Data from: Cognitive tasks could be biased towards generalists: a lesson from wild non-eusocial bees

In this study, we compared associative learning performance and
exploratory tendencies between dietary specialist and generalist bee
(Anthophila) species using a closed-environment task with free-moving bees
called the free-moving proboscis-extension response (FMPER). We found
lower participation rates than expected, especially among specialist
species, which hindered our ability to answer our primary question. Thus,
we combined our data with another published dataset that reported results
from the same learning task but for several different bee species (again
including specialists and generalists) to investigate the relation of diet
breadth with associative learning and exploration across a broader species
assemblage. The data are grouped by individual bee ID, for which each
individual has information about life history traits (taxonomic
information, inter-tegular distance (proxy for body size), diet breadth)
and values for the 8 experimental trials of the FMPER task (time until
drinking the correct strip, time until first drink, if first choice was
correct, participation). We present our files for organizing data,
creating figures, and analyzing the data. Phylogeny-informed generalized
linear mixed models indicate that neither specialists nor generalists
increased accuracy throughout the task, although both diet breadths became
faster at drinking from the rewarding strip. Bees decreased their drinking
latency—a measure of exploration—throughout the experiment, with no effect
of diet breadth. However, specialists became less likely to participate
over the course of the task compared to generalists. Our results suggest
that specialist and generalist bees have experienced similar selection for
associative learning ability, and that specialists are hesitant to
continue interacting with novel stimuli.