Psychotherapy Bulletin

Psychotherapy Bulletin

Seeking the “Perfect” Match

An Empirical Examination of Applicant Differences According to Internship Setting

Despite ongoing efforts to conceptualize and envision possible solutions to resolve the internship imbalance, the problem has continued to escalate (e.g., Baker, McCutcheon, & Keilin, 2007; Grus, McCutcheon, & Berry, 2011; Keilin, Baker, McCutcheon, & Peranson, 2007; McCutcheon, 2011; Rodolfa, Bell, Bieschke, Davis, & Peterson, 2007). The number of students entering the Association of Psychology Postdoctoral and Internships Centers (APPIC) match has increased over the past 10 to 15 years, steadily outpacing the growth of available internship positions (Hatcher, 2011; Larkin 2012). In the 2013 match, upward of 1,000 internship applicants went unmatched, for an overall match rate of 75% (see 2013 Match Statistics). Phase I of the 2014 match process suggests similar trends, with 77.1% of Ph.D. students and 69.0% of Psy.D. students having successful matched, leaving 801 participating applicants (20%) unmatched going into Phase II (see 2014 Match Statistics).

What may be overlooked as one considers these numbers is that not all internships are created equal in terms of their training value to interns with varying needs. The internship is a critical developmental step that serves as a training capstone experience and is often a time of intense and rapid professional growth (Collins, Callahan, & Klonoff, 2007). Completion of internship promotes integration of knowledge and aids in the successful transition from graduate training to emergence into the profession (Collins et al., 2007; Lamb, Baker, Jennings, & Yarvis, 1982). Thus, beyond matching, it is the securing of an internship that is consistent with identified training and career goals that is a vital aim to internship applicants (Stedman, 2007).

Recent reports offer internship applicants some empirical information about the characteristics associated with securing a match (Callahan, Collins, & Klonoff, 2010; Callahan, Hogan, Klonoff, & Collins, in press; Ginkel, Davis, & Michael, 2010); however, there are no empirical studies in the literature exploring whether there are characteristic differences among successfully matched internship applicants according to the type of internship setting. Identifying characteristics associated with successfully matched applicants across internship settings could be beneficial in (1) helping students identify how they may best prepare for the particular type(s) of internship that is most appropriate for their specific training needs and career goals, and/or (2) aiding students and their mentors in the subjective appraisal of readiness to competitively pursue internship at the type of site to which they are hoping to match. In light of the conceptual literature, we hypothesized that there would be characteristic differences among successfully matched interns according to internship site. However, we did not make any hypotheses regarding specific differences.


The dataset for this investigation was constructed by compiling data gathered in two earlier internship match studies (Callahan et al., 2010; Callahan et al., in press), but has not been previously combined or analyzed in the manner presented in the current study. In light of space constraints, a comprehensive review of the methodologies employed in those studies will not be duplicated herein, but interested readers are encourage to consult the original sources for more information.


Our primary goal was to examine whether there are meaningful differences among applicants matched to different internship settings: armed forces medical center (n = 7); child/adolescent psychiatry/pediatrics (n = 182); community mental health center (n = 73); consortium (n = 81); medical school (n = 222); prison or other correctional center (n = 21); private general hospital (n = 64); private outpatient clinic (n = 43); psychiatric hospital (n = 38); psychology department (n = 27); school district (n = 18); university counseling center (n = 59); veterans administration (VA) medical center (n = 246). A comprehensive table with data on all 14 types of sites can be found online (scroll to manuscript citation and click on “Supplemental Material” link at:

The comprehensive table contains averages and standard deviations for interns who were matched to each of the given types of sites, as well as the lowest, highest, and quartile values for each of the variables. This information may help students plan their training experiences to meet the expectations of the types of sites to which they intend to apply. For example, one might notice 75% of all interns who are matched to a VA setting for their internship have at least 694 intervention/assessment hours, 50% have at least 912 hours, and 25% have at least 1112 hours. A student who wants to complete an internship at a VA may want to make sure that he or she accrues at least 694 hours in order to be considered sufficient.

In addition to considering the volume of experiences that are typical for each of the sites, students should also consider the type of experiences that are reviewed favorably. For example, over 75% of interns who are matched to correctional facilities have had some previous practicum experience in a forensic setting. In contrast, few interns at any of the other settings have had a forensic setting practicum. Similarly, over 75% of interns who are matched to a university counseling center have had previous practicum experience in a university counseling center, while 50% of interns or less at each of the other internship sites have had practicum experience at a university counseling center. Also, hospital/medical clinic practicum experience seems important (50% of matched interns have had this experience) for most of the internship sites.

Pairwise comparisons

Direct comparisons of averages between the types of sites may also be useful in identifying characteristics salient to different types of sites. Figure 1 provides a summary of significant findings from these analyses and also presents corresponding sparklines that illustrate the mean score on each variable of interest within each internship setting. Although the sample size associated with armed forces medical centers (n = 7) was too low for inclusion, all other settings were considered in these analyses. Given our exploratory goal, and to avoid overlooking potentially meaningful differences, the results are presented both with and without family-wise error correction. To highlight those results that remain significant following Bonferroni correction (α = .003), an asterisk follows the reported p value.

Figure 1. Results of pairwise comparisons on applicant characteristics by internship setting type. As noted in text, the armed forces medical center setting was not included in pairwise comparison due to insufficient sample size. Sparklines depict means on the variable of interest for each of the 14 internship setting types (left to right: 1 = Armed Forces Medical Center; 2 = Child/Adolescent psychiatry/pediatrics; 3 = CMHC; 4 = Consortium; 5 = medical school; 6 = prison or other correctional center; 7 = private general hospital; 8 = private outpatient clinic; 9 = psychiatric hospital; 10 = psychology department; 11 = school district; 12 = public hospital; 13 = UCC; 14 = VAMC). Note: due to unequal variances among groups being compared, visual representation in the sparklines may occasionally suggest a difference that is not confirmed by t-test.

Figure 1. Results of pairwise comparisons on applicant characteristics by internship setting type. As noted in text, the armed forces medical center setting was not included in pairwise comparison due to insufficient sample size. Sparklines depict means on the variable of interest for each of the 14 internship setting types (left to right: 1 = Armed Forces Medical Center; 2 = Child/Adolescent psychiatry/pediatrics; 3 = CMHC; 4 = Consortium; 5 = medical school; 6 = prison or other correctional center; 7 = private general hospital; 8 = private outpatient clinic; 9 = psychiatric hospital; 10 = psychology department; 11 = school district; 12 = public hospital; 13 = UCC; 14 = VAMC). Note: due to unequal variances among groups being compared, visual representation in the sparklines may occasionally suggest a difference that is not confirmed by t-test.

Quantitative GRE. Interns matched to state/county/public hospital settings had lower quantitative scores on the GRE compared to interns from the other sites, t(509) = 2.01, p = .045, d = 0.26.

Intervention/Assessment hours. While interns matched to sites that focused on children/adolescents were significantly higher than the average intern from other sites in their total intervention/assessment hours, t(695) = 2.25, p = .025, d = 0.19 interns matched to university counseling centers were significantly lower in their intervention/assessment hours, t(695) = 2.64, p = .008, d = 0.41.

Supervision hours. Consortium interns had significantly fewer supervision hours compared to other interns, t(690) = 2.26, p = .024, d = 0.29, and interns in correctional settings had significantly more intervention/assessment hours per hour of supervision compared to other interns, t(690) = 2.43, p = .016, d = 0.61.

Integrated reports. Consortium interns had written significantly more integrated reports compared to interns at other types of sites, t(87.1) = 2.59, p = .011, d = 0.35, and university counseling center interns had written a significantly fewer compared to other types of sites, t(101.28)[1] = 7.08, p < .001*, d = 0.70.

Peer-reviewed publications. Medical school interns had significantly more peer-reviewed publications, t(674) = 3.85, p < .001*, d = 0.30; VA interns also had significantly more peer-reviewed publications, t(674) = 2.15, p = .032, d = 0.18. In contrast, state/county/public hospital interns had significantly fewer peer-reviewed publications, t(137.89)a = 2.00, p = .047, d = 0.21, as did university counseling center interns, t(674) = 2.39, p = .017, d = 0.37, and community mental health center interns, t(704) = 2.03, p = .043, d = 0.25.

Presentations. Medical school interns also had more presentations, t(705) = 3.64, p < .001*, d = 0.29, compared to other interns, while university counseling center interns had fewer, t(77.93) = 4.42, p < .001*, d = 0.53. (Because Levene’s test for equality of variances was significant, the reported t-test statistic does not assume equal variances).

Personality characteristics. Interns matched to child/adolescent sites, t(378) = 2.27, p = .024, d = 0.27, private outpatient clinics, t(378) = 2.99, p = .003*, d = 0.70 and psychiatric hospitals, t(378) = 2.01, p = .045, d = 0.48, were all higher on agreeableness; community mental health clinic interns were higher on openness, t(377) = 2.51, p = .013, d = 0.40; and university counseling center interns were significantly lower in extraversion, t(379) = 2.06, p = .040, d = 0.36, compared to interns matched at other sites.


As with previous research, this study attempted to demystify the process of internship match by investigating student characteristics, experiences and other variables that are relevant during the internship match process (Callahan, Collins & Klonoff, 2010; Callahan, Hogan, Klonoff, & Collins, in press). Given the multitude of elements that are considered when students apply to internship sites, there is a great need to investigate factors that influence the match process so that students can understand the process and plan their experiences accordingly.

Within the fourteen different types of internship settings, overall there were more similarities than differences in terms of the experiences and characteristics of successfully matched applicants. For example, there were no significant Verbal GRE score differences and all but one setting demonstrated similar Quantitative GRE scores. Similarly, most settings fell into a narrow range (2.11 to 2.58) for the intervention/assessment to supervision hours ratio, with only applicants matched to the prison or correctional setting evidencing a significantly higher ratio (3.12).

Although successfully matched applicants appear to be largely similar across settings, more subtle distinctions do suggest differing emphases in terms of the characteristics of successful applicants. As just one example, students matched to child/adolescent psychiatry settings tended to have higher internship/assessment hours. The only noted exception to this trend of subtle distinctions were students matched to university counseling centers, who appeared to have a cluster of differences from students matched to other settings (lower number of intervention/assessment hours, integrated reports, peer-reviewed publications, and presentations). This trend speaks to the need for students to understand the setting to which they hope to apply and focus on specific avenues that are important for students who tend to be matched to those settings.

In particular, several settings appeared to require previous experience in that respective setting. For example, at least 75% of students matched to either a university counseling center or a forensic setting having had previous practicum experience in a similar setting. Not only are students able to see which types of previous experiences are common for students matched to a particular setting, but they can also see the types of experiences that are common for several settings. For 10 out of the 14 settings (armed forces, child/adolescent, consortium, medical school, private general hospital, private outpatient clinic, psychiatric hospital, psychology department, public hospital and VA medical center), at least 50% of students who were matched had previous experiences in a medical/hospital setting.

Unfortunately, data from students matched to armed forces medical centers were not able to be included in the analyses due to insufficient sample size. However, visual inspection of the data associated with this setting (column one of each sparkline) suggests the possibility of potentially meaningful differences. In light of the small sample size, caution in interpretation is merited, but future research would be useful to specifically test hypotheses derived from this study that applicants who successfully match to armed forces medical centers may differ from other successfully matched applicants in a number of ways (i.e., they may be higher in conscientiousness, lower in openness, attain lower quantitative GRE scores, amass more intervention and assessment hours, and/or attain fewer supervision hours).

Finally, differences in personality variables were found between applicants matched to varying settings. As with clinical/research experiences, there appear to be more similarities than differences, although several sites have slight differences. In particular, the personality traits of openness, extraversion and agreeableness were noted differences between students in some internship settings. For example, students matched to child/adolescent psychiatry, private outpatient clinics, and psychiatric hospitals tended to have higher ratings of agreeableness. However, this relationship is complicated; students should not assume a unidirectional influence, where students are rated highly by an internship site and thus more likely to be matched because of their personality. Instead, students with certain personality traits might be more likely to apply to and highly rate certain internship settings during the match process. In addition, there could be a bidirectional relationship, where students with a specific personality trait are drawn to a setting, and that setting is more likely to desire that student because of that personality trait.

Limitations and Future Directions

As previously noted, the sample sizes associated with some of the settings were small. Specifically, armed forces medical centers, prison/other correctional, psychology department, and school district settings all had fewer than 30 matched applicants in this sample. Another limitation relates to the recruitment methods. Participants were recruited through directors of clinical training belonging to the Council of University Directors of Clinical Psychology (CUDCP), resulting in most participants (96.5%) being clinical psychology trainees. Future research investigating characteristics of matched students by internship setting with a sample that includes not only CUDCP, but also the National Council of Schools and Programs of Professional Psychology (NCSPP) and the Council of Counseling Psychology Training Programs (CCPTP) is needed for a more complete understanding.

With respect to the variables considered, it would be helpful if intervention/assessment hours indicated the variation in clinical hours (individual therapy, group therapy, child therapy, family therapy, assessment hours, etc.) accrued by students. As just one example, applicants to child/adolescent internship settings might desire information regarding how much clinical experience with children/adolescents is typical for a successfully matched intern in that setting.

Also, additional inquiries that more specifically focus on identifying changes evidenced by applicants who are unmatched their first year but successfully matched the second year they apply could also be beneficial.

Implications for Students and Mentors

Given the current concern over internship shortages, students may be understandably concerned as they plan and apply for the internship match (Baker, McCutcheon & Keilin, 2007). Accessing the comprehensive table we have prepared and reviewing the results of this study might provide students and their mentors with a tool to help in the formation of realistic, though still subjective, appraisals regarding readiness for internship. By reviewing trends of students who were successfully matched to various types of internship sites, especially at the beginning of doctoral training, students may select clinical experiences and practicum settings that are appropriately aligned with settings likely to foster their training needs and future career goals. In particular, with the data on quartile cut-offs, students can more objectively determine their strengths and identify gaps in their preparation prior to making a decision about applying for the match, so that if a need for corrective action exists, it can be enacted proactively.

Mentors and program faculty could also review this information and use it to inform program expectations and opportunities. Although program requirements regarding minimum clinical hours might not necessarily be changed, mentors can have empirically based conversations with students about the type of characteristics associated with successful applicants being matched to various settings. Similar conversations with collaborating local practica may also be useful, so that students have access to appropriate externships that can prepare them well to advance to their desired internship settings.

Be the 1st to vote.
Cite This Article

Callahan, J. L., Swift, J. K., Hogan, L. R., Tompkins, K. A., Connor, D. R., & Klonoff, E. A. (2014). Seeking the “perfect” match: An empirical examination of applicant differences according to internship setting. Psychotherapy Bulletin, 41(1), 13-18.


Baker, J., McCutcheon, S., & Keilin, W. G. (2007). The internship supply-demand imbalance: The APPIC perspective. Training and Education in Professional Psychology, 1, 287-293. doi: 10.1037/1931-3918.1.4.287

Callahan, J. L., Collins Jr., F. L., & Klonoff, E. A. (2010). An examination of applicant characteristics of successfully matched interns: Is the glass half full or half empty and leaking miserably? Journal of Clinical Psychology, 66, 1-16. doi: 10.1002/jclp.20664

Callahan, J. L., Hogan, L. R., Klonoff, E. A., & Collins Jr., F. L. (in press). Predicting match outcomes: Science, practice, and personality. Training and Education in Professional Psychology.

Collins, F. L., Jr., Callahan, J. L., & Klonoff, E. A. (2007). A scientist-practitioner perspective of the internship match imbalance: The stairway to competence. Training and Education in Professional Psychology, 1, 267-275. doi: 10.1037/1931-3918.1.4.267

Costa, P. T., Jr., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO–PI–R) and NEO Five-Factor Inventory (NEO–FFI) professional manual. Odessa, FL: Psychological Assessment Resources, Inc.

Ginkel, R. W., Davis, S. E., & Michael, P. G. (2010). An examination of inclusion and exclusion criteria in the predoctoral internship selection process. Training and Education in Professional Psychology, 4, 213-218. doi: 10.1037/a0019453

Grus, C. L., McCutcheon, S. R., & Berry, S. L. (2011). Actions by professional psychology training and training groups to mitigate the internship imbalance. Special Section: Struggling to resolve the internship imbalance. Training and Education in Professional Psychology, 5, 193-201. doi: 10.1037/a0026101

Hatcher, R. L. (2011). The internship supply as a common-pool resource: A pathway to managing the imbalance problem. Training and Education in Professional Psychology, 5, 126-140. doi:10.1037/a0024658

Keilin, W. G., Baker, J., McCutcheon, S., & Peranson, E. (2007). A growing bottleneck: The internship supply-demand imbalance in 2007 and its impact on psychology training. Training and Education in Professional Psychology, 1, 229-237. doi: 10.1037/1931-3918.1.4.229

Lamb, D. H., Baker, J. M., Jennings, M. L., & Yarvis, E. (1982). Passages of an internship in professional psychology. Professional Psychology: Research and Practice, 13, 661-669. doi: 10.1037/0735-7028.13.5.661

Larkin, K. T. (2012). Models for reducing the internship imbalance. Training and Education in Professional Psychology, 6, 249-257. doi: 10.1037/a0030762

McCutcheon, S.R. (2011). The internship crisis: An uncommon urgency to build a common solution. Training and Education in Professional Psychology, 5, 144–148.

Stedman, J. M. (2007). What we know about predoctoral internship training: A 10-year update. Training and Education in Professional Psychology, 1, 74-88. doi: 10.1037/1931-3918.1.1.74

Kelley A.Tompkins, M.S., Dana R. Connor, M.S., & Elizabeth A. Klonoff, Ph.D., ABPP


Submit a Comment

Your email address will not be published. Required fields are marked *