Universities can measure career services effectiveness only by separating four questions that are usually merged: what the service offers, which students it actually reaches, how they use it, and whether graduates are demonstrably better off as a result. Counts of appointments, events and platform logins answer the first three questions at best. The fourth requires outcome data linked to service use and a credible comparison group, because students who choose to use career services differ from those who do not. The research evidence is strongest for short-term gains in career decision-making and thinnest for employment and earnings.
Why activity counts are not evidence of benefit
Annual reports from career services often lead with volume: appointments held, workshops delivered, employers on campus, vacancies posted. These figures describe effort and capacity. They do not show that the right students were reached, that the help was good, or that anyone's career turned out differently. A busy service can be reaching mainly the students who would have done well anyway, while those with the least access to professional networks never arrive.
This article uses four terms with distinct meanings:
- Availability: the service formally exists, is staffed and is documented (for example, individual advising, workshops, employer events, placement or vacancy support, a digital platform).
- Accessibility and reach: students know about it and can use it at times, in formats and in languages that fit their circumstances, including part-time, commuting, online and international students.
- Utilization: who actually uses which service, how often, and at what point in their studies.
- Demonstrated benefit: evidence that users are better off on a defined outcome than they would otherwise have been.
Graduate outcomes themselves (employment rates, job quality, further study) are a separate topic, covered in How Should Universities Measure Graduate Employability? This article concerns the narrower link between what a career service does and what changes for students.
What existing survey instruments capture
Large public surveys already touch on career services, but they measure different things, and none measures benefit on its own.
Australia's Student Experience Questionnaire, used in the national Student Experience Survey within the Quality Indicators for Learning and Teaching (QILT) program of the Australian Government Department of Education, asks students how helpful they have found supports and services provided by their institution. One listed item is "career development services (including employability skill development, careers advice and planning)." Responses run from "Not at all helpful" to "Very helpful," with a "Not applicable" option. This is a perception of helpfulness among students who answer the item. It does not record whether a student used the service, why a student did not, or what happened afterward.
A different approach appears in the graduate employment survey run by AQU Catalunya, the quality assurance agency for Catalan universities, which asks graduates how they found their first job and lists university career services as one route among several. That records one specific outcome of one specific service (placement into a first job) but not advising, workshops or the many students who used career services without finding a job through them.
Neither format is a flaw in itself. The point is that a satisfaction item, a route-to-first-job item and an appointment count each answer a different question, and none can be substituted for another.
What the research shows
Participation is lower than provision suggests
Bradley, Quigley and Bailey (2021, Studies in Higher Education) surveyed 258 undergraduate psychology students in a UK department that delivered employability mainly through the central careers service rather than embedding it in the curriculum. According to the abstract, engagement with careers service events was low, typically under 50% attendance, although attendance rose over the course of the degree. The study covers one discipline in one department, so it shows that low use is possible even where services exist, not how common it is.
Jackson, Lambert, Sibson and colleagues (2024, Higher Education Research & Development) surveyed 324 bachelor's graduates from two Australian universities who finished in 2020, 2021 or 2022, with focus groups of 11 graduates. Courses with extensive mandatory placements (such as medicine, nursing and education) were excluded. Graduates rated their engagement in 10 employability-building activities, and mean participation hovered around "to some extent." External paid work and work-related activities such as internships and projects were the most common; entrepreneurship activities and global experiences were the least common. Work and study commitments were rated the most influential barriers; other barriers grouped into personal pressures (such as financial and health concerns) and a lack of confidence, awareness or appreciation of the activities. Barriers varied somewhat by background; for example, graduates with a disability reported higher personal pressures.
Two findings from that study matter directly for career services. First, graduates rated career counselling the least useful activity for developing a sense of professional self and among the least useful for building networks, while work-related activities were rated most useful for securing work. Second, graduates whose parents were university educated found career counselling more useful for understanding their professional self than first-in-family graduates did. Focus group participants also reported that some events felt too broad for their needs and that participation was more likely when activities were embedded in the curriculum. These are retrospective self-reports from two institutions, and the authors note possible effects of COVID-19 lockdowns on responses.
Interventions improve career readiness measures, rarely tested against jobs
Soares, Carvalho and Silva (2022, Australian Journal of Career Development) systematically reviewed published evaluations of career interventions for university students. From 596 articles screened, 26 were retained. Group interventions predominated, and most evaluations used pre-test and post-test designs, some with a comparison or control group. The reported outcomes were generally positive for career decision-making, planning, exploration, adaptability, optimism and perceived employability. The authors also note the scarcity of studies with follow-up assessments and call for more longitudinal designs.
A later review by Sampaio, Taveira, Carvalho and Silva (2024, Societies) examined career interventions delivered at a distance. According to its abstract, 16 articles were included; samples were mostly university students, group formats and pre-test and post-test evaluation predominated, and few studies evaluated distance interventions at all. For universities that have moved advising, workshops and employer events online, this means the evidence on digital delivery is still small.
The common pattern is that the outcomes measured are psychological constructs such as decision-making self-efficacy, captured shortly after the intervention. Those constructs matter, but they are not the same as employment, job quality or career progression.
Evidence on employment and earnings is thin
The clearest employment evidence comes from outside higher education. Liu, Huang and Wang (2014, Psychological Bulletin) meta-analyzed 47 experimentally or quasi-experimentally evaluated job search interventions for job seekers in general. According to the abstract, participants' odds of obtaining employment were 2.67 times those of control groups. Interventions with components such as teaching job search skills, improving self-presentation, boosting self-efficacy and promoting goal setting were more effective, and interventions promoted employment only when they included both skill development and motivation enhancement. The populations were not specifically university students, so the result indicates which components are promising rather than how a given campus service performs.
Within higher education, Scandurra, Hermannsson and Fusaro (2026, International Journal of Manpower) estimated the earnings effect of finding a first job through a university career service, using six waves (2008 to 2023) of the AQU Catalunya graduate survey. Their analytic sample was 5,919 graduates under 30 who moved directly from study to work; 6.89% had found their first job through career services. To address the fact that graduates are not randomly sorted into job-search routes, they used an instrumental variable: the share of peers in the same university, field and survey year who found jobs that way. Their main estimate was an earnings premium of 8.5% (9.2% in a simpler regression), larger for women (15.7%) and for graduates without prior work experience (about 17%); the estimate for men was not statistically significant.
The authors are explicit about the limits. The estimate applies to graduates whose route choice is influenced by peers, not necessarily to everyone. The data are repeated cross-sections, so medium- and long-term returns cannot be assessed. Cohort-level confounding is not fully ruled out, the results may mainly generalize to Southern European systems, and costs were not assessed. The measure also captures entry into a job through career services, not the effect of advising or workshops used at other times.
Randomized evidence for university-level career coaching is still emerging. The research organization MDRC describes an ongoing randomized trial of COOP Careers, a program for unemployed or underemployed college graduates, particularly first-generation and low-income graduates, with employment and earnings to be measured from administrative records over two years. Its project page reports no results yet.
Our interpretation: why usage data mislead in both directions
The studies above point to a structural problem with the most common evidence universities hold, which is a comparison of outcomes between students who used career services and those who did not. That comparison can mislead in either direction.
- Upward bias: motivated, well-networked students may seek out every resource, including career services, and would have done well regardless. Their success then gets credited to the service.
- Downward bias: students who are struggling to find work are more likely to ask for placement help. If they still have weaker outcomes than non-users, the service looks ineffective even if it helped.
- Missing groups: the Jackson study suggests that time pressure, finances, health and low awareness keep some students away. A service measured only on its users never sees them.
Our reading of the combined evidence is that well-designed interventions reliably change how prepared students feel and how they plan, that combining practical job-search skills with motivation support is a promising design principle (based on evidence from job seekers in general), and that robust evidence linking routine campus career services to later employment remains scarce. A university claiming that its career service "improves graduate outcomes" should be able to show which outcome, for which students, compared with whom.
An evidence ladder for career services
The table below is an editorial framework proposed by this article; it is not the UNIRANKS methodology and does not describe how UNIRANKS evaluates institutions. Each level adds a stronger kind of evidence. Most institutions can reach the third level with existing systems, and a growing number can reach the fourth.
| Level | Question answered | Typical evidence | What it cannot show |
|---|---|---|---|
| 1. Provision | What exists? | Service catalogue, staffing, opening hours, formats, languages | Whether anyone can or does use it |
| 2. Reach | Who uses it, and who does not? | Usage records linked to student characteristics (mode of study, first-in-family status, international status, disability, field) | Whether use helped |
| 3. Experience | Did users find it helpful? | Short post-service surveys; national items such as perceived helpfulness | Whether perceptions match later outcomes |
| 4. Linked outcomes | How do users fare afterward? | Usage records matched to graduate survey or administrative outcomes, with adjustment for prior attainment and background | Whether differences are caused by the service |
| 5. Comparative evaluation | Would users have done worse without it? | Randomized or phased rollouts, waitlist designs, credible quasi-experiments | Long-run effects, unless follow-up is extended |
What to record for each type of service
Different services have different plausible outcomes. Judging an employer fair by the share of graduates in employment, for example, attributes to one event an outcome shaped by many factors. The table pairs each service type with a reasonable near-term outcome and a common measurement error. It is also part of the editorial framework above.
| Service | Useful reach measure | Plausible near-term outcome | Common error |
|---|---|---|---|
| Individual advising | Share of each student group with at least one appointment before final year | Change in career decision-making or a concrete plan, recorded at follow-up | Counting repeat visits by the same students as reach |
| Workshops and career modules | Completion rates, split by curricular and voluntary delivery | Pre and post measures with a comparison group | Post-session satisfaction treated as learning |
| Employer events and fairs | Attendance by field and study mode; employer mix relative to student fields | Applications, interviews or internships arising from contacts | Reporting number of employers instead of relevance to attendees |
| Placement and vacancy support | Who receives offers through the service | Route into first job, job match and earnings, ideally with a comparison | Crediting all graduate jobs to the service |
| Digital platforms | Active users rather than registered accounts | Completed actions (applications, bookings) and follow-up outcomes | Logins treated as engagement |
| Follow-up after graduation | Share of recent graduates contacted and responding | Outcomes for graduates still seeking work | Low response treated as representative |
Practical steps
For career services leaders
- Report reach as a share of each student group, not as total appointments, and publish who is not being reached.
- Link usage records to graduate outcome data under appropriate data-protection rules, and adjust comparisons for prior attainment, field and background before interpreting any difference.
- When launching a new program, consider a phased rollout or waitlist so that some students can serve as a comparison group for a time.
- Build follow-up into evaluations; a post-session survey alone cannot show durable change.
- Where participation is low among students with heavy work or caring commitments, test embedding career learning in the curriculum and evaluate whether reach improves.
For university leaders and quality reviewers
- Ask which level of the evidence ladder each claim about career services reaches.
- Treat satisfaction and helpfulness ratings as one input, alongside reach and linked outcomes.
- Fund evaluation as part of the service rather than as an optional extra.
For researchers
The gaps are clear: longer follow-up, employment and job-quality outcomes rather than only psychological constructs, evidence on digital delivery, and designs that handle self-selection. Among the sources reviewed here, evidence on employment and earnings outcomes comes from very few countries.
Students looking for general guidance on career planning can find it in the UNIRANKS career guidance section; this article addresses how institutions evaluate their services rather than individual career decisions.
Limitations
This article draws on six peer-reviewed sources. Three were read in full (Jackson and colleagues; Soares, Carvalho and Silva; Scandurra, Hermannsson and Fusaro) and three from their abstracts only (Bradley, Quigley and Bailey; Sampaio and colleagues; Liu, Huang and Wang). Statements about the abstract-only papers stay within what their abstracts report. The participation studies come from single departments or two institutions in the UK and Australia and rely on self-report. The job search meta-analysis covers job seekers in general, not university students specifically. The earnings study is from one region with distinctive labor-market features, and its authors limit their own conclusions accordingly. The two tables are editorial frameworks that have not been validated as measurement instruments.
Related reading on UNIRANKS
For the wider evidence base on graduate outcomes, see the Employability & Career Success research hub. Work-related activities, which graduates in the studies above rated highly, are examined in Do University Internships Improve Graduate Career Outcomes?
Editorial disclosure
UNIRANKS publishes university rankings, and its methodology includes an Employability & Career Success dimension. This article reviews independent research on the evaluation of university career services. It does not describe how UNIRANKS scores institutions.
Research References
- Bradley, A., Quigley, M., & Bailey, K. (2021). How well are students engaging with the careers services at university? Studies in Higher Education, 46(4), 663-676. https://doi.org/10.1080/03075079.2019.1647416 (verified 22 November 2026)
- Jackson, D., Lambert, C., Sibson, R., et al. (2024). Student employability-building activities: participation and contribution to graduate outcomes. Higher Education Research & Development, 43(6), 1308-1324. https://doi.org/10.1080/07294360.2024.2325154 (verified 22 November 2026)
- Liu, S., Huang, J. L., & Wang, M. (2014). Effectiveness of job search interventions: A meta-analytic review. Psychological Bulletin, 140(4), 1009-1041. https://doi.org/10.1037/a0035923 (verified 22 November 2026)
- Sampaio, C., Taveira, M. do C., Carvalho, C., et al. (2024). Career Interventions at a Distance: A Systematic Literature Review. Societies, 14(11), 230. https://doi.org/10.3390/soc14110230 (verified 22 November 2026)
- Scandurra, R., Hermannsson, K., & Fusaro, S. (2026). The career services premium: how university job-search boosts graduate earnings. International Journal of Manpower, 47(10), 237-256. https://doi.org/10.1108/IJM-10-2025-0875 (verified 22 November 2026)
- Soares, J., Carvalho, C., & Silva, A. D. (2022). A systematic review on career interventions for university students: Framework, effectiveness, and outcomes. Australian Journal of Career Development, 31(2), 81-92. https://doi.org/10.1177/10384162221100460 (verified 22 November 2026)
Sources
- Student Experience Questionnaire (SEQ). Quality Indicators for Learning and Teaching, Australian Government Department of Education. https://www.qilt.edu.au/surveys/student-experience-survey-(ses)/student-experience-questionnaire-(seq)/ (accessed 22 November 2026)
- La inserció laboral dels graduats i graduades de les universitats catalanes (ninth edition). AQU Catalunya. https://www.aqu.cat/en/Studies/Dissemination/Publications/Publications-finder/La-insercio-laboral-dels-graduats-i-graduades-de-les-universitats-catalanes (accessed 22 November 2026)
- Evaluation of COOP Careers. MDRC. https://www.mdrc.org/work/projects/evaluation-coop-careers (accessed 22 November 2026)
