Cambridge Recommendation Letter Example: Jazz musician to music education policy (Score 93)

I am a lecturer in research methods at , where I convene an advanced seminar in which students design and defend a grant-style research proposal over the course of a term. The applicant enrolled in this seminar during [term/year] and submitted a proposal examining the relationship between structured community music education programmes and measurable outcomes in civic participation among adolescents in under-resourced urban settings. I am writing in support of their application to the MPhil in Public Policy at Cambridge. I should be direct about the context of my knowledge. I did not supervise the applicant across a full degree programme. What I can speak to is the quality of their thinking under the specific pressures of proposal design: how they handled identification problems, how they responded when I pushed back on scope, and how they reasoned about what their evidence could and could not support. That is a narrower window than a dissertation supervisor would have, but it is a revealing one. The proposal itself was unusual in its starting point. The applicant drew on direct experience running jazz education sessions in community settings to motivate a policy question about whether structured arts programming produces durable civic skills — not aesthetic ones, but measurable behavioural ones: attendance at community meetings, participation in local decision-making processes, self-reported civic efficacy. Most students in this seminar arrive with a literature-driven question. This applicant arrived with a question rooted in observed practice and then worked backwards to the literature, which created both a strength and an early problem I want to describe honestly. The strength was that the applicant's identification of the treatment was unusually precise. They knew, from the inside, what 'structured community jazz education' actually meant as an intervention — the duration, the pedagogical model, the instructor-to-student ratio, the difference between drop-in and cohort-based formats. When I asked the seminar to define their treatment variable clearly enough that a stranger could replicate it, most students produced vague descriptions. The applicant produced something close to a programme theory. That specificity matters for causal identification, and I noted it. The early problem was scope. The first draft proposed a mixed-methods design spanning three cities, two cohort waves, and a qualitative strand involving semi-structured interviews, all within a notional twelve-month timeline. In our proposal defence session — where I ask each student to justify their timeline and sampling frame in front of peers — I pressed the applicant on this directly. I asked them to identify which strand was load-bearing for the central claim and which strands were aspirational. The initial response was to defend the full design on the grounds that the question required it. I pushed back: the question may require it eventually, but a feasible first study does not. It took two revision cycles before the applicant trimmed the design to a single-city, single-cohort study with a clearly bounded comparison group and a realistic ethics timeline. The final version was tighter and more defensible. I want to be clear that the willingness to revise was genuine — but the instinct to over-scope was real, and I would expect a Cambridge supervisor to hold that boundary firmly in the early stages of the MPhil. What I found more consistently strong was the applicant's handling of the ethics section, which many students treat as procedural. The applicant identified a genuine tension: participants in community music programmes are often minors from families with limited institutional trust, and standard informed-consent procedures may systematically exclude the most relevant participants. They raised this unprompted in a seminar discussion, framed it as a validity problem rather than just a compliance problem, and proposed a modified consent protocol drawing on participatory research literature. I had not assigned that literature. The applicant had found it independently and applied it to a real methodological constraint. That kind of lateral thinking — connecting an ethical problem to a measurement problem — is not common at this stage. The policy memo the applicant produced as a companion piece to the proposal also demonstrated an ability to translate a research design into a form usable by a non-specialist audience. The memo was addressed to a notional local authority commissioner and argued for a pilot evaluation of community music programming using the study design as its evidence base. The argument was clear, the limitations were acknowledged rather than buried, and the recommendations were proportionate to what the proposed study could actually establish. I reviewed a draft of this memo and found it more disciplined than the proposal had been at the same stage. The applicant's academic background is not a conventional policy research pipeline, and I think it is worth naming that directly. The move from practice-based music education to postgraduate policy research requires a bridging of registers that not every applicant manages. Based on what I observed in this seminar, the applicant has begun that bridge — the proposal showed genuine methods literacy, and the policy memo showed communication discipline. Whether the full transition is complete is something a Cambridge supervisor would need to assess across the MPhil year.