Emotional-Cognitive Overconfidence and Gender


We study the role gender plays in emotional and cognitive overconfidence in a consumer-choice environment. Regardless of age, we find in our sample that males show more emotional and cognitive overconfidence than females do. However, older females are more overconfident in both emotional and cognitive terms. This result can be rationalized by the fact that older females have an edge on their immune system.

Share and Cite:

Maia, G. , Bonfim, D. and Silva, S. (2020) Emotional-Cognitive Overconfidence and Gender. Open Access Library Journal, 7, 1-13. doi: 10.4236/oalib.1106116.

1. Introduction

Consumers choose under incomplete knowledge and perhaps biases, and correspondence is not guaranteed between self-assessed validity of knowledge and actual validity [1]. When they think it is, they display overconfidence [2]. In questionnaires, overconfidence refers to a resulting positive value after subtracting how confident one respondent was in a question from the correct response (and underconfidence refers to a resulting negative value). This overconfidence is cognitive. However, emotional overconfidence may also be involved [3]. Emotional overconfidence refers to excessive confidence while responding to an emotional intelligence test, such as the Consumer Emotional Intelligence Scale described next. Both cognitive and emotional overconfidence share the same reserve of mental energy.

Emotional self-confidence matters because consumer behavior is not based exclusively on cognitive knowledge, but is also influenced by unknown factors that can lead to uncertainty feelings that dampen the perception of cognitive knowledge [1]. As a result, consumers have to overcome such an uncertainty to choose, and those consumers with feelings of doubt, frustration and distrust are less emotionally self-confident [3].

Calibrated consumers are those whose (emotional and cognitive) confidence matches their (emotional and cognitive) ability [3] ; that is, subjective knowledge matches objective knowledge. Knowledge calibration discriminates those consumers who perform better on various tasks [4].

Individuals who show high cognitive ability do not follow their intuitions about puzzles and are also not prone to accept other suggestions from their automatic mind; individuals with low cognitive ability are impulsive, impatient, and keen to receive immediate gratification [5]. And emotional ability refers to the ability-based emotional intelligence [6].

Emotional intelligence can be described as “a set of skills that contributes to the accurate assessment and expression of emotion (in oneself and others), the ability to effectively regulate emotions (in oneself and others), and the ability to use feelings to motivate, plan and achieve” [7].

Kidwell and colleagues [3] showed that emotionally calibrated consumers make healthier food choices regardless of their cognitive calibration and cognitive ability, and then identified as a promising line of future research the study of interaction between emotional and cognitive calibration in consumer choice: to what degree does overconfidence in emotional ability relate to overconfidence in cognitive ability? Because miscalibration may be linked to deficits in metacognitive skill [8] , here we study the interplay between emotional and cognitive overconfidence mediated by a cognitive reflection test [5]. In particular, we consider the role of gender and age.

Section 2 of this article presents the procedures employed in our study and the concepts used. Section 3 shows the results found and discusses them. And Section 4 presents the conclusion.

2. Materials and Methods

2.1. Protocol

We submitted three questionnaires through the app Eval & Go (https://www.evalandgo.com/). The first was the Consumer Emotional Intelligence Scale intended to measure emotional overconfidence; the second was the standard test of cognitive overconfidence; and the third was the Cognitive Reflection Test designed to measure cognitive ability. These will be explained next.

We administered the questionnaires to 115 participants. These were initially sent by email to the authors’ contacts. The link was then resent by the authors’ contacts to their own contacts, by email, Facebook and WhatsApp.

2.2. The Consumer Emotional Intelligence Scale

To measure individual differences in consumers’ ability to use emotional information, we conducted the Consumer Emotional Intelligence Scale (CEIS). This questionnaire of 18 items (available at https://docs.wixstatic.com/ugd/bc96ae_03cfcb92b2404c21a64d0dd63bfb7d1b.pdf) is arguably more appropriate for the consumer-choice environment [6] than the more general Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT) [9]. We applied the test through a link to the site https://www.ceis-research.com/ that was embedded in the Eval & Go routine.

From the 18 questions, the first five refer to emotional perception, involve images and what they transmit to an individual. For example, how a respondent is able to identify feelings of excitement when looking at a dessert menu. These questions aim to assess the ability to identify emotions in oneself and others and interpret emotional information―such as, facial expressions, body posture and color―that is critical for accurate social situational appraisal and effective communication [6].

The next four questions refer to emotional facilitation and aim to track the degree to which an individual considers useful to feel certain emotions in some circumstances. For example, a respondent high in using emotion will know that it is important to feel relaxation when interacting with an aggressive salesperson. These questions gauge the ability to generate, use, and feel emotion in order to focus attention, reason and communicate [6].

The next five questions assess emotional understanding and a participant should infer in a hypothetical situation which emotions can emerge in the situation. An example is given by a participant understanding how her emotions can change throughout a buying process and understand how she will feel before, during and after the purchase. These questions track a respondent’s emotional knowledge about the formation of basic and complex emotions, as well as the anticipation of emotional reactions over time [6].

The last four questions are related to emotional management. They present a situation and ask a respondent to evaluate how a predefined behavior is effective in the situation. An example is that of a respondent who is able to control her impulses when ordering from a menu. These questions measure self-control over emotions in oneself and in interpersonal interactions; they involve regulating the expression and experience of emotions [6].

There are no correct or incorrect responses in the CEIS, of course. A participant only receives a score in each of the four types of consumer emotional intelligence described above: emotional perception, emotional facilitation, emotional understanding and emotional management. And then, he or she receives an average score. By design, a score of 100 means average emotional intelligence; 115 or above means high emotional intelligence; and 69 or below means low emotional intelligence.

Consumers with emotional overconfidence are those with high emotional confidence but low emotional ability [6]. One can make a bold decision because of emotional confidence that ends up as a poor decision because of a lack of emotional ability. To measure emotional over- or underconfidence in the CEIS, a participant is further asked how confident he or she was in the responses after responding to each of the four blocks of the 18-item questionnaire. Here, a participant is considered fully uncertain when responding with a value of 50 percent, while responding with a value of 100 percent means a participant is absolutely certain.

2.3. The Cognitive Overconfidence Test

Cognitive overconfidence arises when cognitive confidence is unrealistically high for a given level of accuracy or when accuracy is unrealistically low for a given cognitive confidence level [10]. One individual is perfectly calibrated when her confidence exactly matches her ability [11]. Cognitive overconfidence has a direct influence on consumer decisions and can have a negative impact on future choices, as it leads a consumer to make riskier decisions that should have been avoided [12].

Cognitive overconfidence can be easily assessed through a questionnaire, and we opted to employ the classical one [11]. The description of this questionnaire in detail is provided elsewhere [13]. After responding to a set of 10 true-false questions, respondents indicated how confident they were in the responses given―from 50 to 100 percent. A value of 50 percent means one respondent had no idea what the correct response was―50 percent is a random guess between the choices. A 100 percent, in turn, means completely confident response. After subtracting how confident one respondent was in each of all 10 questions from the correct responses given (in percentages), one gets a measure of overconfidence, in case of resulting a positive value.

2.4. The Cognitive Reflection Test

To measure cognitive ability, we conducted the cognitive reflection test (CRT) [5]. This is a simple test that measures how individuals differ in cognitive ability in terms of the relative powers of their automatic mind (a.k.a. System 1) and analytical mind (System 2) [12]. Individuals scoring high on the CRT show enhanced ability for using their System 2 to override System 1 tendencies.

The cognitive reflection test contains three questions that are conceived to elicit automatic responses that are compelling but wrong.


1) A bat and a ball cost $1.10 in total. The bat costs $1.00 more than the ball. How much does the ball cost?

_____ cents

2) If it takes 5 machines 5 minutes to make 5 widgets, how long would it take 100 machines to make 100 widgets?

_____ minutes

3) In a lake, there is a patch of lily pads. Every day, the patch doubles in size. If it takes 48 days for the patch to cover the entire lake, how long would it take the patch to cover half the lake?

_____ days

The correct responses are 5, 5, and 47, respectively. Commonly, the intuitive (wrong) responses are, respectively, 10, 100, and 24. We requested a participant to respond to the three questions above in less than 30 seconds. This caveat was taken to ensure an automatic choice was made. We also asked whether the participant already knew one or all of the three questions. He or she was left out from the sample if reporting knowledge of at least one of the questions.

3. Results

We approached 143 potential participants, from whom 115 responded to the three questionnaires in full. However, from these 115 respondents, 25 either knew the CRT or did not respond to it in less than 30 seconds. Thus, these 25 were dismissed when the analysis considered the CRT. The table in the Appendix shows the results.

Because there is no truly correct response in the CEIS, the “correct” response shown in the table in the Appendix refers to the percentage of one participant’s score relative to the difference between the maximum score and the minimum score of all participants in the test.

We first compared emotional overconfidence and cognitive overconfidence for 115 participants as in the table in the Appendix, and then we considered the sample with only the valid responses in the CRT (n = 90). Figure 1 shows the

Figure 1. Overconfidence in the CEIS. Most participants showed emotional overconfidence, as displayed by the points to the right of the 45-degree calibration line. Pearson’s correlation between percent correct and confidence was negligible, that is, −0.011.

correct responses plotted against confidence in the responses given (in percentage) for the CEIS, while Figure 2 shows the same for the cognitive overconfidence test. Both figures refer to the sample of valid CRT responses from 90 participants.

Most participants in Figure 1 exhibited emotional overconfidence; only one showed emotional underconfidence. The same was true of cognitive overconfidence (Figure 2); 55 participants showed cognitive overconfidence, 3 were perfectly calibrated, and 32 exhibited cognitive underconfidence. Therefore, more confidence in the responses given could not be associated with accuracy in the responses for both emotional and cognitive overconfidence. Indeed, for emotional overconfidence, Pearson’s r correlation between percent correct and confidence was negligible, that is, −0.008 for the full sample and −0.011 for the subsample of valid CRT responses. As for cognitive overconfidence, Pearson’s r correlation between confidence in the responses and accuracy was weak too, that is, 0.219 for the full sample and 0.217 for the subsample of valid CRT responses.

To compare emotional and cognitive overconfidence by considering the mediation of cognitive ability, we took the full sample (n = 115) and the subsample of valid CRT responses (n = 90), and then segregated this subsample into two groups of low and high CRT scores. Table 1 shows that emotional and cognitive

Figure 2. Overconfidence in the cognitive overconfidence test. Most participants exhibited cognitive overconfidence, as displayed by the points to the right of the 45-degree calibration line. Pearson’s r correlation between percent correct and confidence was weak, that is, 0.217.

Table 1. Correlation between emotional and cognitive overconfidence.

Note: The z-score refers to Fisher’s r to z transformation.

overconfidence presented a positive correlation regardless of CRT performance. Indeed, r = 0.274 for the full sample, and r = 0.348 for the subsample of valid CRT responses. Table 1 also shows there is no statistically significant difference between those who scored low and those who scored high on the CRT. Therefore, cognitive ability did not matter for the correlation between emotional and cognitive overconfidence.

From the 115 participants, 66 were males and 49 were females. The positive correlation between emotional and cognitive overconfidence did not depend on gender, however (males: r = 0.292; females: r = 0.228). In line with the literature [14] , males showed more cognitive overconfidence than females (male average = 9 percent; female average = 5 percent). Moreover, our study extended this finding to emotional overconfidence, in that males showed more emotional overconfidence than females (male average = 41 percent; female average = 36 percent).

We found a negligible correlation (r = 0.060) between emotional overconfidence and age, and a weak positive correlation between cognitive overconfidence and age (r = 0.192). However, considering age in connection with gender produced interesting results. The older a male, the less overconfident he was. This was true for both emotional and cognitive overconfidence. In contrast, older females tended to be more overconfident than younger ones, in both emotional and cognitive terms (Table 2).

In sum, more confidence in the responses given could not be associated with accuracy in the responses for both emotional and cognitive overconfidence. There was a weak positive correlation between emotional and cognitive overconfidence. And males showed more emotional and cognitive overconfidence than females did. Older males showed less emotional and cognitive overconfidence, and the reverse was true for females.

The biological immune system protects us against illness while the psychological immune system [15] has stress-reducing and antidepressive effects. The psychological and the biological immune systems bolster each other, and both are regulated by the hypothalamic-pituitary-adrenal axis. One downside of the psychological immune system is that it also leads to overconfidence [16] , possibly in both dimensions, cognitive and emotional. Moreover, men are at a disadvantage as they grow older. Estrogen boosts female immune system during the

Table 2. Correlation between emotional and cognitive overconfidence, by age and gender.

premenopausal years while testosterone impairs male immune system all lifelong, a circumstance that helps explain why women live longer. So older females have perhaps a relatively better immune system than older males, but this also brings the side effect of more pronounced female overconfidence. This background can imply that females develop more overconfidence than males as they age.

4. Conclusions

We study emotional and cognitive overconfidence in consumer choice and the role played by gender. Cognitive ability is also considered, but this does not affect the found correlation between emotional and cognitive overconfidence.

More confidence in the questionnaire responses could not be associated with accuracy in the responses given, in both emotional and cognitive terms. Emotional and cognitive overconfidence presented a weak positive correlation. Males displayed more emotional and cognitive overconfidence than females did. That males are more overconfident in emotional terms adds to the literature result that males are more overconfident in cognitive terms. Interestingly, the older a male, the less overconfident he was, in both emotional and cognitive terms. And the reverse was true for females. The older a female, the more overconfident she was. This result can be possibly explained by the fact that older females have an edge on their immune system.


Financial support from CNPq and Capes is acknowledged.

Ethical Committee Approval

This experiment is part of a larger project that is registered at Plataforma Brasil (Comissão Nacional de Ética em Pesquisa) under No. 64758617.2.0000.0121.

Appendix. Responses to the Questionnaires

Conflicts of Interest

The authors declare no conflicts of interest regarding the publication of this paper.


[1] Alba, J.W. and Hutchinson, J.W. (2000) Knowledge Calibration: What Consumers Know and What They Think They Know. Journal of Consumer Research, 27, 123-156. https://doi.org/10.1086/314317
[2] Moorman, C., Diehl, K., Brinberg, D. and Kidwell, B. (2004) Subjective Knowledge, Search Locations, and Consumer Choice. Journal of Consumer Research, 31, 673-680. https://doi.org/10.1086/425102
[3] Kidwell, B., Hardesty, D.M. and Childers, T.L. (2008b) Emotional Calibration Effects on Consumer Choice. Journal of Consumer Research, 35, 611-621.
[4] Gershoff, A.D. and Johar, G.V. (2006) Do You Know Me? Consumer Calibration of Friends’ Knowledge. Journal of Consumer Research, 32, 496-503.
[5] Frederick, S. (2005) Cognitive Reflection and Decision Making. Journal of Economic Perspectives, 19, 25-42. https://doi.org/10.1257/089533005775196732
[6] Kidwell, B., Hardesty, D.M. and Childers, T.L. (2008a) Consumer Emotional Intelligence: Conceptualization, Measurement, and the Prediction of Consumer Decision Making. Journal of Consumer Research, 35, 154-166.
[7] Salovey, P. and Mayer, J.D. (1990) Emotional Intelligence. Imagination, Cognition and Personality, 9, 185-211. https://doi.org/10.2190/DUGG-P24E-52WK-6CDG
[8] Kruger, J. and Dunning, D. (1999) Unskilled and Unaware of It: How Difficulties in Recognizing One’s Own Incompetence Lead to Inflated Self-Assessments. Journal of Personality and Social Psychology, 77, 1121-1134.
[9] Mayer, J.D., Salovey, P., Caruso, D.R. and Sitarenios, G. (2003) Measuring Emotional Intelligence with the MSCEIT V2.0. Emotion, 3, 97-105.
[10] Pulford, B.D. and Colman, A.M. (1996) Overconfidence, Base Rates and Outcome Positivity/Negativity of Predicted Events. British Journal of Psychology, 87, 431-445. https://doi.org/10.1111/j.2044-8295.1996.tb02600.x
[11] Lichtenstein, S. and Fischhoff, B. (1977) Do Those Who Know More Also Know More about How Much They Know? Organizational Behavior and Human Performance, 20, 159-183. https://doi.org/10.1016/0030-5073(77)90001-0
[12] Kahneman, D. (2011) Thinking, Fast and Slow. Farrar, Straus and Giroux, New York.
[13] Zindel, M.L., Menezes, E., Matsushita, R. and Da Silva, S. (2010) Biological Characteristics Modulating Investor Overconfidence. Economics Bulletin, 30, 1496-1508.
[14] Barber, B.M. and Odean, T. (2001) Boys Will Be Boys: Gender, Overconfidence, and Common Stock Investment. Quarterly Journal of Economics, 116, 261-292.
[15] Gilbert, D. (2006) Stumbling on Happiness. Knopf, New York.
[16] Mischel, W. (2014) The Marshmallow Test. Little, Brown Spark, New York.

Copyright © 2023 by authors and Scientific Research Publishing Inc.

Creative Commons License

This work and the related PDF file are licensed under a Creative Commons Attribution 4.0 International License.