Appearance
Choose series-level description for homogeneous, high-volume, lower-demand material where context matters more than pinpoint retrieval, and reserve item-level description for unique, heavily used, or heterogeneous records where users genuinely need to find one document among thousands. The decision is fundamentally about cost versus retrieval precision: item-level work is roughly five to fifteen times slower per linear metre, so spending it everywhere is the fastest way to grow a backlog you will never clear.
What is the real difference, in practice?
Series-level description tells the user what is here and why it hangs together: a run of board minutes 1920–1975, arranged chronologically, fourteen volumes, gaps in 1941–1943. Item-level description tells them exactly which document is which: the minutes of 14 March 1934, three pages, discussing the harbour extension. One record versus dozens.
Series level preserves the relationships — provenance, original order, the meaning that comes from context. Item level shatters a series into atoms that are individually findable but can lose the thread that connects them. Neither is "better"; they answer different questions.
Which signals push you toward item-level description?
Drop to item level when the material shows these traits:
- Heterogeneity — a "miscellaneous papers" box where every item is different and a series note would be uselessly vague.
- High research demand — the series people request weekly justifies the index.
- Uniqueness and value — founding charters, signed correspondence, unique photographs.
- Discontinuous content — items that do not share a pattern a series note could summarise.
- Digitisation with public access — when each image needs a citable record anyway.
If two or three of these apply, item level probably pays off.
Which signals push you toward series-level description?
Stay high when:
- The material is homogeneous — uniform forms, identical report types, routine ledgers.
- Demand is low or speculative — described "in case", not because anyone asks.
- Volume is large — hundreds of boxes where item work means years.
- Original order is strong — the arrangement itself does the finding work.
A run of 5,000 near-identical insurance claim forms needs one good series note, not 5,000 thin records.
How big is the cost difference?
This is the number that should drive the decision.
| Factor | Series-level | Item-level |
|---|---|---|
| Relative speed | Fast (baseline) | 5–15x slower |
| Records produced | One per series | One per item |
| Retrieval precision | Coarse (browse) | Exact (search) |
| Context preserved | Strong | Weaker, unless linked |
| Maintenance burden | Low | High (every record ages) |
That speed multiplier is not abstract. A collection you could describe at series level in a week might take three to four months at item level. Multiply across a backlog and the strategic stakes are obvious.
Can you mix levels — and how?
Yes, and a well-built finding aid almost always does. ISAD(G) and DACS are explicitly multilevel: you describe top-down, and you are not obliged to go to the bottom everywhere. The pattern that works:
text
Fonds: Hartley Engineering Works (series-level overview)
Series 1. Board minutes, 1920-1975 (series-level)
Series 2. Project files, 1931-1968 (file-level)
File 2/14. Harbour extension, 1934-1937 (file-level)
Item 2/14/3. Signed contract, 12 Mar 1934 (item-level)
Series 3. Routine correspondence, 1920-1975 (series-level)Series 2 earns file-level detail because it is heterogeneous and used; the single signed contract earns an item record because it is unique and valuable; the routine correspondence stays at series level because nothing is gained by atomising it.
How do you document the decision so it is defensible?
Write the level choice and its rationale into the processing plan, not just the catalogue. A one-line justification per series — "item level: high demand, heterogeneous; series level elsewhere" — makes the decision auditable when a colleague or a reviewer asks why the treatment is uneven. Uneven treatment is correct under multilevel description; undocumented uneven treatment looks like neglect.
Key Takeaways
- Series level gives context cheaply; item level gives precise retrieval expensively.
- Choose item level for unique, heterogeneous, or high-demand material; series level for homogeneous bulk.
- The cost gap is real: item-level work is commonly 5–15x slower per linear metre.
- Mix levels within one finding aid — multilevel description expects exactly that.
- More records is not more quality; over-itemising can bury archival context.
- Document the level decision and its rationale in the processing plan.
Frequently Asked Questions
What is the basic difference between series and item-level description?
Series-level description summarises a whole grouping of related records in one record, while item-level description creates a separate record for each individual document or object. Series level is faster and gives context; item level is slower but supports precise retrieval.
Is item-level description always better quality?
No. More records is not more quality. Item-level description applied where it is not warranted wastes effort, multiplies inconsistency, and can bury the contextual relationships that make archival material meaningful.
How do I decide between the two for a given series?
Weigh research demand, uniqueness, homogeneity, and budget. Homogeneous, low-demand bulk gets series-level treatment; unique, high-demand, or heterogeneous material justifies item-level records.
Can I mix levels in one finding aid?
Yes, and you usually should. Multilevel description in ISAD(G) and DACS is explicitly hierarchical — describe most of a fonds at series and file level, then drop to item level only for the parts that earn it.
How much more does item-level description cost?
Item-level work commonly runs five to fifteen times slower than series-level for the same physical extent, because each item needs its own title, date, extent, and access points. That multiplier is the core trade-off.
Does item-level description help with digitisation?
It can, because each digital object often needs a metadata record anyway. But you can also digitise at file level with sequence numbers, so item-level cataloguing is not a prerequisite for putting images online.