We would like to thank the authors of all of the datasets we reanalyzed, our two reviewers -- and especially Sascha Meyen for suggesting the absolute effect size test. An R package of our novel tests can be found in github.com/mufcItay/sig.... (14/14)

This was a collaboration with the inspiring group: @matanmazor.bsky.social @liadmudrik.bsky.social @nfaivre.bsky.social. (13/14)

However, many of these effects should be interpreted cautiously due to highly important methodological considerations specific to non-directional tests. Accordingly, we provide best practice recommendations for using non-directional tests. (12/14)

In contrast, non-directional effects were found in *94 out of 136* other null effects from the psychological literature. Hence, we show that 1) true effects may be hidden due to inter-individual variability in effect signs, and 2) this is not the case for unconscious processing effects. (11/14)

The results were highly consistent across tests and datasets: non-significant directional unconscious effects were _not_ masked by variability in effect signs, ruling out non-directional unconscious effects. (10/14)

We also implemented two non-parametric tests of our own: the sign-consistency test and the absolute effect size test. We show with simulations that these tests are more statistically sensitive than existing tests. (9/14)

We used three existing non-directional tests: the qualitative individual differences test (QUID; Rouder & Haaf’s 2019), a global null prevalence test (GNT; Donhauser et al., 2018), and the omnibus ANOVA test (OANOVA; Miller & Schwarz, 2018). (8/14)

We re-analysed 26 previously reported non-significant unconscious effects, using five non-directional tests that ask if effects are reliable within individuals rather than at the group level. (7/14)

(for example, if some slow down and others speed up when the prime matches the target). (6/14)

We noted that the statistical tests employed in these studies may be too stringent. They require non-zero group-level effect, but it is equally interesting if group-level effects are zero because participants are reliably affected by the unconscious stimulus in opposite ways. (5/14)

(e.g., academic.oup.com/nc/article/2...; @meganakpeters.bsky.social @ianbphillips.bsky.social @neddo.bsky.social ) (4/14)

In recent years, methodological concerns and replication failures have led some to argue that consciousness is strictly necessary for some functions, with important implications for the field. (3/14)

Unconscious effects on behaviour are cases where participants report not seeing a stimulus, but their behaviour is affected nonetheless. For example, slowing down or speeding up in certain conditions. (2/14)