In this study a respondent conditioning procedure was used to pair three arbitrary symbols indirectly with emotive words. Stimuli consisted of words (A), consonant-vowel-consonant nonsense syllables (B) and symbols (C) and were paired B-A and B-C. Participants were then presented with soft drink samples. A pre-tasting preference assessment was conducted in which the participants were asked which drink they would like to taste first, followed by a tasting and taste-test preference assessment. It was hypothesized that the respondent procedure would facilitate the transfer of function and the participants would prefer the positively labelled drink. A matching-to-sample one-to-many technique was used to test for stimulus equivalence (A-C). Twenty-two participants aged between 19 and 59 took part. Results showed that 13/22 participants met the 65% criterion for equivalence with 17/22 preferring to taste the positively labelled drink first. The results of the taste-test showed no significant effect between the three drink labels for the group results. A preference for the positively labelled drink over the neutral and negatively labelled drink was found among those who demonstrated equivalence and for the group results. Participants preferred the positively labelled drinks, followed by the neutral and negatively labelled drinks. Preference for the positive words was also indicated with 18/22 participants preferring the positive word, three preferring the neutral and one preferring the negative word. The current findings are consistent with previous research that suggests a stimulus pairing procedure is suitable for conditioning preference.