The January 2014 issue of Perspectives in Psychological Science features a special section focused on behavioral priming research and attempts at replication.
The five articles included in the special section explore issues including the potential role of moderators in hampering the replication of priming effects and whether direct replications are truly feasible. In addition, researchers discuss the fundamental importance of theory to understanding when, why, and how priming effects occur.
Special Section on Behavioral Priming and Its Replication
Psychological science has recently focused on the replicability of its research, and the field of priming has not escaped this scrutiny. Several failures to replicate priming effects have led some researchers to conclude that these effects are due to type 1 error or are perhaps too sensitive for adequate study. There are many potential moderators of priming effects, but because the field is so young, few well-developed theories identify these moderators. This means that even the smallest change in methodology during a replication attempt can alter the findings. Because of this, researchers should provide evidence of direct replication upon initial publication.
Roberta L. Klatzky and J. David Creswell
The replication failures often seen in priming may be explained by a model of intersensory integration. In this model, sources of information produce estimates, or bids, for the value of an event that, when combined, produce an estimate of an event’s magnitude along the underlying perceptual dimension. The bidding process then interacts with information from memory or inference, leading to additional variability in outcome behavior. The influence on behavior of what may seem to be remote factors could explain why priming effects seem to lack robustness.
Wolfgang Stroebe and Fritz Strack
Recent instances of fraud and replication failure have led many psychological scientists to call for more exact replications. These problems, if indeed they are present, will not be solved through exact replication attempts. Null findings from exact replications are often uninformative and uninteresting, and tell us little about why or when a certain effect might occur. Rather than focusing on exact replications, researchers should instead work to identify and reproduce the same theoretical constructs and mechanisms examined in original studies.
The evidence supporting behavioral priming is overwhelming, with new empirical articles detailing the effect being published each month. The articles in this section help interpret the value of recent failed replication attempts. As the contributors to this special section point out, it is often impossible to exactly replicate the conditions inherent to an original study. Instead of calling into question priming effects in general, nonreplications should serve as a call for greater research into this phenomenon, with specific attention to its boundary conditions and moderators.
Daniel J. Simons
Reproducibility is a cornerstone of science, and if an effect is real and robust, it should be possible to replicate it using the same procedure with adequate statistical power. Although direct replication is a good first step in verifying a result, it is not enough. The argument that unknown moderators are responsible for failures to replicate is problematic, because it allows replication failures to be summarily dismissed. Unless there is evidence for the existence of a moderator, researchers should assume an effect will reproduce if the same procedures are followed. Although failure to replicate an effect does not necessarily disprove a theory, it does add important information regarding the effect’s reliability.