Incomplete-Block Design¶
Assignment-design method — instantiates Blocking Design
Assigns only a connected subset of the treatments to each block when a block cannot hold them all, arranging the overlaps so every treatment comparison is still recoverable somewhere.
Incomplete-Block Design is the answer to a hard constraint: there are more treatments than any single block can accommodate. Rather than abandon blocking or shrink to a completely randomized layout, it puts a subset of the treatments in each block and chooses those subsets so the design stays connected — every pair of treatments is compared, directly or through a chain of shared blocks. Its defining move is engineering coverage under a feasibility limit: when full completeness is impossible, arrange the overlaps so no contrast is orphaned. In its balanced form, every pair of treatments appears together in a block equally often.
Example¶
A food company wants to rank seven new hot-sauce recipes, but sensory science is clear that a taster's palate fatigues after about three samples — no one can fairly judge all seven in a sitting. A complete block (one taster tries every sauce) is infeasible. So each taster is a block of size three, and the recipes are allotted to tasters so that every pair of sauces is tasted together by the same number of tasters and the whole set stays connected.
No taster ever compares all seven, yet because the blocks overlap in a balanced pattern, every pairwise preference is estimable and the rankings hold together. The design has traded the impossible goal of full coverage for connected coverage — enough shared comparisons to tie the seven sauces into one comparable set.
How it works¶
- Size blocks to the real limit. The block size is set below the number of treatments by a genuine feasibility constraint — palate fatigue, bench space, session length — not by choice.
- Choose connected subsets. Allot treatments to blocks so the design remains connected; in a balanced incomplete-block design, every treatment pair co-occurs equally often, spreading precision evenly.
- Let overlaps carry the load. Comparisons not made inside a single block are recovered through the chain of blocks that share treatments, provided connectivity holds.
Tuning parameters¶
- Block size — how many treatments per block; larger approaches completeness, smaller eases the feasibility limit but demands more blocks.
- Balance — fully balanced (every pair equally often) versus partially balanced, trading uniform precision against feasibility.
- Replication — how many blocks overall, setting how much information accumulates on each contrast.
- Priority contrasts — when full balance is unreachable, which comparisons to favor with more overlap.
When it helps, and when it misleads¶
Its strength is enabling blocked comparison when the treatment set is simply too large for any natural block — many products, limited sessions, small litters, scarce rigs. It preserves most of blocking's precision without demanding the impossible.
It misleads when the design silently becomes disconnected: if some treatments never share a block with others, their contrasts are unidentifiable and no analysis can recover them. It also misleads when the incomplete structure is ignored at analysis time — naive treatment means are biased, because they mix within- and between-block information unevenly.[n1] The discipline is to verify connectedness before running and to analyze with the block structure rather than against it.
How it implements the components¶
block_size_overlap_and_feasibility_design— its reason for existing: block capacity is smaller than the treatment set, so it sizes blocks to a feasible subset and engineers the overlap pattern without excessive exclusion.treatment_coverage_and_cell_completeness— it manages coverage as connectedness rather than completeness: not every treatment in every block, but enough shared blocks that every intended contrast stays estimable.
It does NOT achieve full completeness with every treatment in every block (treatment_coverage_and_cell_completeness in its complete form is Randomized Complete-Block Design's) or recover the contrasts at analysis time (block_aware_analysis_and_weighting_plan, Block-Adjusted Effect Estimator); block construction by unit similarity (block_construction_and_similarity_rule) is Matched-Pair Randomization's, and the precision accounting (precision_power_and_design_effect_assessment) is Randomized Complete-Block Design's.
Related¶
- Instantiates: Blocking Design — the feasibility-constrained case, where blocks are too small for the full treatment set.
- Sibling mechanisms: Randomized Complete-Block Design · Block-Adjusted Effect Estimator · Matched-Pair Randomization · Stratified Randomization Schedule · Cluster or Site Blocking · Permuted-Block Sequence · Covariate-Adaptive Randomization · Time, Batch, Run, or Location Block · Within-Block Randomization Inference
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: Incomplete-Block Design operates as a bounded trial, probe, simulation, or rehearsal that generates evidence from performance because it assigns only a connected subset of the treatments to each block when a block cannot hold them all, arranging the overlaps so every treatment comparison is still recoverable somewhere
Independent corroboration: The frozen evidence defines Incomplete-Block Design as 'Assigns only a connected subset of the treatments to each block when a block cannot hold them all, arranging the overlaps so every treatment comparison is still recoverable somewhere', so its operative form is Experiment, Test & Rehearsal.
Nearest alternative: Structure, Architecture & Configuration — The block arrangement exists specifically to generate recoverable experimental comparisons under feasibility constraints.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Statistics & Experimental Design
Origin pattern: Single lineage
Present-day reach: Multi-domain
Rationale: Balanced incomplete-block designs are a classical experimental-design construction for recoverable treatment comparisons under capacity limits.
Related originating lineages:
- Mathematics — Combinatorial block-design theory supplies the balanced incidence structure.
Review resolution: Both reviewers independently assign statistics_experimental_design as the primary originating domain, so that shared primary is retained. Alternate domains are the union of reviewer-identified formative or independently originating lineages; later application settings alone are excluded. The evidence describes one principal historical lineage. It has established independent use across several domains, but that does not make it domain-free. The encyclopedia entry generalizes the established mechanism without creating a new composite lineage.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
Connectedness is the identifiability condition. A pretty, balanced-looking allotment that happens to split the treatments into two islands leaves the between-island contrasts unestimable — a failure that is invisible until someone tries to compare treatments that never shared a block.
[n1] A balanced incomplete-block design (BIBD) arranges the subsets so every pair of treatments appears together in the same number of blocks; it is the classical way to keep precision even across contrasts when completeness is infeasible, and its analysis must combine within- and between-block information rather than average raw treatment means. ↩