Decision Architecture Entry #0131 Classified Declassified

The reason you decide differently before and after lunch

Across a long session of similar decisions, the safe default is chosen more often as the session goes on. Why that happens is still argued about.

No visual record attached The written record below is complete.
Plate 123 — forty decisions, plotted against the clock rather than the file.

Intuition test — answer before you read on

A reviewer’s outcomes drift towards the low-justification option across a long session of randomly ordered cases. What is the fair reading?

A reviewer works through forty comparable applications in one sitting. The first ten receive detailed notes and a mix of outcomes. The last ten receive shorter notes and lean towards whichever answer requires no justification. The applications were randomly ordered, and nothing about them changed.

What everyone sees

Each decision is treated as independent, and reviewers experience them that way — every case is judged on its file. Position in the queue is not part of the record and does not feel like an input, so drift across a session is invisible from inside it.

What is actually happening

Sequence position in repeated judgement has been observed to shift outcomes towards the option requiring least justification, and Danziger and colleagues reported such a pattern in judicial rulings across a day. The interpretation is contested: Weinshall-Margel and Shapard argued that case ordering and scheduling could produce the same curve, and the broader ego-depletion literature has had mixed replication. What is better supported is that deliberate evaluation is effortful and that effort-saving strategies increase with time on task.

Why it stays hidden

The drift hides for two reasons. Nobody sees their own decisions arranged by position, because the record is organised by case rather than by clock. And every individual decision has a complete, honest, file-based justification, so there is no residue pointing at the sequence. A pattern that only exists in aggregate cannot be noticed one case at a time.

Position in the queue is an unrecorded input. Effort-saving accumulates across a session while each case still gets a full justification.

Position in the queue is an unrecorded input. Effort-saving accumulates across a session while each case still gets a full justification.

The hidden part — entry #0131

Collect this card

Position in the queue is an unrecorded input. Effort-saving accumulates across a session while each case still gets a full justification.

0 / 10,000 collected

Sources & further reading 3
  1. Danziger, Levav & Avnaim-Pesso, "Extraneous Factors in Judicial Decisions", PNAS, 2011
  2. Weinshall-Margel & Shapard, "Overlooked Factors in the Analysis of Parole Decisions", PNAS, 2011
  3. Hagger et al., "A Multilab Preregistered Replication of the Ego-Depletion Effect", Perspectives on Psychological Science, 2016

Circulate this file

Annotations are reserved for archive members.

Sign in to annotate