Choose a museum task rather than assuming every internship is curatorial
The Met's departmental-placement resource describes several parts of museum work and notes that available projects vary. Start with the assignment and learning goal, not only a favorite collection. A museum can need research, documentation, design, visitor-facing work and operational support.
For this comparison, a collections-record exercise focuses on identifying an object and maintaining traceable information. A curatorial research exercise focuses on a question about the object and how evidence supports an interpretation. The tasks can overlap, and the actual department determines responsibilities and supervision.
Neither exercise authorizes handling, treating or moving a real object. You can demonstrate the relevant habits with fictional records and your own writing before obtaining any access to a collection.
Sources[1]
Resolve the record conflict without inventing a fact
The synthetic file contains two versions of OBJ-001 and one record for OBJ-002. The older OBJ-001 title is “Blue bowl,” while a later note proposes “Storage vessel.” The date is a range, 1880–1900. The creator is not recorded in the supplied material.
| Object ID | Version | Title | Date evidence | Attribution evidence | State |
|---|---|---|---|---|---|
| OBJ-001 | 1 | Blue bowl | 1880–1900 | Creator not recorded | Earlier record |
| OBJ-001 | 2 | Storage vessel | 1880–1900 | Creator not recorded | Proposed revision; review needed |
| OBJ-002 | 1 | Printed festival notice | Date not established | Printer named in source note | Current teaching record |
Do not simply replace the older title, turn the date range into a confident 1890 or label the creator “Anonymous” as though that were an established attribution. A proposed title needs its evidence and review state. A missing creator field means the supplied record does not settle the question.
Keep a stable identifier, source reference, version and status. The two OBJ-001 rows are versions of the same object, not proof that the collection contains two objects. Deduplicating by title would fail because titles can change; merging every similarly named item would create a different error.
Write an interpretation whose claims can be traced
Choose OBJ-002 and pose a question: what could the notice tell a visitor about the event's intended audience? Separate what the fictional source explicitly states from what you infer. A printed location or named organizer can be documented; the audience's actual composition would require additional evidence.
Write a short label in plain language and a separate research note containing the evidence, uncertainty and alternative explanation. The label should not become more confident merely because it is shorter. Preserve uncertainty in a way a visitor can understand rather than burying it in an internal appendix.
Use an original description or permitted source material. A digital image available online is not automatically licensed for your portfolio. Keep image availability and permission to reuse as separate fields, and avoid copying a museum's published label as your own analysis.
Make the cataloging checks concrete
Validate unique object identity separately from record-version identity. Check that date ranges have valid order and that a missing bound is not replaced with zero. Confirm that every revision cites the information that motivated it and whether it has been accepted, rejected or remains proposed.
For each public-facing statement, ask whether the source supports the whole sentence. A record establishing a place does not necessarily establish who made an object. A stylistic resemblance is not a documented ownership history. The exercise's goal is not to settle an attribution but to show how you avoid losing the distinction.
Package the old and proposed records with an exception log. A reviewer should be able to understand what changed, why and who still needs to decide, without reconstructing the process from filenames such as final-v7.
Read a specific museum program as its own set of terms
The Met's summer program offers ten-week opportunities for undergraduates, graduate students and specified recent graduates. Its published eligibility includes current undergraduate sophomores through seniors and a recent-graduate window measured against the application deadline. Read the full conditions rather than applying that summary to every museum.
As reviewed September 16, 2026, the application control is closed and the page announces opening in mid-November 2026 and closing January 13, 2027 at 5 p.m. Eastern. It requests essays, a résumé, an unofficial transcript and two recommendations, including one academic reference; a portfolio is requested for Design Department candidates.
Those distinctions affect preparation. Arrange the appropriate references and explain what you want to learn in the actual department. Do not treat an announced future window as a currently open application.
Sources[2]
Show the work you want someone to trust you with
For collections-oriented work, present the record comparison and explain the identity and revision checks. For research-oriented work, present the question, short interpretation and supporting note. Explain which statements remain unresolved and what evidence would be needed to address them.
Use the museum's program instructions to decide whether a work sample is requested. Do not upload a long unsolicited portfolio when the application asks for specific materials only. A brief project example in an essay can still show careful thinking.
The application should connect your interests to a task you are ready to learn. You do not need to imply that a student cataloging exercise made you a curator. A precise account of a record conflict, an evidence limit and a considered revision is enough to make the work concrete.