Clinical Psychology Forum, 175,7-11. Selection for clinical psychology training: Is there evidence of any bias towards applicants from ethnic minorities? Scior,K., Gray,J.S., Halsey,R. & Roth,A.D. Authors Affiliations: Katrina Scior, James S. Gray, Anthony D. Roth University College London Robert Halsey Barnet, Enfield & Haringey Mental Health Trust Contact Address for Correspondence: Dr Katrina Scior, Sub-Department of Clinical Health Psychology, University College London, Gower Street, London WC1E 6BT e-mail: k.scior@ucl.ac.uk 1
Selection for clinical psychology training: Is there evidence of any bias towards applicants from ethnic minorities? This study looked at two cohorts of applicants to the UCL Training course to look for any evidence of inadvertent discrimination against ethnic minority applicants. There is general recognition that clinical psychology services should be accessible to all members of the community. One way of achieving this and promoting diversity is to ensure that as far as possible the profession reflects the make-up of the population. Concern has long been expressed about the under-representation of individuals from Black and Minority Ethnic (BME) groups amongst UK clinical psychologists (Davenhill et al., 1989; Bender & Richardson, 1990; Boyle et al., 1993). While BME individuals make up 9 % of the population of England (Office of National Statistics, 2003), they constitute only 7.4% of clinical psychologists, although their proportion is somewhat greater in younger age cohorts (Department of Health, 2004). The fact that BME individuals constitute 9.4% of applications to clinical psychology training courses, but only 6.4% of successful candidates (Turpin & Fensom, 2004) suggests that they experience some disadvantage when applying to clinical psychology courses. This study tracked the progress of two cohorts of applicants to the clinical psychology training course at UCL through the selection process, in order to identify predictors of success and to explore whether there was evidence of any systematic bias. METHOD A total of 1127 individuals applied to the course for entry in 2002 and 2003 (N=570 for 2002, N= 557 for 2003). Of these, ethnic monitoring information was available for all but 21 applicants who were excluded from all analyses based on ethnicity. Thus 2
the data set for all analyses based on ethnicity is N=1106 for the combined 2002 and 2003 data. All application forms were coded on 36 variables by two independent researchers, including: stage of selection process reached, academic background, employment history, references in the form of numerical ratings made by the academic and clinical referee (from 1 5, where 5 is the highest) and source of reference and ethnic background. Data on ethnicity was provided by the Clearing House. As the other category was over-inclusive, wherever possible, applicants in this category were assigned to one of three main ethnicity categories. Applicants of mixed parentage, for example, were placed in the category of the ethnic minority parent*. We also identified forms where the person had a clearly identifiable ethnic minority name, or which made reference to their BME status within the application form. The data were analysed using SPSS. *We acknowledge that ethnicity is based on self-definition but felt that this was the only way to produce meaningful results in view of the high numbers of applicants who described themselves as other. 3
RESULTS 1. Differences between applicants offered a place and unsuccessful applicants Applicants offered a training place were found to be different unsuccessful applicants on a number of variables. T-tests showed significant differences between the two groups in the number of A-levels (p=0.031) and A-level results (p<0.001). On average successful applicants had 3.3 A-levels and 14.1 total A-level points, compared to 3.0 A-levels and 11.5 total A-level points for unsuccessful applicants (where A=5, E=1). A further nine variables were found to be significant in distinguishing applicants offered a place from unsuccessful ones using chi-squared tests- these are presented in Table 1. Table 1 In addition to better A-level results, successful applicants were more likely to have obtained a first class degree, to have attended an old University, to have better ratings from their academic and clinical referees, and to have been employed as an Assistant Psychologist or as a Research Assistant. In order to establish whether the 11 variables identified as distinguishing successful from unsuccessful applicants are indeed good predictors of success a logistic regression analysis was conducted. Three variables emerged as significant predictors of success, independent of each other: the total number of A-level points and the scores given by the academic and clinical referees. At this point it is worth stressing that having an ethnic minority name was not an independent predictor of success or failure, but emerged as significant in the initial analysis due to its association with other variables. 4
2. Overall pattern of progress through the selection process by ethnicity Table 2 shows the proportions of individuals passing through key stages of the selection process in relation to their ethnic background for the combined 2002 and 2003 data set. Table 2 Chi-squared tests showed no significant differences between the proportions of different ethnic groups reaching each stage of the selection process, whether comparing white to BME applicants (χ²= 4.01, df=2, p =1) or white to Asian, black and other applicants (χ²= 3.74, df=6, p =1). As more detailed data was available for 2003 applicants, a survival analysis was conducted to examine in more detail at which stage applicants from different ethnic groups dropped out of the selection process. A marked difference was evident at the very first stage of the selection process, where the attrition rate was 49 % (N=34) for the 66 applicants from BME and other backgrounds, compared to only 29 % for white applicants. Closer examination showed that these 34 applicants did not meet the basic entry criteria set out by the course, namely possession of the Graduate Basis for Registration with the BPS (GBR); home or EU fee status; a 1 st or 2.1 degree (or 2.2 with clear evidence of higher academic potential); and at least some relevant clinical experience. 5
While it was reassuring to find no statistical differences in overall progression through the selection process for the combined 2002 and 2003 data, the differential attrition rates at different stages encountered in the survival analysis suggested that there are differences between applicants from different ethnic groups which merit further investigation. 3. Differences between White and BME applicants There were significant differences between applications from white and BME candidates on a number of indicators, mostly those concerned with academic achievement. T-tests showed that white applicants had significantly higher A-level points (12.0 vs 9.9 points, p<0.001). More white applicants had gained a 1 st class degree and fewer a 2.2 degree. The fact that 16.8 % of BME applicants applied with a 2.2 degree is important, given that a 2.1 is the usual minimum criterion for entry. This may explain the greater prevalence of Master s degrees among BME candidates, presumably attempting to demonstrate an appropriate level of academic ability. In addition, white applicants were significantly more likely to have attended old universities and were given higher scores by their clinical referees. Table 3 DISCUSSION The current study set out to examine whether there is any evidence of a bias operating against applicants from BME backgrounds to the UCL clinical psychology course. 6
The finding that the proportions applying were similar to those selected suggests no immediate evidence of a bias towards BME applicants. However, a more detailed survival analysis of 2003 applicants revealed that BME applicants were far more likely to be rejected at the earliest stage of the selection process due to a failure to meet the basic course entry criteria. This suggests that BME applicants overall may be less likely to access good careers guidance services or have less opportunity to seek advice from those inside the profession. This suggestion is further supported by the higher rates of 2.2 or 3rd class degrees and of Master s degrees amongst this group, yet their lack of success in applying for clinical training. Alongside referee ratings, A-Level points emerged as independent predictors of success. Notably, BME applicants overall tended to have lower A-Level points than their wide counterparts. Shiner & Modood (2002) note that nationally BME students have lower average A-level scores which are partly accounted for by their tendency to study fewer subjects. A strong emphasis on academic excellence is not surprising in view of the high demands of the Doctorate. While a lowering of academic entrance requirements might promote greater diversity amongst the trainee body, it might also have the undesired effect of increasing drop-out rates which might in turn affect different ethnic groups disproportionately. However, in the absence of other clear ways of distinguishing amongst the very large numbers of applicants with 2.1 degrees, selectors may overly rely on academic indicators, particularly A-Level results, which may be strongly affected by ethnicity and social class. Thus more detailed information about the quality of 2.1 degrees, such as degree transcripts which will be requested by the Clearing House in future, may help to reduce the risk of potential over-reliance on scholastic achievement (see also Turpin & Fensom, 2004). 7
With regard to the broader question whether the profession is showing at least some success in attracting more practitioners from BME backgrounds, the present findings suggest that at least for the UCL course (which does attract around a third of all applicants nationally) we may be seeing a very slow move in that direction. Bender & Richardson (1990) reported that of clinical psychologists who trained between 1978 and 1984, 5.8 % in metropolitan areas came from ethnic minorities. The present finding that 7.4% of trainees across two cohorts came from BME backgrounds could be seen as encouraging. However, it also indicates that the profession continues to be highly unrepresentative of the local population- around 29% of the greater London population are from BME backgrounds (ONS, 2003), although it needs stressing that applicants to the course live in all parts of the UK. The question of the appeal of psychology as a degree subject and clinical psychology as a career path amongst individuals from BME backgrounds has been raised in trying to explain their comparatively low rate amongst trainee clinical psychologists. Turpin & Fensom (2004) found that in terms of overall numbers psychology is more popular amongst BME undergraduates than more traditional healthcare professions such as medicine, dentistry and pharmacy, although BME students make up a higher proportion of students in these subjects. However, while 12% of psychology students are from BME backgrounds only 9.4% of applicants to clinical psychology courses are. It has been suggested that the level of awareness of clinical psychology as a possible career is lower amongst certain ethnic groups (Helm, 2002). In addition, our finding of a difference in the quality of applications suggests that perhaps the profession fails to attract BME graduates with strong academic records because it is 8
perceived as less prestigious than other careers, such as pharmacy, medicine, law and business studies (see Turpin & Fensom, 2004). This indicates a need for much more proactive promotion of clinical psychology amongst BME undergraduates and most likely well before (for fuller discussion of these issues see Helm, 2002). Finally, our finding that BME applicants were almost twice as likely to have dependants is striking and merits further investigation. There are a number of limitations to the current findings. They are based on only two cohorts of applicants and when comparing the 2002 and 2003 data there is marked variation in the course s success in attracting an ethnically diverse trainee body. Therefore a more longitudinal evaluation is likely to produce more reliable results. The main limitation though lies in the small number of successful BME applicants which render any analysis and interpretation tentative. Furthermore, the current study did not examine differences in the qualitative accounts provided by applicants on the application form. Undoubtedly these should also be expected to have a strong influence on selection decisions, although this impact is much more difficult to examine in a valid and reliable fashion than the quantitative information which formed the focus of the current study. Word Count: 2115 words including references 9
References Bender, M.P. & Richardson, A. (1990). The Ethnic Composition of Clinical Psychology in Britain. The Psychologist, 6, 250-252. Boyle, M., Baker, M., Bennett, E. & Charman, T. (1993). Selection for Clinical Psychology Courses: A Comparison of Applicants from Ethnic Minority and Majority Groups to the University of East London. Clinical Psychology Forum, 56, 9-13. Davenhill, R., Hunt, H., Pillay, H.M., Harris, A. & Klein, Y. (1989). Training and Selection Issues in Clinical Psychology for Black and Minority Ethnic Groups from an Equal Opportunities Perspective. Clinical Psychology Forum, 19, 13-17. Department of Health (2004). NHS Hospital and Community Health Services Non- Medical Workforce Census England: 30 September 2003. London: Department of Health. Helm, M. (2002). Promoting Diversity in Clinical Psychology, Unpublished Document. Canterbury Christ Church University College. Office for National Statistics (2003). The 2001 National Census: Ethnicity and Religion. London: Office for National Statistics. Shiner, M. & Modood, T. (2002). Help or Hindrance? Higher Education and the Route to Ethnic Equality. British Journal of Sociology of Education, 23(2), 209-231. Turpin, G. & Fensom, P.(2004). Widening Access within Undergraduate Psychology Education and its Implications for Professional Psychology: Gender, Disability and Ethnic Diversity. Leicester: British Psychological Society. 10
Table 1: Differences between Applicants Offered a Place and Unsuccessful Applicants Variable Offered a Place Unsuccessful χ² value p School Type Public 35.9%* 19.5 %* 14.10 0.001 State 57.7% 64.6 % Degree Class 1st 35.9 % 15.6% 24.14 <0.001 Degree from Old University 2.1 61.5 % 73.3% 92% 62.6% 24.14 <0.001 Past Employment as Assistant psychologist Past Employment as Research Assistant Average Score given by Academic Referee Average Score given by Clinical Referee Clinical Psychologist as Clinical reference 78.2 % 51.7 % 20.39 <0.001 53.8 % 39.5 % 0.72 0.013 4.5 4.38 49.25 <0.001 4.38 4.0 13.59 0.035 69.6 % 49.5 % 11.96 0.003 Ethnic Minority Name 1.3 % 8.2 % 4.98 0.026 *percentages do not add to 100% due to applicants educated outside the UK educational system 11
Table 2: Selection Stage reached by Ethnicity Applied (N=1106) Interviewed (N= 213) Offered Place* (N=81) White 962 (87.0%) 193 (90.6 %) 75 (92.6 %) BME & Other** 144 (13.0%) 20 (9.4%) 6 (7.4%) Asian*** 76 (6.9 %) 11 (5.2 %) 2 (2.5%) Black **** 45 (4.1 %) 6 (2.8 %) 3 (3.7%) Other 23 (2.1 %) 3 (1.4%) 1 (1.2 %) *Includes anyone offered a place, regardless whether subsequently accepted place. **Summary category for all Asian, Black and other applicants. ***This category includes all applicants of Indian, Bangladeshi, Pakistani or of Mixed White & South Asian Parentage ****This category includes all applicants of Black British, Black African, Black Afro-Caribbean or of Mixed White & Black Parentage 12
Table 3: Variables that Distinguished White and BME Applicants Variable White Ethnic χ² value p minority* Degree Class 1 st 18.4% 8.4 % 14.12 0.007 2.1 72.1 % 74.8 % 2.2 7.5 % 15.9 % Master s degree 19.2 33.6 13.99 <0.001 Degree from Old 73.2 % 49.1 % 28.20 <0.001 University Average Score given 4.11 3.76 27.30 <0.001 by Clinical Referee Dependants 5.8 % 10.3 % 3.99 0.046 *This category includes all applicants whose ethnic origin falls into the Black, Asian and other categories 13