Evaluating the Spidey sense: The metacognition of intuition

doi:10.21203/rs.3.rs-1538076/v1

Download PDF

Article

Evaluating the Spidey sense: The metacognition of intuition

https://doi.org/10.21203/rs.3.rs-1538076/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The notion of intuition is a common phenomenological description of how people feel they make some decisions. Although previous work has demonstrated an objective lab-based method to measure intuition, it remains unknown how such objective data might be related to real-world phenomenology of intuition, such as metacognition. To investigate this, we compared performance in an objective lab-based intuition task and its physiological responses with off-line self-report measures of intuitive thinking. We found that intuitive phenomenology was positively correlated with objective intuition and physiology in the lab-based task. Further, we found that those categorized as ‘intuitive thinkers’ performed better on the lab-based intuition task, relative to ‘rational thinkers’. Interestingly, individuals categorized as neither intuitive nor rational thinkers demonstrated decisional adaptability in their degree of intuition, depending on the availability of conscious evidence. These findings support a link between objective psychophysical lab-based measures, physiology, and subjective phenomenology of intuition in decision making: good metacognition.

intuition

decision-making

consciousness

individual differences

thinking style

Rational-Experiential Inventory

Stories describing so called ‘gut-feelings’ or ‘spidey-sense’ have attracted much interest of late, and consequently helped form the commonly held notion that we can make accurate decisions without carefully evaluating the options – a subjective sensation people often refer to as intuition (Gigerenzer, 2007; Gladwell, 2007). In fact, it has been suggested that gut-feelings or intuition can be preferable in high complexity tasks (Hogarth, 2001; Shirley & Langan-Fox, 1996), with high levels of uncertainty (Khatri & Ng, 2000), or in rapid judgment tasks (Klein, Calderwood, & Clinton-Cirocco, 1986).

A recent study provided robust behavioral, physiological and modelling evidence that differential non-conscious emotional information can provide a rapid boost to decision accuracy, reaction times and even confidence (Lufityanto, Donkin, & Pearson, 2016). Skin conductance measures demonstrated an interaction between the non-consciousness emotional information and conscious decisional evidence. This work used a noisy sensory decision-making task that required subjects to make a simple categorical judgment (left/right) on a snow-storm of moving dots (Newsome, 1997; Shadlen & Kiani, 2013; Vlassova, Donkin, & Pearson, 2014). While making these conscious sensory decisions, subjects were presented with suppressed non-conscious emotional images. The valence of the images was always contingent on the motion direction, e.g. left-negative, right-positive, and the data and modeling suggested that this differential non-conscious emotional information was accumulated alongside the conscious sensory evidence to boost overall performance, without individuals being aware of this process or the emotional information. Importantly, this was not simply an effect of priming, as the emotional information was concurrent and differential (both positive and negative).

These data provide evidence that non-conscious emotional information can be accumulated along with conscious sensory information to boost performance. This dovetails nicely with what many think-of as intuition. However, it remains unknown how such lab-based data, physiology and modeling might relate to real-world, everyday decisions and subjective reports about feelings of intuition. Traditionally, intuition has been investigated by directly asking people about how they remember making their decisions. Do their decisions feel more rational or more intuitive? Based on such self-report data, people can be divided into two primary categories of information processing style (Epstein, Pacini, Denes-Raj, & Heier, 1996). One category is characterized as fully conscious, rational, deliberate, rule-based decision making, or memory of decision making. While the other primary category is described as a non-conscious, intuitive, automatic, and associative-based, decision making style (De Neys, 2006; Evans, 2008). Without objective measures of intuition, it remains unknown how accurate these self-report measures might be. In other words, do we have good metacognition for how intuitive we are? Or are subjective reports of intuition based more on memory, contextual effects at the time of report or some other signal in the brain unrelated to the actual decision-making process.

Our previous study provided evidence for an objective psychophysical lab-based task and novel mechanism of human intuition (Lufityanto et al., 2016). We wondered if this mechanistic lab-based approach might be related to an individual’s self-reported information processing style, based on how an individual reports making everyday real-world decisions. If the boost in decision performance based on non-conscious emotional information (our lab-based task), and subjective phenomenology of intuition rely on common neural information, then the two measures should be related suggesting we have good metacognition of intuition.

To investigate the relationship between our objective sensory decision making ‘intuition task’ and self-reports of everyday decision making styles we utilized the Rational-Experiential Inventory (REI; Epstein et al., 1996) questionnaire to measure the subjective preference for the different modes of thinking. We then compared both behavioral and skin conductance data to individual differences in thinking styles and group differences.

Methods

Participants

Sixty-seven (29 males and 38 females) right-handed, healthy individuals participated in this experiment. This number of participatns was based on our previous study using similar paradigms (Lufityanto et al., 2016), which showed the effect size of the experiments was always larger than 0.45. Taking that into account, we run a priori power analysis which gave us an estimation that 20 participants per cell was sufficient to get the expected effect size (d = .45) with (1 – β) at .95 and α at .05 level. Based on a previous study which utilized the REI questionnaire (Shiloh et al., 2002), there would be potentially three groups created from median split of the REI (See Psychological Questionnaire for more details) Therefore 20 participants x 3 groups= 60 is sufficient for running this study.

All participants were university students who received course credit in exchange for participation. All experiment protocols were approved by the University of New South Wales Institutional Reviewer Board. In addition, the informed written consent was attained according to procedures of the University of New South Wales ethics committee. All methods were carried out in accordance with relevant guidelines and regulations.

We applied an exclusion criteria based on the monotonicity of the sensory decision data, this enabled us to ensure that participants were doing the task as required e.g. processing the decision information (see Lufityanto et al., 2016 for a detailed discussion of this criterium). Rather than using all the motion coherencies we based the monotonicity check only on 6%, 17%, and 32% motion coherence in all experimental conditions. Accordingly, we excluded 7.5 % of participatns in this experiment, who did not show monotonic relationship between decision accuracy and motion coherence. We assume that these subjects did not rely on the given decision information to perform the task. This left a final sample of 62 (28 males and 34 females).

Among total 62 participants whose behavioral data were eligible for further analysis, physiological data was collected from a subset of 48 participants. 14 participants did not undergo physiological measurement due to a technical issue, i.e. limited access to the measurement device.

Visual Stimuli

We used similar visual stimuli to our previous study (Lufityanto et al., 2016). Visual stimuli were presented on a 20” Sony Multiscan G520 CRT monitor (resolution = 1,024 × 768, refresh rate 75Hz). Subjects’ heads were stabilized by a chin rest and positioned 57cm away from the monitor. Stimuli were presented using The Psychophysics Toolbox, version 3, for MATLAB (Kleiner et al., 2007) on a Macintosh MacPro running Mac OS X.

Our random dot motion (RDM) stimuli consisted of 60 white dots (each a 2 x 2 pixel square) moving inside a black background circular aperture with diameter of 3.98°. The motion coherence was pseudo-randomly selected from a pool of 6 coherence levels (6%, 11%, 17%, 24%, 32% and 39%) and the motion direction was equally distributed between left and right directions.

Emotional images were retrieved from International Affective Picture System (IAPS) (For a review, see Lang, Bradley, & Cuthbert, 1999) presented at 3.54 x 3.86°. Six images illustrating positive events (e.g. baby, puppy) and another six images with negative events (e.g. gun, snake) in total were used in our experiments. Mean pleasure ratings for the positive and negative images were 6.98 and 3.05 respectively, and mean arousal ratings were 5.04 and 6.37 for positive and negative images respectively based on the IAPS norms (Lang et al., 1999). Two different images for each valence were used in each block to prevent any effect of image-specific learning.

To form a baseline condition we used a spatial phase-scrambling technique (Maudsley, 1988), we removed semantic content of the images while at the same time maintaining the low-level image characteristics such as contrast, color and spatial frequency. The scrambled images acted as baseline condition relative to the intact emotional images, as no meaningful information can be retrieved post scrambling

The emotional and the phase-scrambled images were transformed into sepia color, so they were easily distinguished from the continuous flash suppression (CFS) stimuli – intentionally designed to suppress visual objects from conscious perception (Tsuchiya & Koch, 2004). Using a mirror stereoscope, CFS and the emotional images presented to each eye, such that they overlapped at a single spatial location allowing CFS to suppress the emotional images from awareness. The 10Hz CFS stream comprised of colored rectangles, triangles and circles in 6 bright colors not including the sepia or any color associated with it (i.e. black, brown, orange, and dark green). The size of CFS stimulus was 4.37° x 4.77° of visual angle, slightly bigger than the images to cover the whole area, including the image edges.

Physiological Measurement

We measured skin conductance response (SCR) as in our previous study (Lufityanto et al., 2016), which found a link between SCR score and the utilization of non-conscious emotional information in the decision task. SCR was recorded using the ADInstruments PowerLab 16/30 system, following the standardized published guidelines (ADInstruments, 2009). Electrodes were placed on the middle phalanges of the index and the second fingers of left hand (i.e. non-dominant).

Following our previous study procedure, we calculated SCR mean value starting from stimuli onset up to 6000ms after the response to capture the entire dynamics of the SCR. We then binned the individual SCR values per 400ms window for each coherence. Outlier values (which fell outside mean ±2.5 SD) were removed; we surmised that they might be due to confounding events, e.g. random musculoskeletal response or any response which is not related to our intended experimental manipulation.

Psychological Questionnaire

We used the Rational-Experiential Inventory (REI) which was constructed based on Cognitive-Experiential Self Theory (CEST; Epstein, 1998). This theory proposes there are two separate cognitive systems used in information processing: (i) rational system, characterized as analytical-based, deliberate, and time-consuming, and (ii) experiential system which is more intuitive, automatic, and rapid.

We used the REI because it has been largely utilized in studies that attempt to measure the dual processing account of human behavior (Shiloh, Salton, & Sharabi, 2002; Sladek, Bond, & Phillips, 2010).

The REI consists of two subscales: Need for Cognition which is aimed to measure individual preference toward rational thinking style (e.g. “I would prefer complex to simple problems”), and Faith of Intuition, which aims to assess susceptibility toward experiential/intuitive thinking (e.g. ‘When it comes to trusting people, I can usually rely on my ‘gut-feelings’’). Subjects were asked to rate how accurately each statement describes themselves using 5-point Likert response (1 = strongly disagree and 5 = strongly agree). We used the latest version of REI (Pacini & Epstein, 1999), which consists of 40 items that distinguish between engagement and ability for both rational and experiential dimensions. The latest REI version comprises four sub-scales: Rational Ability, Rational Engagement, Experiential Ability, and Experiential Engagement –each consisting of 10 items. The ability sub-scales refer to the capacity of an individual to apply certain dimensions, whereas the engagement subscale refers to a reliance on certain dimensions. The composite rational and experiential scores are simply computed by summing scores for the ability and engagement sub-scales (Hodgkinson, Sadler-Smith, Sinclair, & Ashkanasy, 2009).

In Experiment 1 we split subjects into two groups, i.e. rational and intuitive thinkers, following previous study procedures (Shiloh et al., 2002). We took the median score from each rational and experiential scale independently. The median scores from our empirical data were 3.75 and 3.2 respectively for rational and intuition sub-scales. This gives two categories for each scale: high vs low scores. ‘Rational thinkers’ are defined as those who scored high in the rational scale and low in the experiential scale. Intuitive thinkers are those who score low in rational scale and high in the experiential scale. With this procedure, we ended up with 18 rational thinkers and 18 intuitive thinkers for. The other 26 subjects either scored high in both scales, or low in both scales, referred as non-differentiated (Shiloh et al., 2002). Using exact the same procedure, among 48 participants who underwent physiological assessment, we ended up 14 rational thinkers, 14 intuitive thinkers and 20 non-differentiated thinkers. Please note that the median scores of rational and intuitive sub-scales did not change. We further analyzed both behavioral data and physiological data in three group subjects: intuitive thinker, rational thinker, and non-differentiated thinker separately.

Procedure

The experiment consists of two tasks in which subjects were required (i) to perform sensory decision-making task first and (ii) to fill out REI questionnaire afterward. The order of this task was always constant for all subjects. REI questionnaire was put last simply to keep subjects focused on the sensory decision task.

The procedure largely followed our previous study (Lufityanto et al., 2016). Subjects were presented with RDM stimuli and simultaneously accompanied by emotional images suppressed by CFS stimuli for 400ms (see Fig. 1). Subjects were instructed to respond signaling the global direction of the RDM stimuli by pressing left/right arrow buttons on the computer keyboard as fast and accurately as they could. The binary emotional valence of the images (positive or negative) was 100% contingent with the direction of the RDM (left or right) across 6 levels of motion coherence. The contingency was counterbalanced across subjects to prevent specific direction-valence bias.

Further, subjects were asked to press the spacebar whenever they saw any sepia color at the CFS location, i.e. a break in suppression. The trials with suppression breaks were immediately stopped and then randomly reintroduced amongst future trials in the same block. Only trials without suppression breaks were included in the analyses.

As assurance that a participant’s criteria for reporting incidental suppression breaks was reliable, we included catch trials, intermixed with the experimental trials. In the catch trials, the contrast level of CFS stream was lowered, therefore inducing image suppression breaks. We used novel neutral images for the catch trials to preserve awareness of the association between dot motion direction and emotional valence in the normal trials. Only subjects who responded >90% to the catch trials across the 3 blocks, were included in the analysis, similar to the threshold applied in our previous study (Lufityanto et al., 2016).

In total, participants completed 3 blocks of trials. Each block consisted of 104 trials, 48 emotional intact condition, 48 trials of the emotional phase scrambled, and 8 catch trials. The trials were presented to subjects in pseudo-randomly order.

At the end of the experiment, participants were given a separate suppression test in which all images used in the experiment were again presented to subjects, suppressed by CFS, without the moving dot stimuli. Subjects were asked to indicate whether those images contained an object (emotional or neutral images) or an abstract pattern (scrambled images). Skin conductance response was recorded during performance across three blocks of experiment and the suppression test.

Results

We replicated our previous finding (Lufityanto et al., 2016) that non-conscious emotional information aided decision making, indicated by higher decision accuracy in the intact suppressed emotional image condition compared to a phase-scrambled image condition, with no emotional information (F_(1,61)=11.986, p=.001, h_p²=.164; Fig. 2a). The utilization of emotional images in the conscious random dot decision was also interacted with the motion coherency (F_(5,57)=9.496, p=.000, h_p²=.135). Across all participants, reaction times for the intact emotional images were faster than for the scrambled images (F_(1,61)=6.352, p=.014, h_p²=.094; Fig. 2b). This finding replicates our previous study (Lufityanto et al., 2016).

Next, we wanted to test if there was a relationship between our quantitative psychophysical measure and the Rational-Experiential Inventory (REI), commonly used to assess subjective intuitive thinking styles. Across all the conditions in these data, we see the largest difference in decision accuracy between emotional and scrambled images for the lower coherence, more difficult decisions. This makes sense as there is less conscious decision relevant sensory information in the low coherent trials, inducing a reliance on the non-conscious emotional information. Accordingly, we compared the gap in decision accuracy between the intact emotional images and phase scrambled-image trials for the low coherence (6%, 11%, 17% coherency) trials to different subsets of Rational-Experiential Inventory. We subtracted the score in adjacent coherencies between two conditions (e.g. score in 6% coherency intact emotional images condition was subtracted with score in 6% coherency scrambled images), before taking the average score for each subject.

The intuitive scale subset was positively correlated with the gap in decision accuracy between emotional and scrambled images for the low coherencies (r=.601, p<.000; Fig. 2c), while negative correlation was found in the subset rational scale (r=-.283, p=.027; 2d). When the pool of subjects was split using the median of the REI scale (on both intuitive and rational scale subsets; Supp. Table. 1), into three groups (See Psychological Questionnaire for the median split method), one group, referred to as ‘intuitive thinkers’, showed a large boost in decision accuracy in the non-conscious emotional image condition, across all coherences (F_(1,17)=34.744, p=.000, h_p²=.671; Fig. 3a). In contrast, no main effect was observed for the ‘rational thinkers’ (F_(1,17)=.214, p=.65, h_p²=.012; Fig. 3b). The interaction between emotional images and the motion coherence was absent in both intuitive thinkers (F_(5,13)=2.038, p=.081, h_p²=.107) and rational thinkers (F_(5,13)=1.993, p=.088, h_p²=.105).

Next, we calculated the decision accuracy for all 26 ‘non-differentiated thinkers’, before then splitting them in two sub-group of participants who: (i) scored high in both rational and intuitive sub-scales, and (ii) scored low in both two sub-scales. Among all the 26 ‘non-differentiated thinkers’, we found no boost in decision accuracy in the non-conscious emotional image condition (F_(1,25)=1.056, p=.314, h_p²=.041; Supp. Fig. 1), but there was an interaction between the utilization of non-conscious emotional information and the motion coherency (F_(5,21)=5.701, p=.000, h_p²=.186), such that participants tended to only use the non-conscious emotional information when the motion coherency was low, i.e. indicating lack of related evidence to initiate decision. Further, this interaction effect was apparently largely driven by participants who scored high in both rational and intuitive sub-scales (F_(5,8)=14.605, p=.000, h_p²=.549; Fig. 3c). There was no interaction effect found in the participants who scored low in both rational and intuitive sub-scales (F_(5,8)=0.337, p=.088, h_p²=.027; Fig. 3d). In addition, neither participants who scored both high (F_(1,12)=1.166, p=.301, h_p²=.089) and low (F_(1,12)=.189, p=.671, h_p²=.016) in the two subscales showed differential non-conscious emotional effect relative to the phase-scrambled image condition in the conscious decisions.

Next, we calculated the decisional reaction time based on thinking style. The ‘intuitive thinkers’ made faster decisions in the intact emotional image condition compared to the phase-scrambled condition (F_(1,17)=16.684, p=.001, h_p²=.495; Supp. Fig. 2a). However, this effect was not observed for (i) the ‘rational thinkers’ (F_(1,17)=.572, p=.460, h_p²=.033; Supp. Fig. 2b), (ii) ‘non-differentiated thinkers’ who scored high in both subscales intuition and rational (F_(1,12)=1.022, p=.332, h_p²=.078; Supp. Fig. 2c), and (iii) the other type of ‘non-differentiated thinkers’ who scored low in both subscales (F_(1,13)=.591, p=.457, h_p²=.047; Supp. Fig. 2d). Thus, our data provides evidence that the utilization of non-conscious emotional information, reflected by more rapid decisional time, is predicted by an individual’s self-reported thinking style.

Among the 48 participants for whome we collected physiologica data, we calculated the skin conductance response (SCR) while performing the decision task. First, we ran a correlation between the average score of SCR and the decision accuracy on the low motion coherences (i.e. 6%, 11%, and 17%) and found that that those two variables were positively correlated (r =.256, p=.044, Fig. 4a). We found that SCR for the emotional image condition was significantly higher than the phase-scrambled condition (F_(1,47)=9.904, p=.003, h_p²=.197, Fig. 4b), but did not interact with the motion coherency or task difficulty in the conscious decisions (F_(5,43) =2.126, p=.063, h_p²=.043). When splitting into the different thinking style groups the skin conductance data also showed a significant difference between the emotional and scrambled image conditions for ‘intuitive thinkers’ (F_(1,13)=11.495, p=.008, h_p²=.561; Fig. 4c). However, this effect was not significant for (i) the ‘rational thinkers’ (F_(1,13)=0.827, p=.380, h_p²=.060; Fig. 4d), (ii) ‘non-differentiated thinkers’ who scored high in both subscales intuition and rational (F_(1,9)=1.729, p=.220, h_p²=.162; Fig. 4e), and (iii) the other type of ‘non-differentiated thinkers’ who scored low in both subscales (F_(1,9)=3.763, p=.084, h_p²=.0295; Fig. 4f).

Next, we calculated the gap in accuracy between intact and phase scrambled emotional images across all 6 motion coherence levels for each subject, then averaged across all coherences resulting a single value for decision accuracy, each for intact and phase-scrambled image conditions. We did a similar procedure for the reaction time data. We found that decision accuracy was negatively correlated with reaction time (t_(1,21)=-.294, p=.020; Supp. Fig. 3) indicating that perhaps ‘intuitive decisions’ as we are defining them here, would not be exhibited if people took longer to make decisions. Importantly, this hypothesis links in with the common notion that intuitive decisions are associated with rapid ‘gut-based’decisions.

To further test the prediction that an intuitive decision strategy would be replaced by a more rational cognitive approach when time constraints are removed, we conducted Experiment 2 in which the decision stimulus stayed on the screen until participants responded. In our previous experiments, stimuli were presented for 400ms (i.e. rapid) and were then removed. Without such time constraints, participants should have more time to accumulate the decisional evidence and therefore not resort to “intuition” to tap into the non-conscous emtoinal information.

Methods

Forty-eight (22 males and 28 females) first-year university students participated in this experiment in exchange for a credit course. Further we excluded 8% of subjects due to non-monotonic exclusion criteria, resulting a final sample of 44 (19 males and 25 females).

The procedure was similar to that in experiment 1. The only difference was that the stimuli remained on the screen until participants made a decision. By extending the stimuli presentation duration, participants should be able to collect more sensory decisional evidence. However, we did limit the stimuli display time to a maximum of 1000ms, to prevent familiarity of the suppressed images. If the subjects did not respond within the 1000ms time window, the trial was stopped and reintroduced back in amongst later trials.

Next, we split participants using REI scale following a similar procedure to that in Experiment 1, resulting in the same proportion of participants in each group: 13 rational thinkers, 13 intuitive thinkers, and 18 non-differentiated thinkers. Among the total non-differentiated thinkers, 9 participants scored high on both intuitive and rational subscale, and the other 9 participants scored low on both sub-scales (See Supp. Table 2 for the categorization).

Results

As we predicted, participants in Experiment 2 took longer to respond (See Supp. Fig. 4). Importantly, the longer reaction times did not affect breaks in suppression as the incidental rate of suppression breaks remained below 5% across all subjects (See Supp. Fig. 5).

For all participants, unlike the time-limited data these longer response time data did not show a difference in accuracy between intact emotional and scrambled image condition (F_(1,43)=3.642, p=.063, h_p²=.078; Fig. 5a). We also found no difference in the reaction time between the two conditions (F_(1,43)=.551, p=.462, h_p²=.013; Fig. 5b). However, we wondered whether the influence of thinking style disposition in decisions could still be observed in a situation in which the use of ‘intuition’ was not necessarily needed.

Looking at the data from Experiment 2 split by thinking style, suggests that non-conscious emotional information still boosted accuracy for ‘intuitive thinkers’, even with the longer reaction times (F_(1,12)=17.565, p=.001, h_p²=.594; Fig. 6a). While the ‘intuition effect’ was not observed for (i) the ‘rational thinkers’ (F_(1,12)=4.412, p=.058, h_p²=.296; Fig. 6b), (ii) ‘non-differentiated thinkers’ who scored high in both subscales intuition and rational (F_(1,8)=1.143, p=.316, h_p²=.125; Fig. 6c), and (iii) the other type of ‘non-differentiated thinkers’ who scored low in both subscales (F_(1,8)=2.766, p=.135, h_p²=.0257; Fig. 6d).

The evidence for the utilization of non-conscious emotional information was also present in the skin conductance response data, exclusively for the ‘intuitive thinkers’ (F_(1,13)=6.975, p=.022, h_p²=.368; Fig. 6e). In contrast, we did not find the same effect for (i) the ‘rational thinkers’ (F_(1,13)=.704, p=.418, h_p²=.055; Fig. 6f), (ii) ‘non-differentiated thinkers’ who scored high in both subscales intuition and rational (F_(1,8)=1.476, p=.259, h_p²=.156; Fig. 6g), and (iii) the other type of ‘non-differentiated thinkers’ who scored low in both subscales (F_(1,8)=.625, p=.452, h_p²=.072; Fig. 6h).

Here across two different experiments, using validated questionnaires, our data suggests that our objective lab-based psychophysical measure of benefit from non-conscious emotional information or intuition, can predict an individual’s self-reported thinking style. Specifically, both the behavioral and SCR data were related to self-reported intuitive thinking styles. Further, when time-limitations were removed, only intuitive thinkers utilized the non-conscious emotional information. These findings are interesting as they suggest we have good general metacognition for intuition. How we feel we make everyday decisions (from the questionnaire data) is related to how well we utilize non-conscious emotional information for concurrent conscious decisions.

In Experiment 1, where the amount of task-related evidence was restricted to 400ms, we found non-conscious emotional information boosts decision accuracy for ‘intuitive’ and ‘flexible thinkers’. Likewise, objective skin conductance measures also supported the utilization of non-conscious emotional information within those two groups, but not for rational thinkers. In Experiment 2, when participants were given flexibility to sample more sensory evidence over longer periods, the effects were reduced and only intuitive thinkers benefited from the non-conscious information.

These data also provide insight into the characteristics of ‘rapid’ intuitive decision making (Evans, 2008; Hogarth, 2001; Klein et al., 1986). We demonstrated that restricted time to collect task-related decision evidence promotes intuitive decisions. When the amount of task-related time is reduced, people tend to rely more on non-conscious emotional information, as our previous study suggested (Lufityanto et al., 2016). Emotional information can be processed rapidly due to affect salience and possibly efficient neural-architecture (Liddell et al., 2005; Vuilleumier et al., 2002). On the contrary, when conscious decisional information is abundant, emotional information is perhaps not required or simply ignored.

Conceptual frameworks, such as Cognitive-Experiential Self Theory (CEST: Epstein, 1998) and Dual-Theory of Information Processing (Evans, 2008) have proposed two cognitive systems, namely rational and experiential. Our study has demonstrated distinct behavioral and physiological evidence for a dissociation of the two systems. While ‘intuitive thinkers’ utilized the non-conscious emotional information, the ‘rational thinkers’ did not to the same degree.

While static trends of behavior were exhibited by both ‘intuitive thinkers’ and ‘rational thinkers’, ‘non-differentiated thinkers’ who scored high in both rational and intuitive sub-scales tended to change their use of the non-conscious information across the two experiments. In Experiment 1, when the stimuli duration was short, this specific sub-sample showed an interaction effect between motion coherence and image condition. Non-conscious intact emotional images were advantageous only when the task was difficult. In contrast, when the sensory evidence was more available, this sub-sample seemed to no longer use the non-conscious emotional information. Although previous studies (e.g. Shiloh et al., 2002) have already acknowledged the presence of the ‘non-differentiated thinkers’, our pilot study shows that these participants have the capability to change the sources of information informing the particular decision task.

Importantly, the limitations of this study include the number of participants and future studies should replicate these findings. It is however worth noting that some of our findings were previously documented and indeed replicated multiple times (Lufityanto, Donkin, & Pearson, 2016).

In summary, our study provides a congruency between behavioral, physiological and self-report data for the ‘intuitive thinkers’, further supporting our proposed framework: that the productive utilization of non-conscious emotional information in decisions – as a basic ingredient of intuitive decisions – can predict everyday thinking styles. Or to put it another way, our study suggests a proposition that humans may have good metacognition for their intuition. Therefore, our lab-based objective paradigm complements self-report measures of intuition, but offers the possibility of measuring intuition objectively, without an individual knowing the purpose of the test. A more objective approach may address some limitations of self-report techniques, such as social desirability (Arnold & Feldman, 1981; Van de Mortel, 2008) and any other contextual biases (Adams, Soumerai, Lomas, & Ross-Degnan, 1999).

Author contributions: GL and JP designed the studies; GL collected the data and analysed the it; GL and JP wrote the paper.

Acknowledgements. This work was supported by Australian NHMRC grants GNT1046198 and GNT1085404 and ARC discovery projects DP140101560 and DP160103299. JP was supported by an NHMRC Career Development Fellowship GNT1049596.

Availability of Data and Materials

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Adams, A., Soumerai, S. B., Lomas, J., & Ross-Degnan, D. (1999). Evidence of self-report bias in assessing adherence to guidelines. International Journal for Quality in Health Care, 11(3), 187–192.
ADInstruments, P. L. (2009). ADInstrument PowerLab/30 Series Owner's Guide. Australia: ADInstruments Pty Ltd.
Arnold, H. J., & Feldman, D. C. (1981). Social desirability response bias in self-report choice situations. Academy of Management Journal, 24(2), 377–385.
Boyle, G. J. (1995). Myers-Briggs Type Indicator (MBTI): Some Psychometric Limitations. Australian Psychologist, 30(1), 71–74.
Choi, I., Koo, M., & Choi, J. A. (2007). Individual differences in analytic versus holistic thinking. Personality and Social Psychology Bulletin, 33(5), 691–705.
De Neys, W. (2006). Dual processing in reasoning two systems but one reasoner. Psychological Science, 17(5), 428–433.
Epstein, S. (1998). Cognitive-experiential self-theory Advanced personality (pp. 211–238): Springer.
Epstein, S., Pacini, R., Denes-Raj, V., & Heier, H. (1996). Individual differences in intuitive–experiential and analytical–rational thinking styles. Journal of personality and social psychology, 71(2), 390.
Evans, J. S. B. T. (2008). Dual-processing accounts of reasoning, judgment, and social cognition. Annu. Rev. Psychol., 59, 255–278.
Franken, I. H., & Muris, P. (2005). Individual differences in decision-making. Personality and Individual Differences, 39(5), 991–998.
Gigerenzer, G. (2007). Gut Feelings. Short Cuts to Better Decision Making. New York: Penguin Books.
Gladwell, M. (2007). Blink: The power of thinking without thinking: Hachette Digital, Inc.
Hodgkinson, G. P., Sadler-Smith, E., Sinclair, M., & Ashkanasy, N. M. (2009). More than meets the eye? Intuition and analysis revisited. Personality and Individual Differences, 47(4), 342–346. doi: http://dx.doi.org/10.1016/j.paid.2009.03.025
Hogarth, R. M. (2001). Educating Intuition. Chicago: The University of Chicago Press, Ltd.
Khatri, N., & Ng, H. A. (2000). The role of intuition in strategic decision making. Human Relations, 53(1), 57–86.
Klein, G. A., Calderwood, R., & Clinton-Cirocco, A. (1986). Rapid decision making on the fire ground. Paper presented at the Proceedings of the Human Factors and Ergonomics Society annual meeting.
Kleiner, M., Brainard, D., Pelli, D., Ingling, A., Murray, R., & Broussard, C. (2007). What’s new in Psychtoolbox-3. Perception, 36(14), 1.1–16.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1999). International affective picture system (IAPS): Technical manual and affective ratings: Gainesville, FL: The Center for Research in Psychophysiology, University of Florida.
Liddell, B. J., Brown, K. J., Kemp, A. H., Barton, M. J., Das, P., Peduto, A.,.. . Williams, L. M. (2005). A direct brainstem–amygdala–cortical ‘alarm’system for subliminal signals of fear. Neuroimage, 24(1), 235–243.
Lufityanto, G., Donkin, C., & Pearson, J. (2016). Measuring Intuition Nonconscious Emotional Information Boosts Decision Accuracy and Confidence. Psychological Science, 27(5), 622–634.
Maudsley, A. (1988). Dynamic range improvement in NMR imaging using phase scrambling. Journal of Magnetic Resonance (1969), 76(2), 287–305.
Newsome, W. T. (1997). Deciding about motion: Linking perception to action. Journal of Comparative Physiology, 181(1), 5–12.
Pacini, R., & Epstein, S. (1999). The relation of rational and experiential information processing styles to personality, basic beliefs, and the ratio-bias phenomenon. Journal of personality and social psychology, 76(6), 972.
Pittenger, D. J. (1993). Measuring the MBTI… and coming up short. Journal of Career Planning and Employment, 54(1), 48–52.
Shadlen, M. N., & Kiani, R. (2013). Decision making as a window on cognition. Neuron, 80(3), 791–806.
Shiloh, S., Salton, E., & Sharabi, D. (2002). Individual differences in rational and intuitive thinking styles as predictors of heuristic responses and framing effects. Personality and Individual Differences, 32(3), 415–429.
Shirley, D. A., & Langan-Fox, J. (1996). Intuition: A review of the literature. Psychological Reports, 79(2), 563–584.
Sladek, R. M., Bond, M. J., & Phillips, P. A. (2010). Age and gender differences in preferences for rational and experiential thinking. Personality and Individual Differences, 49(8), 907–911.
Tsuchiya, N., & Koch, C. (2004). Continuous flash suppression. Journal of Vision, 4(8), 61–61.
Van de Mortel, T. F. (2008). Faking it: social desirability response bias in self-report research. Australian Journal of Advanced Nursing, The, 25(4), 40.
Vlassova, A., Donkin, C., & Pearson, J. (2014). Unconscious information changes decision accuracy but not confidence. Proceedings of the National Academy of Sciences, 111(45), 16214–16218.
Vuilleumier, P., Armony, J., Clarke, K., Husain, M., Driver, J., & Dolan, R. (2002). Neural response to emotional faces with and without awareness: event-related fMRI in a parietal patient with visual extinction and spatial neglect. Neuropsychologia, 40(12), 2156–2166.

No competing interests reported.

SupplementaryMaterialslatest.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Evaluating the Spidey sense: The metacognition of intuition

Status:

Version 1

Abstract

Figures

Introduction

Experiment 1

Experiment 2

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1