Does a Working Memory Load Really Influence Semantic Priming? A Self-replication Attempt

Abstract

The present paper describes two attempts to replicate a recent study of ours in the semantic priming domain (Heyman, Van Rensbergen, Storms, Hutchison, & De Deyne, 2015). In that study, we observed that semantic priming for forward associates (e.g., panda-bear) completely evaporated when participants' working memory was taxed, whereas backward (e.g., baby-stork) and symmetric associates (e.g., cat-dog) showed no ill-effects of a secondary task. This was the case for relatively long and short stimulus onset asynchronies (i.e., 1,200 ms and 200 ms, respectively). The results thus suggested that prospective target activation is, contrary to what some theories of semantic memory posit, not an automatic process. However, the two replication studies reported here cast serious doubt on this conclusion. A Bayesian analysis of all the available data indicated that there is at least substantial evidence for a priming effect in every condition, except for forward associates in the short SOA condition. The null hypothesis is still supported in the latter condition, though the replication studies weakened the evidence for a null effect. The theoretical implications of these findings are discussed.Keywords: semantic priming; automaticity; working memory; lexical decision IntroductionThe semantic priming effect is an often studied phenomenon by cognitive psychologists, presumably because it tells us something about the structure of semantic memory (among other things). It refers to the improvement in speed (and accuracy) when responding to a target stimulus that is preceded by a semantically related prime stimulus (see McNamara, 2005, for a review). For instance, people tend to recognize the word dog faster when they first see the related word cat. Over the years, several accounts of the semantic priming effect have been advanced, many of which were inspired by Collins and Loftus' (1975) spreading activation model of semantic processing. According to this model, conceptual knowledge is stored in a network of interconnected nodes, where each node corresponds to a concept (e.g., cat). If one processes a concept, for instance, by reading the word cat, the matching node gets activated and activation will spread to connected nodes, which entails that semantically related concepts such as dog are (partly) activated. A spreading activation mechanism can readily explain the semantic priming effect if one assumes that

The impact of spatial and verbal working memory load on semantic relatedness judgements

“…The WM demands inherent in relatedness judgements appear to make semantic processing in the SRT more sensitive to concurrent WM load. Whereas the LDT is appropriate for exploring the differences between automatic and strategic processes involved in semantic priming (Heyman et al, 2015(Heyman et al, , 2017, the SRT is likely more suitable for studying possible effects of different types of WM load on semantic processing.…”

“…The first experiment was conducted in English, whereas the second experiment was a very precise replication of the original study in Dutch, but with a larger sample. Despite the similarities, Heyman et al (2017) failed to replicate Heyman et al's (2015) crucial three-way interaction between load, relatedness and the type of association. As acknowledged by the authors, a possible reason for the unstable findings may be the non-verbal nature of the dot memory task used to manipulate WM load in both studies.…”

“…Consistent with Hutchison et al (2014), who found a correlation between the level of attentional control and semantic priming, Heyman et al's (2015) results revealed that priming effects were almost eliminated for FA pairs in the high-load condition compared to the low-load condition, whereas priming effects were the same for BA and SYM pairs in both load conditions. Heyman et al (2017) conducted two further semantic priming experiments aiming to replicate Heyman et al's (2015) findings. Both experiments involved the same design and procedure as the original study.…”

“…Some previous studies used the SRT to explore semantic processing in word pairs that were directly or indirectly semantically related (Kuperberg et al, 2008) or that differed in association strength (Ortu et el., 2013), but they did not investigate the impact of WM. Studies that focused on the effect of WM on semantic processing did not distinguish between spatial and verbal domains and used only related and unrelated word pairs (Heyman et al, 2015(Heyman et al, , 2017. To explore the impact of verbal and non-verbal WM tasks on relatedness judgements, the present study manipulated WM type (spatial vs. verbal) and WM load (low vs. high) in an SRT with strongly related (SR), weakly related (WR) and unrelated (UR) pairs.…”

Differential effects of working memory load on priming and recognition of real images

“…Such findings would question accounts of priming that assume automatic activation of target representations. However, a more recent paper by these authors attempted to replicate these findings (Heyman, Goossens, Hutchison, & Storms, 2017) and partly failed to do so when a Bayesian analysis was performed on data from two new experiments. In the replication study there was substantial evidence for priming regardless of load and it was suggested that the non-automaticity account claimed in the previous study could be ruled out.…”