Twenty‐five recreational gamblers were initially asked to place bets on either red or black positions on a roulette board in a simulated casino setting. Each participant was then exposed to a stimulus pairing observing procedure which attempted to develop equivalence classes between one color (black or red) and traditionally positive words (e.g., love, happy, sex) and another color (black or red) and traditionally negative words (e.g., death, cancer, taxes), in the absence of consequence manipulations. Twenty‐one of the twenty‐five participants demonstrated greater response allocation to the color position on the roulette board that participated in a relational network with the positive words. Variations in sequencing of experimental conditions had no impact on poststimulus‐pairing wagers, but did impact tests for equivalence accuracy.