The relevance and the pitfalls of replications: a commentary on “The convergent validity and nomological net of two methods to measure retroactive influences” by Matt C. Howard, 2018

This commentary aims to demonstrate that Howard’s (2018) studies are a neither a “close” nor a “conceptual” replications of Bem’s test of retroactive influence (RI) as far as it concerns task 1 and that all tasks in the three studies can be classified as “very far conceptual replication” of a general class of phenomena classified as forced-choice precognition (conscious cognitive awareness) of a future event that could not otherwise be anticipated through any known inferential or perceptual process. As to the results related to the convergent validity (correlation) between the tasks used to test the RI, we posit that without a measure of intra-task reliability any measure of inter-task reliability is uninterpretable.