Student selection process effectiveness : Correlations between task performance and undergraduate success

INTRODUCTION: This article is a case study of student selection process effectiveness in an undergraduate social work degree at Unitec in Auckland, Aotearoa New Zealand. Addressing an internationally under-researched topic, it examines whether admitted candidates’ performance on selection day tasks correlate with their success in the programme. METHODS: Applicant selection data were analysed for 2012 and 2013 cohorts (N = 196). Student success is measured in relation to outcomes across eight courses, considering both course completions and grades received. Correlational analyses were performed to address the research aim. The project also examined whether existing demographic data correlate with success, as these data represent potentially confounding variables. FINDINGS: Performance on selection activities conducted as part of application to this degree does not correlate with course completions, and correlates only moderately with higher grades received for courses. Findings also show that students who are wage-earning or self-employed on admission, studying full time, and/or admitted well before their studies begin are more likely to succeed. No other demographic factors are correlated with student success. CONCLUSIONS: This study supports existing literature documenting the ineffectiveness of selection criteria for social work programmes, and also addresses a gap in scholarship by examining the value of specific selection tools and measures. Factors that do correlate with student success suggest that being well-resourced as a student is paramount. Supporting students to succeed during the course of their study may be of higher value than attempting to select “suitable” students in the first place.


Introduction
This article describes a case study of selection task effectiveness for an undergraduate social work degree at Unitec in Auckland, Aotearoa New Zealand.It addresses gaps in the literature identified in a 2016 international literature review (Hughes, McNabb, Ashley, McKechnie, & Gremillion, 2016) conducted by a team of co-authors which includes four of the authors involved in the current study.This literature review sought to determine whether social work programmes are using effective tools to select students who will subsequently succeed, or whether there is little or no predictive reliability in the selection process.Key findings include the following: …there is no simple formula for selecting students who will go on to complete their social work programme, and… many of the tools used during the selection process are unreliable predictors of eventual student success.Findings also indicate that existing research is limited; further research is needed, particularly on the value of specific selection tools and measures.(Hughes et al., 2016, p. 94) In addition, Hughes at al. (2016) found that student success in social work programmes internationally may depend on a complex set of factors including professional socialisation.Professional socialisation is regarded as the assimilation of knowledge, values and skills pertaining to a profession into an individual's identity during their course of study (Cornelissen & van Wyk, 2007).These findings, coupled with the fact that certain selection measures "such as GPAs from previous academic study have been consistently shown … to exclude suitable students from programmes" (Hughes et al., 2016, p. 104), provide the motivation for the present project.
The authors of this case study are staff members in the social work programme at Unitec who have regularly taken part in selection processes since the current version was instituted in 2011.We were initially prompted to collect data on and review our selection processes because, while they are time-consuming and resource-heavy, we have had up until now no reliable information about the efficacy of the selection tools currently utilised.An external driver reinforcing the need for review of applicant selection is the expectation that many graduates will gain registration with the Aotearoa New Zealand (ANZ) registration body: the Social Workers Registration Board (SWRB).Internal drivers to review applicant selection include: meeting Unitec's goals for success and retention of target groups including Máori and Pacific Island students, as well as students under the age of 25, and the long-standing and ongoing desire to select students who are ultimately suitable for practice.This latter point raises a number of questions and tensions, addressed in the following section.
The regulation of social work education in Aotearoa New Zealand Apaitia-Vague, Pitt, and Younger (2011) point to a central contradiction surrounding efforts to admit students to social work programmes who will succeed and be fit to practise: a need for "accountability to clients to ensure that workers are safe to practice, and social work values such as inclusion, social justice, the right to education, and a belief in the power of transformative change" (Hughes et al., 2016, p. 103; see also Dillon, 2007;Pelech, Stalker, Regehr & Jacobs, 1999;Sowbel, 2012).How can admission criteria help ensure student success, without screening out potential students who may succeed in programmes when given the proper support and encouragement?While a detailed consideration of the latter is beyond the scope of the present study, this question points to the importance of critically examining selection processes to ensure that they make a difference in relation to the suitability of students selected, keeping in mind a burden on programmes to scaffold learning for cohorts that include a range of abilities, backgrounds and constraints.
Writing in 2003, Nash (2003) noted the tensions and, at times, competing priorities amongst social work as a profession, the state, social service providers, and social work educators.Arguably these tensions have increased over time, considering market-driven pressures encouraging higher student numbers as well as increasing interest in social work professionalisation, both of which highlight the need for wellqualified graduates.Notably, social work as a profession was recently removed from the Long-Term Skills Shortage list of the New Zealand government.The number of graduates from social work programmes in Aotearoa New Zealand has increased in recent years with the latest figures indicating a corresponding increase in the number of applications for SWRB registration: from 542 in 2013/2014, to 666 in 2014/2015, to 737 in 2015/2016(SWRB, 2015-2016)).Between February and April 2017, further work was conducted; a summary of main points notes the recommendations to pursue mandatory registration and protection of title, and to put in place an efficient regulatory regime.As well, the summary mentions additional support to improve the quality and professionalism of social work services as a requirement. 1  After reviewing the options outlined in the Select Committee's report to government and the government's response documents, the Minister for Social Development determined a preferred option that would extend the coverage of the regulatory regime, but not to the extent suggested by the SWRB.Under this option, only registered social workers would be able to use the title "social worker."This tightening of title use would mean that approximately two thousand currently employed, unregistered, social workers would need to become registered or leave their employment.Social workers would need to be registered as opposed to registrable.The Select Committee proposed that these changes would result in an economy of scale thereby reducing costs for registration (Ministry of Social Development, 2017).The committee's discussion of licensing the roles and tasks of social work resulted in a recommendation that task-based licensing not be supported, as it would not only limit the work currently done by social workers but also shut out other practitioners who may need to carry out some of the roles and tasks of social work.

ORIGINAL ARTICLE
According to the Ministry of Social Development (2017), preparation for mandatory registration has raised concerns about the quality of training being provided in social work education.Additional accountability and oversight of educators has been recommended as part of the legislative change to the SWR Act.In a mandatory environment, employers will be seeking only those employees who are registered at the completion of their tertiary study.This fact, alongside protection of title, has major implications for student selection as there are ethical tensions surrounding the admission of students who may not meet the criteria for registration.If students are ineligible for social work registration, they will not be eligible for employment as social workers.
One of the reasons for social work graduates not gaining registration is failure to meet the requirements of current fit and proper criteria, a somewhat subjective determination based on several factors including a history of previous convictions (SWRB, 2015).However, to improve the transparency of the fit and proper criteria, Cabinet have: propose [d] that the SWR Act adopt a similar approach to the HPCA Act and provide a more comprehensive list of the factors the Board can consider in the exercise of its discretion (including an assessment of any mental or physical

QUANTITATIVE RESEARCH
health issues and disciplinary history).(Office of the Minister for Social Development, 2017, p. 11) It remains to be seen whether the factors that can be considered will allow applicants to gain access to social work training with some certainty of their eligibility to gain registration, or will further exclude applicants with potential who will therefore be ineligible for registration.
Contemporary social work education remains a contested site in the wider political context and within the profession itself (Beddoe, 2014).The SWRB reviewed its expectations for social work programmes over 2012-2013 and created a new minimum level of a four-year undergraduate programme for all tertiary providers (SWRB, 2013).This requirement was fought by Unitec and other tertiary providers on the grounds that many three-year degrees were robust, and that economically disadvantaged students should not have to pay for a further year's study when other options such as employer-funded internships and post-graduate study were possible (McNabb, 2014).The latter concern is heightened by the recent introduction of limitations to student funding for loans and allowances (Ministry of Social Development, 2013).These tensions surrounding accessibility of social work education and eligibility for registration highlight the importance of both appropriate and fair student selection criteria and adequate study support systems and processes.2013).However, participation rates since 2012 appear to have decreased slightly (Ministry of Education, 2017) -although it is difficult to compare reports using a range of measures.

Indigenous populations
The Pasifika Education Plan has seen a similar overall increase in Pasifika enrolments from 14% in 2011 to 15% in 2012 (Horrocks, Ballantyne, Silao, Manueli, & Fairbrother, 2012).However, updated information is lacking, as is information on completion rates for, as well as selection of, these students.It is noteworthy that there are significant pressures on Pasifika students including high parental expectations and aspirations (Madjar, McKinley, Jensen, & Van Der Merwe, 2009), which Greenwood and Te Aika (2009) note exist for Máori students as well.A report by the Ministry of Education (2009) found that family expectations and commitments more generally, including child care responsibilities, disproportionally affect Pasifika students (see also Madjar et al., 2009), with implications for time and energy to focus on tertiary studies.
As noted in Hughes et al. (2016, p. 102)and the point is also applicable to Pasifika students (see Horrocks et al., 2012):

ORIGINAL ARTICLE
Existing literature on Máori student experiences and success focuses on recruitment and retention, not selection processes.Such a focus is arguably appropriate in the current environment, in which there are efforts to create programmes that are successful for Máori, rather than fit Máori students into existing programmes (see Beddoe, 2007).It remains to be seen whether certain selection criteria are relevant specifically for the admission of Máori students who will be successful in social work programmes.(Hughes et al. 2016, p. 102) The Bachelor of Social Practice degree at Unitec

Background and student success and retention
The Bachelor of Social Practice (BSP) degree at Unitec was established in 1992.Until 2015 students enrolled in a three-year programme (if undertaken full time) providing a social work registrable qualification.At the start of 2014, when data entry for this study began, the BSP programme included 357 Effective Full Time Students (EFTS) and 22 Full Time Equivalent (FTE) staff (who also taught across two small postgraduate programmes provided by the Social Practice Department).
The following information includes figures from 2014, when students whose data are included in this research were still studying on the degree.The BSP programme consists of a range of assessment methods that are not exam-based, and is made up of mainly full-time students -90% of its EFTS.The programme enjoys 90% success (passed course) and 93% retention (re-enrolled in the following semester) rates (Social Practice, 2015).As noted above, two key target groups are Máori and Pacific students.Máori students in the programme have an 86% success rate and a 90% retention rate.Pacific students have an 86% success rate and a 91% retention rate.Graduate employment figures overall exceed 80% in the year following graduation, which compares favourably to other professional degree programmes across Unitec.

Selection day activities and scoring
Selection day (SD) activities for admission into the BSP degree include an individual interview, a team activity and a writing task.The activities chosen for SD reflect the nature of assessment used in the degree; written assessment, group presentations, and individual assessment in work-based activities.These activities are described in more detail below.Applicants are given a score out of five on each activity, with "one" indicating the lowest possible score and "five" the highest, so that the total maximum score an applicant can receive is 15.
Once the applicants have left on SD, staff hold a review meeting where scores for all activities are recorded and archived on the SD spreadsheet.Most selection decisions are made at that meeting, and if follow-up is required, the Programme Leader gains clarity from staff about the areas of concern.Applicants are placed in one of four initial categories: accept, require more information, waitlist, and decline.
Except in exceptional circumstances, applicants with a combined score of eight or below are not admitted.For applicants who score above eight, additional criteria are taken into consideration as well: particularly the nature and date of any convictions (discussed below), and reference feedback.

Interviews
Individual, 10-15 minute interviews are carried out in pairs including a Social Practice staff member and, when possible, a community volunteer affiliated with a fieldwork placement agency.The score for the interview is based on candidate responses to the following questions: • Can you tell me why you are interested in studying in the Social Practice programme?• What do you understand about social justice?

QUANTITATIVE RESEARCH
• Why do you think we have a commitment to biculturalism at Unitec? • Do you think it is important for social workers to know about the Treaty of Waitangi?Why?
Notes are taken and, after the interview, the two interviewers negotiate a mark out of five for the candidate's answers.If negotiation fails, the candidate is given the average of the two marks.

Team activity
The team activity is a group discussion of a provided scenario in which several of the characters take actions resulting in a man being murdered.This activity is used to assess the five applicant attributes listed below.
Applicants are asked in groups of six to rate the characters in the scenario in terms of culpability or innocence.They must present as a team and are given marks individually on their participation.Staff members observe the group discussions and score individuals for each of the following criteria: • maturity, motivation and resilience; • vulnerability awareness; • interpersonal understanding; open and non-judgmental attitudes; • self confidence; • analytical ability, conceptual thinking, task focus.
Although applicants tend to think they are being marked on contribution content, they are actually being scored on their involvement in, and contribution to, the discussion.Staff members negotiate or average their total scores per applicant, then divide the total by five to produce a final mark.

Writing task
The writing task is conducted in a computer lab under examination conditions.Candidates are given a temporary login to the institute's shared drives and are asked to create a word document.They are given a hard copy of a written exercise and 15 minutes to read it carefully.Marking criteria are explained to the candidates, who are then given 20 minutes to write their response to the question, 'How might you respond to a social issue' described in the handout.At the end of the 20 minutes, candidates are asked to "save as" their document and to log off.
The writing task is scored according to the following criteria: structure, logic and academic presentation.Two staff members mark the written work and negotiate or average a final agreed mark.Table 1 shows that the team activity has the highest scores and the least variability while the writing task has the lowest scores and the most variability.

Measures of success in the programme
Success in the programme is measured in terms of course completions and grades received across eight compulsory courses in Year One of the BSP degree (totalling 120 credits, or one year's full-time study).
Overall student GPA across these eight courses is an additional data point per student.People-Soft, the student management program at Unitec, was used to access academic transcripts.
All courses that students completed were marked with scores ranging from A+ to E (fail grade).Each letter grade was assigned a corresponding numerical mark at the midpoint of the numerical range that a given letter grade represents.A non-completion mark of "0" was assigned when a course QUANTITATIVE RESEARCH

ORIGINAL ARTICLE
grade was DNC (did not complete), W (withdrawn), or blank (did not enrol in a course, though admitted to the programme).Cross credits (0.05% of the total) were awarded a B grade.

Other measures
Enrolled students provide additional information which we included in our analyses, as these data represent potentially confounding variables.We examine whether the following are correlated with student success: age, gender, culture, conviction history, mental health or disability issues, date of intake, full-or part-time student status, and work status immediately prior to enrolment.The latter is comprised of three groups of students: from school or tertiary education; wage-earner/self-employed; or unemployed, not employed, or beneficiary.Gender are categorised as "man" or "woman": Unitec data do not include transgender or "other" identifications.Culture is comprised of four categories: Máori, Pacific Island, Pákehá/ European, and African/Asian.African and Asian students comprise a single category for analysis purposes because of relatively low numbers within each of these cultural groupings.November intakes (for those starting the degree in February of the following year) -our largest intakes -are separated out from the "top-up" intake dates taking place during the year of enrolment, and the latter are combined into one group.
Conviction histories are taken from a required applicant report and confirmed by a police check procedure.Convictions range from a few serious charges such as assault, theft and fraud through to trespass, drunk in charge of a vehicle and minor driving offences.
Candidates are asked to declare any mental health issues, which include depression, anxiety, and the effects of trauma.These are not necessarily diagnoses, and indeed include at times significant signs of distress, confusion, or disconnection noted by Social Practice staff members during SD tasks, and reported during the SD review meeting.Note that conviction histories and mental health and disability issues were recorded for students admitted in 2012 (N = 150), and were not recorded for students admitted in 2011 (46 students).

Methodology Research aim
To determine whether performance on selection day (SD) activities is correlated with student success in the Bachelor of Social Practice degree.
As noted above, this project also examined whether certain demographic variables correlate with success, as these variables are potentially confounding ones in relation to our research aim.

Sample
Applicant selection and demographic data were collected for three cohorts entering the programme in 2012 (November 2011, January/February 2012 and mid-year 2012

QUANTITATIVE RESEARCH
intakes) and for one cohort entering in 2013 (November 2012 intake).All students admitted during this time, including those who withdrew from the programme or did not complete particular courses, were included as participants (N = 196).All participants, including those enrolled part-time, had ample opportunity to complete all eight courses included in the analysis before data entry for this project was complete.

Ethical considerations
Ethics approval is not required at Unitec for projects such as this one, entailing analysis of de-identified programme data for a secondary (research) purpose.On the advice of Unitec's Research Ethics Committee chair, we consulted Unitec's legal counsel to ensure that publication of our findings would protect student privacy.This latter was confirmed: no students are identifiable, even to themselves (no findings relate to sub-groups of students smaller than 20), and legal constraints surrounding the secondary use of data do not apply in this case, as no identifiable personal information is reported.
A rigorous process was followed in the handling of the project data to preserve participants' confidentiality, as per standard procedure for any evaluative analysis within Social Practice.First, student names were removed from all pieces of data.
In her role at the time as Programme Leader, the first author (Hughes) had access to student grades in the normal course of her work, so she undertook the task of removing student names from transcripts.Next, in the process of entering the data into a spreadsheet, student ID numbers only were used by two research team members strictly for the purpose of linking together the different pieces of data per (un-named) student.Finally, prior to data analysis, student ID numbers were permanently removed from the collated data set, rendering the data fully un-identifiable.

Data analysis and fi ndings
Non-completions and student success In order to isolate the effects on our findings of the large group (79 students) who did not complete one or more courses, we examine correlations between selection day performance and grades received both with this group included and excluded from the analyses performed.Amongst the group who completed all eight courses are a small number of students (nine) who had failing GPAs after submitting all required work for all eight courses.Note that non-completion and failure rates here do not equate with the attrition rate in the programme: a number of students who did not complete one or more courses, or failed one or more courses after submitting work for them, are eligible to re-enrol in these courses in the future.Most students who did not complete one or more courses did complete several courses; only nine students who enrolled in all eight compulsory courses did not complete any of them.
The justification for isolating the noncompletion group is that nearly 90% of failing GPAs in our analysis are due to non-completion.This fact accounts for much of the difference between semester one and semester two grade averages shown in Figure 1 (a 12.4-point difference, p < 0.00 vs. a 4.7-point difference, p < 0.00 when non-completions are excluded).

Success in relation to demographic data
In the demographic data we collected there are many variables that could influence the grades that students achieve.To eliminate their impacts on the key question of whether intake assessment tasks are correlated with success, we first examine the effects of these demographic variables, both when we include and exclude non-completion cases.Note.* = p < 0.05 (two-tailed), i.e., this is a significant correlation, either positive or negative, with the chance of error being less than 5%.

ORIGINAL ARTICLE
intakes in five courses and overall -mean grade = 56.0 vs. 65.4(Nov 2011) and 68.0 (Nov 2012).These three effects -work status at intake, studying full-time and date of intake -are independent of each other as none of the inter-correlations between them are significant.
When non-completions are removed from the data (see Table 5), work status at intake and being full time no longer have a significant impact on grades, with only two significant correlations.However, intake date still impacts on grades for the same five courses as shown in Table 4, and overall there is still a significant disadvantage associated with enrolling in top-up intakes (mean grade = 77.4vs 80.6, p < 0.01).These findings indicate that being employed at intake and being a full-time student are associated with success only when we consider the entire sample (including noncompletions).Unlike the effect of intake in November, they are not associated with degree of success for students who completed all eight courses.
There are two remaining demographic features to consider: any record of convictions and mental health issues.While these factors have moderate-to-strong correlations with a few other demographic features, neither correlate significantly with student success.
We are now in a position to examine the effect on student success of selection day (SD) scores.Table 6 shows that total SD scores and the writing task influence the degree of success for students who complete Note.* = p < 0.05, #= p < 0.01; ^= p < 0.005 (two-tailed), i.e., this is a significant correlation either positive or negative with the chance of error being less than 5%, 1% and 0.5%, respectively.ote.* = p < 0.05, # = p < 0.01; ^ = p < 0.005 (one-tailed), i.e., this is a significant a priori positive correlation with the chance of error being less than 5%, 1% and 0.5%, respectively.Negative correlations are ignored as they were not predicted.all courses, but do not predict success when non-completions are included in the analysis.

QUANTITATIVE RESEARCH
In the latter case, there is only one significant correlation out of 44.When non-completions are excluded, however, the writing task and the total intake scores have moderate to strong correlations across a number of courses, and across semester mean grades and overall mean grades.In addition, all the SD tasks positively correlate with the grades for at least one course.
Recall that, unlike any other demographic factor, date of intake also has a significant influence on grades once non-completions are excluded.In addition, there are four significant correlations between intake date and SD scores for students who complete all courses, as shown in

ORIGINAL ARTICLE
course grades are significant, most at p < 0.01 or lesser.Therefore, in order to clearly demonstrate the correlation between SD scores and grades (when all courses are completed), we need to remove the effect of intake date.The final step in our analysis is to determine the effect size and practical consequences of significant SD score correlations.Lenhard and Lenhard (2016) categorise the effect size of correlations as follows: small, r = 0.10-0.22;moderate, r = 0.23-0.35;and large, r = 0.36-0.45.Of the significant correlations in Table 8, 25% are small, 69% are moderate (including the mean Year One/ total score correlation) and one (6%) is large.
The box and whisker plot in Figure 2 shows the relationship between mean GPA scores and total intake scores, demonstrating the variability in the data which characterises a moderate effect size.Low total intake scores (9.5-10.5)average in the B to B+ range, mid-range scores (11.5-13.0)average B+ and ote.* = p < 0.05, # = p < 0.01; ^ = p < 0.005 (one-tailed), i.e., this is a significant a priori positive correlation with the chance of error being less than 5%, 1% and 0.5%, respectively.Negative correlations are ignored as they were not predicted.

QUANTITATIVE RESEARCH
the highest scores (13.5-15.0)average A-.Thus, while the correlations demonstrate an effect of moderate power for SD scores, the consequences in terms of GPA operate within a narrow (5 to 10-point) range and thus have little practical effect.

Limitations of the study
A limitation of the present study is the nonstandardised nature of the scoring process for SD tasks.Although a moderation process is in place -and final decisions are made at a team review meeting, which mitigates assessor subjectivity (Watson, 2002) -the criteria for marking are inherently subjective and, inevitably, different markers are used for different selection days.Also, staff have a bias towards scoring high, particularly on the team activity and interview tasks, assigning low scores mostly to applicants who perform very poorly.The median scores show that over 50% of applicants received the maximum mark in the team exercise, and over 4.5 in the interview.However, in the writing task and total scores, the median is well below the maximum, indicating a reasonable degree of discrimination in scoring.
Inconsistent scoring processes are most apparent when we examine mean writing scores (see Table 7) across intake dates in relation to student grades achieved.The November 2011 and top-up intake groups evidence the highest writing scores, yet the November 2012 group achieved the highest GPAs (79.8,77.4,and 81.5,respectively).

Conclusions
The aim of this research was to determine whether selection tasks for the Bachelor of Social Practice degree at Unitec in Auckland, Aotearoa New Zealand are correlated with student success in the programme.The findings show that, while the total selection score achieved and performance on the written task, in particular, do influence how well students will perform in the programme when they complete all (compulsory) first-year courses, they do so only to a moderate degree and they do not predict whether students will fail to complete one or more courses.The latter finding is particularly striking given our relatively large (N = 196) and diverse sample size, and the large number of students who did not complete one or more courses (79).Notably, the vast majority of students who fail do not complete.
We have shown that there are inconsistencies in the application of selection measures across BSP degree intakes, and some limitations in the scoring of the interview and team activity in particular.Given that the latter are shown to have little or no correlation with course grades, respectively, these limitations may well have affected our results.That said, these tools are inherently open to bias and difficult to improve in this respect (Ross, 2010;Watson, 2002), though admittedly research on this topic is very sparse (Hughes et al., 2016).
Overall, this case study confirms previous research suggesting that selection tools generally speaking are flawed and unreliable predictors of student success in social work programmes (Hughes et al., 2016;Poole et al., 2012;Ryan, Cleak, & McCormack, 2006).This study also fills a gap in the literature by exploring the value of specific selection tools and measures (Hughes et al., 2016) in this case in relationship to students' results in eight core courses.While the written task and its selection suite as a whole do appear to have some (limited) discriminating value, the interview and especially the team activity are clearly ineffective measures of how well students will do in the BSP programme, although they may be effective screening devices for excluding unsuitable applicants (a topic which the current study does not accommodate).As Sowbel (2012) notes, a gatekeeping role on admission is an ethical obligation to protect future clients.It also embodies an obligation to students who may not be able to engage a fieldwork placement or secure employment after graduating (due to certain criminal convictions, for example).

ORIGINAL ARTICLE
However, this gatekeeping function can be part of a less resource-intensive selection process, potentially (pending future research) with selection tools that are more effective than the ones currently in use in the BSP programme.Possible alternative tools include psychometric testing and the biographical questionnaire, both of which require further research to determine their effectiveness.The biographical questionnaire has been shown to have content validity, but its predictive validity in terms of student success is as yet unclear (Ross, 2010).Dillon (2007) suggests that psychometric testing allows for a relatively reliable identification of candidate traits that are more likely to predict success than a range of other selection measures.Harder measures such as GPAs from previous study are not recommended, because they "have been consistently shown … to be unreliable predictors of student success" (Hughes et al., 2016, p. 104).In addition, although more objective selection measures are easily standardised, the way the results are used can be quite variable (Adam, Dowell, & Greatrix, 1999).
As noted in the introduction to this article, harder measures have also been consistently shown to exclude suitable students from programmes (Hughes et al., 2016), a point which highlights a key issue raised in this article.It is important that any selection tool is used in such a way that guards against the exclusion of students who could be supported to do well in their studies, and to become appropriately fit and proper social work practitioners in the process.This issue is particularly important for a profession that espouses social justice, inclusion, and the possibility of transformative change (see Apaitia-Vague et al., 2011).
To eliminate their impacts on the question of whether selection tasks are correlated with student success, this study also examined whether a range of demographic factors are correlated with success.Three such factors were shown to do so, independently of each other.These are studying full time, wage-earning or self-employed status on admission, and/or admission during a November intake, the latter allowing a large time gap between admission and starting the BSP programme.Notably, employment status and full-time student status have no impact on degree of success for those who complete all courses; unlike performance on selection day tasks, they are correlated with the key success goal of course completion (and with this goal only).
Our findings speak to the value of supporting students to succeed during the course of their study, and raise caution around gatekeeping admission processes too strongly.Further research is needed to explore the significance of the demographic factors shown in the current project to correlate with success -factors which are beyond the scope of our research aim and are not normally considered in relation to student success and retention.We speculate that they are linked to being relatively wellresourced for study: via income support, contextual (e.g., family) support to study full time, and/or other forms of support in one's life allowing for advanced planning.Future research might investigate, for example, the importance of resource use during students' course of study, and of pastoral as well as life skills support to enable student success.This case study suggests that we should be especially mindful of applicants who perform poorly on selection day and might not be predicted to succeed in the programme, and yet might well succeed if provided with support along the way.
A focus on supporting students to succeed once they are admitted fits well with appropriate strategies surrounding the success and retention of Máori and Pacific students, two key priority groups within the New Zealand Government's Tertiary Education Strategy.As noted earlier in this paper, existing literature on Máori and Pacific student experiences and success in social work programmes focuses not on selection processes, but on student retention (as well as recruitment).Given new funding

QUANTITATIVE RESEARCH
constraints for student loans coupled with pressure to increase enrolments, such a focus is also likely to benefit an increasing number of students within a diverse and inclusive student body.

Figure 1
Figure 1 shows all the GPA (0-100) data broken down by course.Grade averages range from 75.4 for Whanaungatanga, Gender & Social Practice to 54.7 for Discourses of Social Practice, with the overall GPA being 63.1.

Figure 2 .
Figure 2. Mean GPA score for the eight Year 1 courses as a function of the total selection day score, excluding dropouts.

Table 1 :
Selection Day Scores

Table 2 :
Correlation Coefficients for All Students

Table 3 :
Correlation Coefficients for all Students Excluding Non-Completion

Correlation coefficients (r) of age, gender and culture with course grades, excluding those who did not complete courses (N = 117)
* = p < 0.02 (two-tailed), i.e., this is a significant correlation either positive or negative with the chance of error being less than 2%.

Table 4 :
Work Status and Course Grades

Correlation coefficients (r) of work status at intake, being full-time and date of intake with course grades for all students, including those who dropped out of courses (N = 196). Work status at intake full time (N = 147; 75%) Date of intake Courses From school or tertiary ed. (N = 63, 32%) Wage earner/self- employed (N = 80; 41%) Unemployed, not employed or beneficiary (N = 53; 27%)
* = p < 0.05, #= p < 0.01; ^= p < 0.005 (two-tailed), i.e., this is a significant correlation either positive or negative with the chance of error being less than 5%, 1% and 0.5%, respectively.

Table 5 :
Work Status and Course Grades-Non-Completions Removed

Table 6 :
Selection Day Scores and Student Success

Table 7 :
Means of SD Scores

Means of selection day scores and correlations with intake dates, excluding those who did not complete courses. Date of intake team activity interview Writing task total score
Table 7 below.

Table 8 :
Selection Day Scores Without Intake Date Effect