All 60 food items were rank ordered based on WTP from highest to lowest
(Figure 2A). Twenty four of the 30 high-value items were placed in one of four
training conditions in a 2 x 2 design (Go/NoGo x Spaced/Massed, Figure 2B) to
allow for four different comparisons at probe: 1) Massed Go vs. Massed NoGo, 2)
Spaced Go vs. Spaced NoGo, 3) Spaced NoGo vs. Massed NoGo, 4) Spaced Go vs.
Massed Go. Each pair type at probe was composed of nine unique pairs. This item
selection procedure ensured that items that were paired during probe were
equated for WTP such that participants should be indifferent to the choice
between items in each pair given their stated pre-experimental preferences. In
order to ensure that only 25% of all trials were Go trials, in accordance with
standard cue-approach training, we included 24 randomly selected low-value
items (from the bottom 30 items in Figure 2A) as Spaced and Massed NoGo items
during training, but these items (along with the 6 non-selected high-value
items) were never seen during probe. Item assignment (based on rank order
number) to each of the four training conditions (Go/NoGo x Spaced/Massed) for
items that appeared during probe was counterbalanced across participants.