From Acceptance and Commitment Therapy (ACT) and from the Relational Frame Theory (RFT), the defusion is a process that creates a new verbal context that separates the functions of words, diminishing the emotional responses or discomfort with those words. The exercise of “milk- milk-milk” that use ACT is a typical verbal task to get that defusion. Nevertheless, it is not clear what are the mechanisms of acting in this type of verbal exercise. The influence of therapist, his/her instructions and shaping during therapy could be one of those key variables.
We present a comparative study with 4 experimental conditions, in order to test what of them produces defusion and diminishes the discomfort with words. It was used a factorial design (4x2) between the 4 groups and pre-post comparisons. They were measured different variables: descriptive valuation of words, the emotional valuation of words, and time-latency in different sentences self-referred. They had been used a set of phrases positives and negatives, and also a word self-referred that was chosen by the participant. Also, some questionaries’ (AAQ, ERQ, EOSS) were applied before and a[er the exercises. Had participated 105 students randomized in the 4 conditions: Group 1 that made the typical defusion exercise repeating “milk-milk-milk” and the word self-referred; Group 2 received direct instruc4ons by computer about the independence between words and emotions; Group 3 received direct instructions about anti-defusion, that is, assuring the close relationship between words and emotions; and Group 4 as control without verbal exercises.
The results did not show differences between the 4 groups, there were not statistical significance. The discomfort and emotional valuation of sentences did not decrease, not general neither self-referred. They were statistical differences in the pre-post latency, but all the groups showed the same change because of repetition effect.
There arise the question of differences with the Masuda experiments (2004, 2009, 2010), that found to decrease in discomfort of self-referred words. We think our experiment was best controlled, with four conditions including control, measurement pre-post, more number of stimuli, and applied to computer without verbal intervention of the researcher. Our study is a direct replica4on of Masuda experiments, genng beoer the experimental control, but the supposed effects of defusion did not appear.