The Slips-Of-Action test is listed as a paradigm under the construct of habits in the RDoC matrix. This paper examines the validity of the Slips-Of-Action test as a measure of habits. It is proposed that the construct of modelfree learning, rather than being equal to habits, is actually a possible confounding variable in the Slips-Of-Action-test. it is suggested that stimuli and udfald are likely to be encoded differently, increasing the risk of the Slips-Of-Action test being a measure of model-free learning rather than habits. A new version of the slips-of-action is considered, which provides a better control of the construct of model free learning.