Interleaved Competitor Retrieval Test¶

Diagnostic estimation — instantiates Encoding–Retrieval Context Alignment

Tests recall with the real look-alikes and sound-alikes mixed in, so you find out whether a cue points uniquely to the target or also fires for its competitors.

The Interleaved Competitor Retrieval Test is a stress test that deliberately mixes the target with its confusable competitors — items that share a name, a cue, or a context — and measures whether retrieval still lands on the right one. Its defining move is that it does not ask "can you recall the target?" in isolation; it asks "can you recall the target when its rivals are present and share the cue?" That exposes cues which are non-diagnostic because they also point at competitors — false cues — and yields a discrimination signal: how often retrieval hits the target versus intruding on a rival. Where the Cue-Diagnosticity Ablation Test removes cues to find the load-bearing ones, this test adds competitors to find the ambiguous ones.

Example¶

A pharmacist recalls "give hydrOXYzine" perfectly when it's the only thing on her mind. The test interleaves its real-world rival: hydrALAZINE — same opening letters, stocked near it, reached for in similar order contexts. Under interleaving, the cue "hydr-, that patient, that shelf" fires for both drugs, and the intrusion rate jumps. The signal is unambiguous: that cue is non-diagnostic against this competitor, and the fix is a more diagnostic cue — tall-man lettering, physical separation, an order-entry alert. The collision was only visible because the rival was present in the test; tested alone, the target came back flawlessly and the danger stayed hidden.^[1]

How it works¶

You enumerate the real competitors that share cues with the target, interleave them into the retrieval task under realistic timing and context, and measure target-hits versus competitor-intrusions. A cue that yields intrusions is false or ambiguous; the intrusion rate is the outcome signal. The distinguishing discipline is that the competitor set must be the one the actual retrieval environment presents — not arbitrary distractors and not an absence of rivals — because interference that isn't staged can't be measured.

Tuning parameters¶

Competitor set — which rivals and how many; they must be the ones present in the real context, or the test either cries wolf or misses the true collision.
Interleaving density — how close together competitors appear; tighter spacing makes interference bite harder.
Similarity dimension — whether rivals share a name, a context, a cue, or a time — test the dimension that actually collides.
Timing and pressure — interference sharpens under time pressure, so test at the tempo of real use.
Intrusion scoring — counting competitor-intrusions and near-misses, not just target hits, so the failure mode is visible.

When it helps, and when it misleads¶

Its strength is surfacing interference-driven errors that isolated testing hides — the "fine in practice, wrong under load" failures where similar traces compete for the same cue. It misleads when the competitor set is unrepresentative: test against rivals nobody meets and it raises false alarms; omit the real rival and it certifies a cue that will fail in the field. The classic misuse is running the test without the real competitors present — so it passes — and then shipping into an environment full of them. The guard is to build the competitor set from the actual retrieval environment and to carry the intrusions (not just the hit rate) forward into cue redesign.

How it implements the components¶

The test realizes the interference-diagnosis side of the archetype — probing whether cues discriminate the target from its rivals:

interference_and_false_cue_check — it is the check for interference and for cues that fire for competitors as readily as for the target.
retrieval_outcome_signal — the target-hit-versus-competitor-intrusion rate is the outcome signal it emits, which downstream cue design and fading gate on.

It does not isolate each cue's positive contribution (diagnostic_cue_set, from Cue-Diagnosticity Ablation Test) or separate not-stored from not-accessible (that's Free-Recall-Then-Recognition Probe); it adds rivals to expose ambiguous cues.

Instantiates: Encoding–Retrieval Context Alignment — the test supplies the interference check the alignment needs before trusting a cue that has competitors.
Sibling mechanisms: Cue-Diagnosticity Ablation Test · Free-Recall-Then-Recognition Probe · Scenario-Based Retrieval Test · Representative-Environment Simulation · Context Reinstatement Protocol · Context Translation Card · Context-Switch Recall Drill · Cue-Fading Schedule · Environmental Retrieval Cue · External Checklist or Job Aid · Mnemonic Cue Pairing · Post-Event Re-Encoding Debrief · Spaced Retrieval Scheduler · Transfer-Appropriate Processing Rehearsal · Varied-Context Retrieval Practice

Notes¶

This test and the Cue-Diagnosticity Ablation Test are complementary halves of cue diagnosis: ablation removes cues to find which ones recall depends on; the competitor test adds rivals to find which cues fail to discriminate. A cue can be load-bearing yet non-diagnostic — it passes ablation and fails here — so a cue is only trustworthy once it has survived both.

References¶

[1] The fan effect — the more distinct items are associated with a single cue, the slower and less reliable retrieval through that cue becomes (J. R. Anderson). Interleaving competitors is how you provoke and measure it. ↩