Attention please: Evaluative priming effects in a valent/non-valent categorisation task (reply to Werner & Rothermund, 2013)

Abstract

It has previously been argued (a) that automatic evaluative stimulus processing is dependent upon feature-specific attention allocation (FSAA) and (b) that evaluative priming effects can arise in the absence of dimensional overlap between the prime set and the response set. In opposition to these claims, Werner and Rothermund (2013) recently reported that they were unable to replicate the evaluative priming effect in a valent/non-valent categorisation task. In this manuscript, I report the results of a conceptual replication of the studies by Werner and Rothermund (2013). A clear-cut evaluative priming effect was found, thus supporting the initial claims about FSAA and dimensional overlap. An explanation for these divergent findings is discussed.

Keywords:

• Evaluative priming effect
• Automatic attitude activation
• Attention
• Affective priming
• Automatic affective stimulus processing
• Selective attention

I would like to thank Klaus Rothermund, Benedikt Werner, Jan De Houwer, and one anonymous reviewer for helpful comments on earlier versions of this article.

I would like to thank Klaus Rothermund, Benedikt Werner, Jan De Houwer, and one anonymous reviewer for helpful comments on earlier versions of this article.

Notes

¹ In line with Fiedler, Bluemke, and Unkelbach (2011), I use the term “evaluative priming” rather than the more commonly used term “affective priming”. While the latter term may be used to refer to priming phenomena that concern multiple dimensions of affective meaning, the former is better suited to describe priming along a single, evaluative stimulus dimension (positive vs. negative).

² It is perhaps worth mentioning that I actually ran this experiment twice. In a first attempt, participants completed this experiment after they had already participated in another evaluative priming study (N=76). This experiment produced reliable evaluative priming effects, both in the error rates, F(1, 75)=6.66, p=.01, and the response latency data, F(1, 75)=13.80, p=.0004. As I was worried that prior experience with the evaluative priming paradigm might have contributed to the emergence of these effects, I decided to replicate this experiment using participants that had no prior experience with the evaluative priming paradigm. Reassuringly, the present findings clearly demonstrated that my initial findings were not contingent upon participants having prior experience with the evaluative priming paradigm.