training emerged with no such criteria (Committee on Training in Clinical Psychology, 1947).
From their beginnings, the structure and content of doctoral programs have undergone changes that have had and continue to have implications for postdoctoral training. The average number of predoctoral supervised hours before the predoctoral internship is now in the range of 2,000–2,500 hr (Alden et al., 2000; Clay, 2011). When added to the 2,000 supervised hours typically required in a predoctoral internship (Dittmann, 2004), this amounts to approximately 2 full years, or the amount APA’s Model Act for State Licensure of Psychologists currently posits as the minimum requirement for licensure (American Psychological Association, 2011). Because the average number of expected hours exceeds those recommended by national conferences by approximately 50%, Ko and Rodolfa (2005) suggested that the number of practicum hours has replaced the real issue of students obtaining professional competence. In terms of competence, there is little evidence to support the idea that increasing the quantity of hours increases one’s performance abilities (Beutler & Kendall, 1995; Bickman, 1999; Dyck & O’Donovan, 2003).
These issues are magnified at the postdoctoral level. Since their origination, positions for organized postdoctoral training programs were scarcely available to graduates seeking training for licensure (Schonbar & Shoben, 1957). The Association of Psychology Postdoctoral and Internship Centers (‘$$$’)currently lists 152 postdoctoral training facilities in the United States, 46 of which are accredited (Association of Psychology Postdoctoral and Internship Centers, 2014). Compared with predoctoral internships, the information available on the status of postdoctoral programs indicates even fewer opportunities for training (Clay, 2000; Robiner & Grove, 2001; Stewart & Stewart, 1998), less organization and consistency in responsibilities and start/end times (Robiner & Grove, 2001), and lower student-perceived importance (France & Wolf, 2000). The APA has placed increased emphasis on postdoctoral training because the number of APA accredited sites has increased from approximately 43 in 2008 to 101 in 2014, with 57 of those sites in clinical psychology and the remainder in the specialty areas of clinical child psychology, clinical health psychology, clinical neuropsychology, and rehabilitation psychology. Given that the mean number of residents at any one site ranges from three to six (American Psychological Association, Commission on Accreditation, 2012), this still means that the number of residents in accredited sites is far below the 2,700 doctoral candidates that graduate each year.
On the basis of the recommendations made by the APA Committee on Education and Training Leading to Licensure in Psychology in 2000 and again in 2005, the APA Council voted to change the APA Model Licensing Act in 2006. The Model Licensing Act provides for the elimination of postdoctoral training as a requirement for licensure and instead allows for appropriate and qualifying hours accrued during predoctoral training or those from postdoctoral training to fulfill the 2 years of supervised experience needed for licensure. One of these years must be a predoctoral internship. The APA still encourages postdoctoral training as an important part of a psychologist’s professional development (American Psychological Association, 2011) because it is still recognized that postdoctoral training is necessary for such things as board certification and specialty clinical training. Currently, 10 states, including Arizona, Connecticut, Indiana, Kentucky, Maryland, North Dakota, Ohio, Utah, Washington, and Wyoming, have incorporated such a change into law (Clay, 2011), and Alabama has never required postdoctoral training.
Advantages of the Licensure Change
The licensure change increases the flexibility in the sequencing of licensure requirements. Those who complete 2 years of predoctoral supervised experience, along with the other requirements for licensure, may be licensed to practice upon receiving the graduate degree. This in itself has several benefits. First, the psychologists who begin practice will earn an income and will be able to pay back their student loans. Currently, psychology students in PhD and PsyD programs alike can accumulate large amounts of debt (De Vaney Olvey, Hogg, & Counts, 2002). The APA Center for Workforce Studies (CWS) reported a median debt of $60,000 for recent psychology doctorate recipients (American Psychological Association, Center for Workforce Studies, 2009), with nearly 23% exceeding $100,000. However, graduates are unable to defer loan payments or qualify for loan forgiveness while in postdoctoral training. Dr. Katherine Nordal, Executive Director of Professional Practice for the APA, stated that the change will hopefully help “early career psychologists . . . reduce debt more quickly” (Clay, 2010, p. 38).
Second, given that students currently obtain substantially more supervised hours than required for licensure, the amount of work that psychologists need for licensure may be reduced to be more commensurate with the pay they receive. Graduates completing the proposed licensure requirements will have a minimum of 4,500 supervised hours from practicum, predoctoral, and postdoctoral experience. Psychologists have among the greatest licensure and board certification requirements compared with other professions, but they are near the bottom for median earnings, earning less pay than occupational and physical therapists, optometrists, and podiatrists, among others (Bureau of Labor Statistics, U.S. Department of Labor, 2014; De Vaney Olvey et al., 2002).
Third, for licensed psychologists who seek postdoctoral training, the facilities at which they train will be able to receive reimbursement for their work, a practice that is currently limited because unlicensed or provisionally licensed persons generally cannot bill for their services (De Vaney Olvey et al., 2002). This has the potential to increase the number of positions in postdoctoral training facilities and to provide higher quality and less variable experiences than what is currently offered.
Disadvantages of the Licensure Change
A potential disadvantage of the licensure change is with inter jurisdictional mobility (‘ASPPB’). A lack of consistency between state licensure regulations makes it difficult for those who have become licensed in states without postdoctoral training to practice in states that still require the postdoctoral training for licensure (Hall, 2013). Ross, Shah, and Rodolfa (2010) surveyed U.S. jurisdictional executive officers and reported that nearly 50% of states that require the accrual of postdoctoral hours for licensure will not allow psychologists to provide temporary services if they were licensed in other jurisdictions and did not obtain these hours. Furthermore, they stated that should such psychologists, after being independently licensed, obtain postdoctoral training in the form of a supervisory relationship with another psychologist and then present documentation of this as evidence of the completion, nearly 20% of the states will discount this as postdoctoral training. Schaffer and Rodolfa (2011) recommended that for those psychologists who desire to be mobile, they should still seek to obtain postdoctoral training experiences regardless of whether or not their states require postdoctoral training.
Another possible disadvantage is that eliminating the postdoctoral year of training may be perceived as reducing the amount of training that psychologists obtain to be licensed. Although the number of supervised hours presented in the APA Model Act has not actually been reduced with the removal of the postdoctoral requirement, the possible licensure change may appear as if psychologists are obtaining less than the necessary experience for the professional competency at the doctoral level.
Despite the controversial nature and potential impacts of the licensure change, there are few data related to the concerns of various stakeholders. The purpose of this survey was to provide information regarding stakeholders’ perceptions and attitudes on aspects of the licensure change as well as information on any adverse effects related to the licensure change.
Three hundred seventy stakeholders of the licensure requirements for psychologists served as participants for this survey. Included were persons from state psychology associations, state psychology licensing boards, and the faculty and students of doctoral training programs in clinical and counseling psychology. Table 1 provides demographic data for the participants.
In addition to the demographic data in the table, participants’ debt from graduate training related to participants’ terminal degrees was compared with data collected by the American Psychological Association, Center for Workforce Studies (2009). In terms of expected or owed debt for students and nonstudents for their graduate training, the average reported debt was $74,500 (SD = $70,263). Students had significantly greater debt (M = $127,400, SD = $59,800) than nonstudents (M = $59,700, SD = 65,800), t(259) = 6.99, p < 001. The APA CWS showed that recent doctoral graduates in health service provider subfields reported an average debt of approximately $78,000 (American Psychological Association, Center for Workforce Studies, 2009), less than students in this sample. However, most students in this sample reported pursuing a PsyD over a PhD, whereas most students in the APA CWS reported pursuing a PhD. Given that the APA survey also showed that students with PsyDs reported having more debt and a growing trend for students obtaining a PsyD, the student sample in this survey has a large amount of debt, but not unexpectedly so.
A three-part questionnaire was constructed for this survey. The first portion included three multiple-choice questions that assessed the participants’ knowledge of APA’s Model Licensure Act related to the postdoctoral requirement and the number of states that have incorporated this change into state law. The second portion included 11 statements regarding aspects of the licensure change for which participants were asked to rate their agreement/disagreement on an 11-point Likert scale with a range of 0–10. Using a content validity approach, these items were an attempt to include the issues and concerns described in the existent literature along with the reasons provided in the justification for the change in the APA Model Licensure Act. Four items were framed in a reversed direction to minimize the likelihood of a response set. The third portion included demographic questions and questions regarding any knowledge of complaints possibly related to the licensure change. The questionnaire was reviewed and piloted by six clinical psychologists. Each was asked to evaluate whether any important issues had been neglected or omitted, whether any items needed to be stated differently, and whether anything else needed to be changed with regard to the questionnaire. Revisions were made. The questionnaire was distributed via an online questionnaire program.
Three primary stakeholder groups were targeted. For students and faculty in clinical and counseling doctoral programs, contact information was obtained through online searches and by requesting the information from various institutions. To obtain responses from faculty and students in clinical and counseling doctoral programs, the link to the questionnaire was first distributed via e-mail to 103 program directors. Programs were selected from the list of APA Accredited Programs in Clinical and Counseling Psychology (American Psychological Association, 2014) and included for initial contact when accurate contact information, such as phone and e-mail address, could be found for a program director. At least two programs were chosen from each state. The directors were then contacted via telephone by the researchers and asked to complete and forward the questionnaire to other faculty and students. For clinicians, the link to the questionnaire was distributed via e-mail to all state psychology associations and the District of Columbia using the e-mail addresses available on the associations’ websites with executive directors contacted whenever possible. These associations were asked to distribute the link to their members. A similar process was used with the third stakeholder group, state psychology licensing boards. Individuals associated with these stakeholder groups were also contacted via telephone by the researchers and asked to complete and forward the questionnaire to other members of state psychology associations, members of state psychology licensing boards, and other practicing psychologists. Finally, to increase response rates, the link to the questionnaire was redistributed a second time to those whom the researchers were unable to contact via telephone. Three hundred seventy participants completed some portion of the questionnaire, and 287 completed every item.
The first three questions within the questionnaire assessed participants’ knowledge of the current APA Model Licensure Act (see Table 2). Fewer than half (42%) of participants selected the correct response for what the APA currently recommends that states require for the supervised professional experience to be licensed. Most participants (77%) selected the correct response when asked about the primary reason for APA’s change in the Model Licensure Act. When asked to select the number of states that, at the time of the survey, required a year of supervised postdoctoral training for licensure, very few (4%) responded with the correct number of 11 states and only 22% were within 2 (9–13) of the correct number. Nearly half (47%) answered within the range of 42–50 states as requiring supervised postdoctoral experience, which has been the traditional model.
Chi-square tests were performed to examine the relation between participant subgroups and answers for the factual items. For question 1, which asked about the current APA Model Act, several differences were detected. Of nonstudents, 48% selected the correct response compared with 27% of students, χ2 (1, n = 313) = 8.82, p = .003. Of those who were residents of or licensed in states without the postdoctoral training requirement, 64% selected the correct response compared with 41% of those who were residents of or licensed in states with the postdoctoral training requirement, χ2 (1, n = 305) = 6.04, p = .014. Of participants who have held a position within their state psychology licensing board, 63% selected the correct response compared with 44% of those who had not, χ2 (1, n = 250) = 5.85, p = .016. Finally, 59% of participants who had held leadership positions within their state psychology associations selected the correct answer compared with 44% of those who had not, χ2 (1, n = 251) = 4.02, p = .045.
The second part of the questionnaire assessed participants’ perceptions of various aspects and consequences of the removal of the postdoctoral year for licensure (see Table 3). On average, the descriptive statistics showed that participants neither strongly agreed nor strongly disagreed with most statements, suggesting that, as a whole, participants perceived many aspects of removing postdoctoral training as neither favorable nor unfavorable. The most notable exceptions in which participants gave higher levels of agreement were with statements that addressed opinions for future direction rather than possible consequences. Specifically, participants gave their strongest agreement with question 9, indicating that they believed that licensure requirements should be consistent across states. Likewise, participants showed agreement with question 10 that predoctoral practicum and internships should have more specific requirements if the postdoctoral year is not required.
Other questions with more moderate levels of agreement or disagreement included statements about possible consequences of removing the postdoctoral requirement. According to the entire participant sample, the most notable tradeoff of perceived consequences for removing the postdoctoral requirement would be helping to reduce student debt, but hurting psychologists’ practice mobility. In addition, on the whole, participants were less inclined to perceive removal of the postdoctoral requirement as beneficial for the growth of postdoctoral training facilities.
To determine if overall mean differences existed between subgroups on perceptions of aspects of the postdoctoral requirement (items 1–10), multivariate tests with Hotelling’s Trace statistic were conducted. An overall difference was found between students and nonstudents, F(10, 289) = 4.05, p < .001; Hotelling’s Trace = .14, partial η2 = .12. Table 3 shows the individual significant differences in the direction of students responding more favorably to aspects of the removal of the postdoctoral requirement.
Overall mean differences across perception-based items 1–10 were also found when comparing participants who answered correctly compared with incorrectly on the first two questions of the factual items (the current APA Model Act and the APA’s main reason for the change), F(10, 292) = 2.32, p = .012; Hotelling’s Trace = .08, partial η2 = .07; F(10, 292) = 2.32, p = .041; Hotelling’s Trace = .07, partial η2 = .06 (respectively). Table 4 shows participants who gave the correct responses for the first two questions responded slightly, but significantly, more favorably toward several aspects of the removal of the postdoctoral requirement than those who answered incorrectly.
Several other significant differences were detected in perceptions. Considering the number of items within the questionnaire, the researchers expected at least a small number of differences to exist between various groups. Participants who reported having held leadership positions within their state psychology associations were slightly, but significantly, more favorable toward some aspects of the removal of postdoctoral training than those who did not report having held leadership positions. Specifically, those who had held leadership positions perceived the removal of the postdoctoral requirement as more helpful for reducing student debt (M = 7.14, SD = 2.33; M = 6.03, SD = 2.46, respectively), t(248) = 3.14, p = .002, less likely to hurt the psychologists’ movement toward the physician definition of Medicare (M = 5.10, SD = 3.10; M = 6.14, SD = 2.55, respectively), t(92) = 2.41, p = .018, and as more overall favorable of the APA Model Act (M = 5.97, SD = 3.64; M = 4.87, SD = 3.45, respectively), t(248) = 2.19, p = .029.
Certain groups were also more in favor of the predoctoral practicum and internship having more specific requirements if the postdoctoral year were removed. Specifically, members of state psychology associations (M = 7.82, SD = 2.09) agreed more than nonmembers (M = 7.25, SD = 2.27), t(305) = 2.27, p = .024, and those on licensing boards (M = 8.35, SD = 1.98) agreed more than those who were not on licensing boards (M = 7.63, SD = 2.16), t(248) = 2.20, p = .029. Licensing board members (M = 7.39, SD = 2.48) also showed less agreement with having a national standard for licensure than nonboard members (M = 8.27, SD = 2.28), t(248) = 2.47, p = .014, although both groups gave moderate agreement. Finally, and not surprisingly, those who were residents of or licensed within states without the postdoctoral requirement (M = 6.79, SD = 3.29) were overall more favorable of the APA Model Act’s removal of the postdoctoral requirement than those from states with the postdoctoral requirement (M = 5.46, SD = 3.49), t(303) = 2.09, p = .037. Despite the difference, it might be noted that the distribution of responses for this item was platykurtic or flatly distributed with a slight positive skew. The most frequent response on the scale was 10 (n = 64), and the frequencies of all other responses (0–9) were within the range of 19–33.
Although results suggested that some participants had concerns that eliminating the postdoctoral requirement would lead to lower competency, only one participant described a specific ethical violation by an individual who had not completed a postdoctoral year. Participants were also asked to contribute their comments, and nearly one third (n = 90) did so. Two of the researchers placed the comments into individually created themes, resulting in the following themes listed in order of descending frequency.
- Various opinions for change and solutions (e.g., more rigid licensure requirements, postdoctoral training reserved for specialty training and board certification, among many others). Example: “If the license is provided with a predoctoral internship, then perhaps some form of credentialing could be developed that requires at least one year of postdoctoral supervision . . . Postdoctoral fellowships should place the emphasis on specialty training and board certification.”
- Concerns about negative consequences for removal of postdoctoral training (e.g., decline of competent practice, lack of readiness for recent graduates). Example: “I believe the way to be competitive in the health marketplace is to be the best trained clinicians . . . Less training is not better and we are missing the big picture.”
- Comments on benefits of having a postdoctoral internship (e.g., obtaining valuable clinical experience, helping to protect the public). Example: “Having the postdoctoral year allows a person to become more independently confident in practice and allows for additional training if needed.”
- Feedback including appreciation, criticism, and suggestions for change regarding the questionnaire. Example: “I am glad that this data is being collected and that alternatives to the current postdoctoral requirement in most states are being seriously considered. I supervise postdoctoral employees and find that they are usually ready, at the beginning of the year, to be entry-level psychologists . . .”
- Desire for nationally regulated licensure requirements for psychologists. Example: “Consistency for licensure requirements across states would significantly decrease the burden on early career psychologists.”
- Students or recent graduates expressing negative impacts of the postdoctoral requirement (e.g., difficulty of becoming licensed because of being unable to find a postdoc position, not knowing which state’s licensure requirements to pursue because of uncertainty of job availability). Example: “In California, it is hard to find a postdoc position . . . The postdoc positions are setting higher requirements because there are so many applicants making it impossible to apply.”
The statistical results described previously suggested that participants’ perceptions, as a whole, were moderate about many aspects of the removal of the postdoctoral requirement. However, those participants who provided this large number of comments were more polarized because they strongly advocated for or against the supervised postdoctoral year of training.
The results of the first portion of the questionnaire, which assessed how knowledgeable the sample was of the APA Model Act for State Licensure of Psychologists, revealed that participants generally lacked awareness of the current “prototype” (American Psychological Association, 2011, p. 1) of licensure requirements and the number of states that have incorporated it into law. The lack of knowledge of the APA Model Act is surprising considering many of the participants who made up this survey’s sample are considered “leaders” within the field (psychology licensing board members, state psychology association members, and faculty of doctoral programs). Although this information is relevant for students who are pursuing licensure, students were even less informed than nonstudents. Possible explanations for the information gap include poor dissemination and seeking of such information. Another reason may be that states have historically had strong independence in creating their own laws, making APA’s leadership role less clear. The guidelines and prototypes set by the APA may not be widely supported and, consequently, not sought out or understood.
A strong theme that emerged was that, with the exception of students (who were less informed than nonstudents), those who were more informed about the APA Model Act responded more favorably toward aspects of the removal of the postdoctoral year than those who were not informed. This was especially true for those who have held positions of leadership within their state psychology associations and those who were residents of or licensed in states that have made the legal change to remove postdoctoral training. The impression given is that being informed serves to direct those in the field to be stronger advocates of the licensure change. If this is true, then the implication for those who advocate less or oppose the licensure change is that they may do so partly because they are less informed. Perhaps the tendency for people to like that which is familiar and dislike that which is unfamiliar relates to this finding. If state associations wish to garner support to remove the postdoctoral requirement, then these results suggest that they should educate their members and licensees about the number of training hours now being accrued by students in predoctoral years. Psychologists who obtained their degrees many years ago may not be aware of the changes that have occurred in training.
The second portion of the questionnaire assessed current attitudes and perceptions of the aspects of removing the postdoctoral year as a requirement for licensure. There was a clear difference in the level of support between students and nonstudents. Students showed a more favorable response and believed that removing the postdoctoral requirement will be beneficial for helping to reduce their debt. As previously mentioned, students and recent graduates currently have substantial debt, and their ability to pay is directly affected by their ability to become licensed and bill for services. Because many are either unable to obtain postdoctoral training or receive little pay during postdoctoral training, it is consistent that this group would see the ability to become licensed upon graduating as an attractive option.
In terms of competency, students also moderately agreed that they have sufficient skills for entry-level professional practice without having had a year of postdoctoral training. The issue of competency was addressed in many of the comments, and many opinions were given for what training constitutes a competent professional within the field. Although this issue has been discussed and debated since the initial development of graduate psychology programs, it continues to remain an important topic. Issues such as the Patient Protection and Affordable Care Act and integrative care (Chor, Olin, & Hoagwood, 2014; Rozensky, 2013) present new challenges that need to be addressed in a comprehensive training model with the postdoctoral requirement being one of many concerns. As previously mentioned, participants as a whole were split on their perceptions of many aspects of postdoctoral training and gave neither strong agreement nor disagreement with most statements in the questionnaire. The area with the most conclusive agreement was the desire for all states to have the same licensure requirements. This was also a theme within the participants’ comments. It seems that the desire for consistency in the language of what constitutes a licensed psychologist is strong within the field, but there seems to be little agreement on whether or not this consistency includes a year of supervised postdoctoral training.
The major limitation of this survey is related to the composition of the sample. For example, although data were collected from participants in 47 states and territories in the United States, including students in 20 states, approximately 32% of participants reported being from or licensed in the state of Missouri where the survey originated. Although not ideal, no significant differences in perceptions between participants in Missouri and those in other states were detected. In addition, the data were obtained through various electronic requests, which in no way assured a representative sample. The demographic data, such as student debt, were similar to what has been obtained in other, larger samples, but they still may not be representative. Finally, most of the attrition in this survey occurred within the first portion of the questionnaire (see Table 2). One suspected reason for this is that participants did not want to respond to questions for which they did not know the answers. In the early distribution of questionnaires, several potential respondents contacted the authors, stating that they did not feel it was appropriate that they should answer the questionnaire given that they knew so little about the topic. A sentence was inserted into the instructions for the questionnaire asking respondents to complete the questionnaire regardless of their confidence in responding accurately to the initial questions. It may be that those who did not know the answers to these questions are different from those who responded to the entire questionnaire. In any case, although possibly not truly representative of all stakeholder groups, the current survey offers a snapshot of the profession’s view of the postdoctoral requirement.
In addition to questions concerning the representativeness of the sample, the reliability and validity of the questionnaire itself could be questioned. It may be that the questionnaire did not adequately sample the universe of items related to the postdoctoral issue or items may have not have been correctly worded. This survey provides an initial, exploratory picture of the limited knowledge and differing attitudes that members in the field have in regard to the APA Model Licensure Act and aspects of postdoctoral training. This picture matches and gives insight into the disarray experienced with the varying requirements for licensure of psychologists. An unanswered question that remains important is to what extent do the perceived benefits and costs of the removal of the postdoctoral year of training compare with the actual consequences. To effectively inform the field of real impacts, whatever they may be, seems to be a key step in moving toward the consistency that is desired by most.
- Full Credit Given to: Boon, A. T., Lutz, D. J., & Marburger, K. M. (2015). Eliminating Postdoctoral Training as a Requirement for Licensure: Perceptions and Anticipated Impacts. Professional Psychology: Research & Practice, 46(1), 62-69. doi:10.1037/a0038198
- Biographical Information for Authors: Austin T. Boon earned his master’s degree in clinical psychology from Missouri State University. He provides counseling services and facilitates domestic violence prevention programs at a community mental health center and serves as a per course instructor at Missouri State University in Springfield, Missouri.
- David J. Lutz earned his PhD in clinical psychology from the University of Kansas. He is currently a professor in the Department of Psychology at Missouri State University and is in private practice specializing in disability and vocational assessment. His research interests focus on professional and ethical issues in psychology.
- Katie M. Marburger earned her bachelor’s degree in psychology from Missouri State University and plans to pursue a master’s degree in public health.
- Acknowledgement: The authors thank Robert Cramer, Steve Capps, Tim Daugherty, Emil Rodolfa, and three anonymous reviewers for their comments on previous drafts of this article.