The effect of time ambiguity on choice depends on delay and amount magnitude

Abstract

Abstract Time ambiguity—that is, having partially/fully incomplete information about when an outcome will occur—is common in everyday life. A recent study showed that participants preferred options with time‐exact delays over options with time‐ambiguous delays, a phenomenon they called time‐ambiguity aversion. However, the empirical robustness and boundaries of this phenomenon remain unexplored. We conducted three online studies: Study 2 (n = 118) was a replication of Study 1 (n = 76) using preregistered analyses; Study 3 (n = 202; preregistered) was a follow‐up study suggested during review. In Studies 1 and 2, participants completed hypothetical choices between €5 today versus later‐but‐larger (LL) rewards that systematically varied in their amount, delay, and time‐ambiguity level (e.g., for a 180 day delay, time ambiguity varied from 179 to 181 to 0–360 days). Effects of time ambiguity on choice were best encoded in an absolute, dose‐dependent manner and depended on delays and amounts: Increasing time ambiguity led to more time‐exact LL choices at shorter delays but more time‐ambiguous LL choices at longer delays. Additionally, time‐ambiguity ranges including today were chosen more frequently than ranges excluding today, akin to the present bias in intertemporal choice. Lastly, evidence suggested that more time ambiguity was preferred for smaller LL amounts yet disliked for larger LL amounts. Study 3 demonstrated that time‐risk and time‐ambiguity preferences are differentiable by giving participants choices involving hypothetical time‐exact, time‐ambiguous, and time‐risky options. Taken together, our results extend the nascent literature on time ambiguity by showing that (i) time‐ambiguity preferences are distinguishable from both time‐risk and delay preferences and (ii) time ambiguity is not generally aversive, but its impact depends on delay and amount magnitude.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 3 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.