Social anchor effects in decision-making under ambiguity

I experimentally examine whether feedback about others' choices
provides an anchor for decision-making under ambiguity. In a
between-subjects design I vary whether subjects learn choices made
individually by a "peer" in a first part when facing the same task
a second time, and whether prospects are defined over gains or
losses. My key findings are that the relative ambiguity attitude
(compared to the peer's) significantly matters for shifts in
individual attitudes, and that dynamics considerably differ between
gain and loss domains. For gains, learning to be comparably
ambiguity averse increases the likelihood for such shifts, relative
to the individual condition; for losses, this likelihood decreases
only if peers learn to exhibit exactly the same attitude. Further,
I observe imitative shifts towards the peer's attitude in the gain
domain, but only towards neutrality in the loss domain. Shifts
towards neutrality for losses also appear significant without
social anchor suggesting that ambiguity seeking might not be
particularly robust. Moreover, cognitive ability positively
correlates to shifts towards neutrality in the gain domain, but has
no impact in the loss domain.