Attribute conditioning: Changing attribute-assessments through mere pairings

Abstract

We propose Attribute Conditioning (AC) as a form of learning that refers to changes in people's assessment of stimuli's (CSs) attributes due to repeated pairing with stimuli possessing these attributes (USs). We review the available evidence and, based on this review, delineate three open questions and investigate them experimentally: a) the moderating role of CS-US similarity; b) the possibility of blocking; and c) the possibility of extinction. Five experiments conditioned health and athleticism. We measured AC effects on direct and indirect dependent variables (direct ratings and semantic misattribution). Experiment 1 shows that CS-US feature similarity does not moderate AC. Experiments 2 and 3 show that AC effects are insensitive to blocking; and Experiments 4 and 5 show that AC effects are resistant to extinction. These five experiments show that AC depends on CS-US contiguity, but not on CS-US contingency. Thereby, the study establishes AC as a simple learning phenomenon describing how people, stimuli, and concepts acquire specific attributes in people's minds due to mere pairings.

Keywords:

• Attribute conditioning
• Similarity
• Extinction
• Blocking
• Evaluative conditioning

Notes

¹We do not strictly exclude one-trial learning; yet in all reported empirical studies, AC effects resulted from repeated pairings.

²We use the term “evaluation” when we refer to evaluation of stimuli's valence, whereas we use the term “attribute-assessment” when we refer to assessment of stimuli's attributes.

³There are some other notable studies showing changes in CSs’ assessments (e.g., Stevenson, Boakes, & Prescott, 1998; Stevenson, Boakes, & Wilson, 2000). Yet, for various reasons they do not fit the present definition of AC.

⁴The unrelated rating task, completed by measurement group only-after in Experiments 2 and 4, constituted a pre-test of USs health used in Experiments 1, 3, and 5.

⁵ CS1-CS2-after and CS1-before-CS2-after analyses are partly redundant as both contain and analyse CS2s ratings (or SMP data) of assessment group only-after. Nevertheless, running both analyses is beneficial as this enables dissociating between different effects.

⁶See note 5 above.

⁷ All-after and before-after analyses are partly redundant as both contain and analyse CSs ratings (or SMP data) of assessment group only-after. Running both analyses is beneficial as they enable dissociating between different effects.

⁸See note 7 above.