Introspection - people reporting their subjective experience of their own mental processes - was a favoured technique among psychology's founding fathers. Today, by contrast, it has a poor reputation, often dismissed as unreliable and unscientific. But in a new paper, Sebastien Marti and colleagues argue that introspection can be accurate and illuminating, providing a useful addition to objective measures.
Ten participants completed a simple dual-task paradigm. First they listened to an auditory tone and pressed one of two keys as fast as possible to indicate whether the tone was high or low pitch. Straight after, they pressed one of two keys as fast as possible to indicate whether a 'Y' or 'Z' had subsequently appeared on a computer screen. When the second decision stage comes too soon after the first, reaction times to the second stage are prolonged - an established effect known as 'the psychological refractory period'.
The key twist in this study is that the researchers didn't just record participant reaction times, they also asked them to make several subjective estimates after each trial: how long they'd taken to respond to the tone; how long they'd taken to respond to the letter; how soon the letter appeared after the tone; and whether the letter appeared before or after they'd made their decision about the tone. Reaction times didn't vary on introspection versus control trials, suggesting, importantly, that introspection didn't interfere with the basic cognitive processes required to complete the task.
Participants displayed the usual 'psychological refractory period' and their subjective estimates of their own reaction times and other factors, although under-estimates, were mostly highly correlated with the objective measures. The accuracy of introspection only went awry when the letter appeared too soon after the auditory tone or simultaneously with it. On these trials, not only did participants' reaction times to the letter slow down, it seems they weren't able to start an internal recording of the duration of their reaction time (to the letter) until they had finished processing the tone. Their estimates of the gap between the tone and letter also became inaccurate. It's as if they weren't able to consciously perceive the letter until they'd finished processing the tone. It was a similar story regarding their judgment about whether the letter appeared before or after their auditory decision. Participants were accurate when there was a big enough time delay between tone and letter, but their insight was compromised when the letter appeared too early.
'For the first time, we were able to reconstruct the sequence of conscious events in a psychological refractory period trial based on subjects' introspection,' the researchers said. 'Overall, the present study showed that quantified introspection is a powerful tool. After each trial, participants can answer multiple questions that provide remarkably coherent data which are not always objectively true, but can be used to paint a consistent picture of the subjective phenomenology of an average trial during a cognitive task.'
Marti S, Sackur J, Sigman M, & Dehaene S (2010). Mapping introspection's blind spot: reconstruction of dual-task phenomenology using quantified introspection. Cognition, 115 (2), 303-13 PMID: 20129603