Psypost.org reports on a recent study of the processes underlying the Implicit Association Test (IAT). The IAT is the best known method for measuring presumably unconscious attitudes. The critical effect is slower response times when stereotype incompatible stimuli are shown together, like happy words with black faces.
Participants consistently slowed their responses down during the incompatible blocks to ensure they were categorizing the images and words correctly. ... The standard D-score metric conflates [decision ease, response caution, and non-decision time], meaning that the test may be grading people on how cautiously they completed the task rather than measuring their automatic prejudices.
Disclaimer: While I've done some IAT research, I'm not an expert on decomposing IAT findings into different processes, such as the use of racing diffusion models. So, I'll limit this to drawing a parallel with other consciousness research.
In unconscious perceptual processing research, misleading results that look like unconscious processing can occur when the participants adopt a conservative decision criterion. This leads to situations where participants deny seeing a stimulus, yet can correctly identify what they've seen in forced-choice tests. An example from everyday life is that participants will say that they can't see anything on the bottom row of an eye testing chart. However, they can "guess" at the letters better than chance would predict. This conservative reporting is most likely to occur in social settings in which responding can be influenced by the opinions of others. We don't want others to think badly of ourselves, so we censor our responding in various ways.
This IAT report sounds like a similar phenomenon. Conservative decision criteria might impact response time results. This could be deliberate and intentional, not unconscious processing.
The reality is that all of our actions are a complex mix of conscious and unconscious processes. For us scientists, it would be nice to tease this mix apart apart, isolating each component. I'm pessimistic though about reducing these into components. While some investigators call this contamination, I see this mixing of conscious and unconscious processes as being the normal state of affairs.
Results like these could be a step forward in accepting that conscious and unconscious processes are always mixed together, at least at the psychological level.
“In short, an IAT score reflects a blend of factors, and part of that blend is more deliberate than the ‘implicit’ label suggests.”
That's how I see it too. This essay is not an effort to discredit the IAT or claim that it is driven entirely by conscious processing. It's merely pointing out that the situation is complicated.