Controlled Vocabulary Indexing

Exhaustivity is about how many concepts an indexer captures from a document — its depth of indexing. Indexing a document with many terms (high exhaustivity) represents minor themes and tends to raise recall, while indexing with few terms captures only the dominant topics. Specificity is about how finely each captured concept is expressed — choosing the most precise term that covers it rather than a broader one. The two are independent policy and judgement choices: exhaustivity controls how much of a document you index, specificity controls how precisely you name what you do index.

Consistency is the degree to which different indexers assign the same terms to the same document; it matters because inconsistent indexing scatters similar documents under different terms, so searchers miss material even when the vocabulary is sound. It is commonly measured with coefficients such as Hooper's, which divides the number of terms two indexers agree on by the total number of distinct terms they assigned (intersection over union). Tracking consistency lets an organization detect drift and improve it through indexing guidelines, training, and review.

Free-text (or derived) indexing uses the document's own words as access points; it is cheap and current but suffers from synonymy and ambiguity. Controlled vocabulary indexing assigns terms from a managed list, gaining precision and the power of a structured vocabulary at the cost of manual effort. Automatic indexing applies algorithms — from term-frequency methods to machine learning — to assign either derived or controlled terms at scale. In practice the approaches are increasingly combined: controlled indexing supplies precision and ground truth, while automatic and assisted methods extend coverage to collections too large for purely manual work.