Individual differences and cognitive reflection across gender and nationality the case of the United Arab Emirates

Mallek, Ray Saadaoui; Al-Baity, Mohamed Article Individual differences and cognitive reflection across gender and nationality the case of the United Arab Emirates Cogent Economics & Finance Provided in Cooperation with: Taylor & Francis Group Suggested Citation: Mallek, Ray Saadaoui; Al-Baity, Mohamed (2019) : Individual differences and cognitive reflection across gender and nationality the case of the United Arab Emirates, Cogent Economics & Finance, ISSN 2332-2039, Taylor & Francis, Abingdon, Vol. 7, Iss. 1, pp. 1-14, https://doi.org/10.1080/23322039.2019.1567965 This Version is available at: https://hdl.handle.net/10419/245194 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=oaef20 Cogent Economics & Finance ISSN: (Print) 2332-2039 (Online) Journal homepage: https://www.tandfonline.com/loi/oaef20 Individual differences and cognitive reflection across gender and nationality the case of the United Arab Emirates Ray Saadaoui Mallek & Mohamed Albaity | To cite this article: Ray Saadaoui Mallek & Mohamed Albaity | (2019) Individual differences and cognitive reflection across gender and nationality the case of the United Arab Emirates, Cogent Economics & Finance, 7:1, 1567965, DOI: 10.1080/23322039.2019.1567965 To link to this article: https://doi.org/10.1080/23322039.2019.1567965 © 2019 The Author(s). This open access article is distributed under a Creative Commons Attribution (CC-BY) 4.0 license. Published online: 31 Jan 2019. Submit your article to this journal Article views: 841 View related articles View Crossmark data Citing articles: 1 View citing articles FINANCIAL ECONOMICS | RESEARCH ARTICLE Individual differences and cognitive reflection across gender and nationality the case of the United Arab Emirates Ray Saadaoui Mallek 1* and Mohamed Albaity 1 Abstract: This study aims to investigate the effect of gender differences and nationality on the Cognitive Reflection Test (CRT), as well as behavioural biases. A sample of 770 questionnaires was collected from undergraduate business students in public and private universities across the United Arab Emirates. The results suggested that low CRT values were dominant in males and females alike for nonUAE citizens in the age group of 18–22 years old, of whom at least one of the parents had a college degree. Subjects with higher cognitive reflection scores were significantly more likely to exhibit overconfidence, risk preference, and risk illiteracy. In addition, females scored lower than males, and non-UAE citizens scored higher than local citizens in the CRT. However, further analysis of the interaction between gender and nationality revealed that local females scored on average higher than non-UAE females. Subjects: Middle East Studies; Gender Studies; Individual Differences/IQ; Cognition & Emotion; Behavioral Psychology; Investment & Securities Keywords: Cognitive Reflection Test; gender effects; dual processes; risk literacy; risk preference ABOUT THE AUTHORS Dr. Ray Saadaoui Mallek obtained her Ph.D. from the Faculty of Economics and Management of Sfax, Tunisia. Currently, she is a Lecturer of finance in the College of Business Administration at the University of Sharjah, UAE. Her research interests are Behavioural finance, Financial Markets, Portfolio Management, Financial Stability, and finance. Dr. Mohamed Albaity obtained his Ph.D. from the University of Malaya, Malaysia. Currently, he is an Assistant Professor of finance in the College of Business Administration at the University of Sharjah, UAE. His research interests are Behavioural finance, macroeconomics, the stock market, Islamic banking, and finance. The authors work in behavioral finance is more focused on the emerging markets and developing economies. They believes that the area of behavioral finance is challenging in these markets and it requires considerable investigation to be recognized. PUBLIC INTEREST STATEMENT The cognitive Reflection Test is a commonly used test to measure to what extent and how individuals are able to override their initial decision and move to a more rational decision when faced with problems. Our study investigates the disparity in gender and nationality on the Cognitive Reflection Test (CRT), as well as behavioural biases based on 770 questionnaires collected from undergraduate business students in public and private universities across the United Arab Emirates. The results suggested that subjects with higher cognitive reflection were significantly more likely to exhibit overconfidence, risk preference, and risk illiteracy. In addition, females scored lower than males, and non-UAE citizens scored higher than local citizens in the CRT. However, further analysis of the interaction between gender and nationality revealed that local females scored on average higher than non-UAE females. Mallek & Albaity, Cogent Economics & Finance (2019), 7: 1567965 https://doi.org/10.1080/23322039.2019.1567965 © 2019 The Author(s). This open access article is distributed under a Creative Commons Attribution (CC-BY) 4.0 license. Received: 08 July 2018 Accepted: 07 January 2019 First Published: 10 January 2019 * Corresponding author: Ray Saadaoui Mallek, University of Sharjah, Sharjah, United Arab Emirates E-mail: [email protected] Reviewing editor: David McMillan, University of Stirling, UK Additional information is available at the end of the article Page 1 of 14 1. Introduction The assumption of rationality in the classical economic school of thought predicts that if the economy is in a downturn it will automatically adjust and normalise the negative impact. Similarly, if the economy is in an upswing it is not possible for investors to make abnormal profits in the asset market. Another theory that manages to explain why these outcomes occur is related to behavioural economics. Behavioural economics has shed light on this issue by combining psychology with economics (DellaVigna, 2009; Gul & Pesendorfer, 2008; Mullainathan & Thaler, 2001). The theory suggests that people’s decisions are influenced by heuristics and behavioural biases. Behavioural biases such as framing, anchoring, herding, regret, availability, and representativeness have been heavily investigated in the past decade especially after the global financial crisis (Cueva & Rustichini, 2015; Gerardi, Goette, & Meier, 2013; Kiss, Rodriguez-Lara, & Rosa-García, 2016). Behavioural biases affect people’s process of decision-making and in turn their final decision. Studies have also shown that cognitive ability is a very important factor that affects people’s lives and their achievement (Benjamin, Brown, & Shapiro, 2013; Choi, Lee, & Cioffi, 2011; Jensen, 1998; Oechssler, Roider, & Schmitz, 2009; Van Rooij, Lusardi, & Alessie, 2011a,2011b). People will react differently to the same problem and therefore their decisions will have a different impact on their lives (Korniotis & Kumar, 2013). This can sometimes be explained by their cognitive ability. Studies show that cognitive ability influences behavioural biases (Agarwal & Mazumder, 2013; Benjamin et al., 2013; Gerardi et al., 2013; Grinblatt, Keloharju, & Linnainmaa, 2012; Oechssler et al., 2009). One commonly used test to measure cognitive ability, developed by Frederick (2005), is the cognitive reflection test (CRT). According to Kahneman (2011), cognitive reflection measures to what extent and how individuals are able to override their initial decision and move to a more rational decision when faced with problems. Frederick (2005) showed that in financial decision-making, risk and time preferences were closely related to cognitive ability. In other words, individuals with a high CRT score are more patient and risk-loving compared to individuals with a low CRT score. Some studies have indeed found a relationship between the CRT score and behavioural biases (Albaity, Rahman, & Shahidul, 2014; Cokely & Kelley, 2009; Toplak, West, & Stanovich, 2011). For example, Corgnet, Hernán-González, Kujal, and Porter (2014) found a positive relationship between CRT scores and the earnings in asset markets. Frederick (2005) and Toplak et al. (2011) showed a positive relationship between CRT scores and general measures of intelligence. Other studies documented that risk and time preferences were linked with CRT scores (Brañas-Garza, Garcia-Muñoz, & González, 2012; Frederick, 2005; Oechssler et al., 2009). With regards to gender differences, various studies have investigated and confirmed the difference of CRT scores between males and females in different samples (Campitelli & Gerrans, 2014; Pennycook, Fugelsang, & Koehler, 2015; Primi, Morsanyi, Chiesi, Donati, & Hamilton, 2016). Although a difference in CRT scores exists between the genders little research has carried out an in-depth analysis of the source of the gap in CRT scores due to gender (Primi, Donati, Chiesi, & Morsanyi, 2018). In addition, previous research has reported that a difference in CRT scores exists based on gender, race and nationality. For instance, Albaity et al. (2014) and Albaity, Rahman, and Isa (2015) investigated gender, race, age, and nationality in Malaysia and found that males scored significantly higher than females and that Chinese students scored higher than Malay students in the CRT. Frederick’s(2005) results showed that the difference between males and females in the CRT persisted, even along with the control of SAT math scores. Several independent studies have since integrated the gender effect (Albaity et al., 2015; Campitelli & Gerrans, 2014; Pennycook et al., 2015; Primi et al., 2016; Zhang, Highhouse, & Rada, 2016). Some have postulated that the gender gap in CRT performance may be ascribed to intensified anxiety toward solving mathematical problems (Primi et al., 2018). Albeit the gender effects in the CRT seem to be robust across various studies, not many studies have considered the basis of the gender gap in performance. Though the research on gender dissimilarities in decision-making styles has been inconsistent, Mallek & Albaity, Cogent Economics & Finance (2019), 7: 1567965 https://doi.org/10.1080/23322039.2019.1567965 Page 2 of 14 several studies have explained that women score higher than men on intuitive type 1 processes, and lower on rational type 2 processes (Najmaei, 2014). This study seeks to understand the basis of gender differences across nationality in the CRT and on behavioural biases. Eckel and Grossman (2008) and Croson and Gneezy (2009) concluded in their research that women were more riskaverse when compared to men. On the other hand, Sarin and Wieland (2016), Eckel and Grossman (2008) and Tanaka, Camerer, and Nguyen (2010) found that there was no significant difference in gender with regards to risk aversion. Some might argue that the level of financial literacy or risk literacy plays a role in people’s ability to make rational decisions. Studies have found that a higher level of financial literacy is correlated with savings behaviour for retirement and holding stocks (Fornero & Monticone, 2011; Lusardi & Mitchell, 2011; Van Rooij et al., 2011a). Others have found that low levels of numeracy and financial literacy are related to low stock market participation, higher rates of loan defaults and high levels of misunderstanding about investment risk (Bateman et al., 2015; Christelis, Jappelli, & Padula, 2010; Gerardi et al., 2013; Van Rooij et al., 2011a). Another important variable that is crucial to decision making is the level of confidence. Confidence is a very important trait in any context. For example, confidence is a driver of career success (Koellinger, Minniti, & Schade, 2007) as well as for individual well-being (Pirinsky, 2013; Taylor & Brown, 1988). On the other hand, overconfidence is a type of bias that might lead to a deviation of one’s own abilities as well as relative to peers. Being overconfident might lead to positive as well as negative outcomes. For example, overconfidence can lead to positive impacts on ambition, morale and persistence (Johnson & Fowler, 2011; Ring, Neyse, David-Barett, & Schmidt, 2016). On the other hand, Barber and Odean (2001) found that overconfident investors diluted their payoff by trading excessively. Others found that those who were overconfident about their chances of success tended to engage excessively in competitive situations or they had less reaction to negative news leading to lower earnings (Camerer & Lovallo, 1999; Trinugroho & Sembel, 2011). With regards to the overconfidence link to the CRT, Hoppe and Kusterer (2011) found that subjects with higher CRT scores had significantly more precise self-assessment. In other words, individuals with high cognitive ability are not as overconfident as those with low cognitive ability. When testing gender differences Ring et al. (2016) found that males thought that they were better performers than their peers, however the same was not true among females. Therefore, the purpose of this study is twofold. Since there are no previous studies linking gender and nationality to CRT performance, risk literacy and time and risk preferences, the first is to test the effect of gender and nationality differences on CRT performance and behavioural biases in a representative United Arab Emirates (UAE) college student sample. The second purpose is to investigate whether the CRT scores and behavioural biases are different across the interaction between gender and nationality. The contribution of this study is as follows. First, the study contributes to the current literature by examining the differences in gender and nationality in the UAE. This is important for two reasons. First, the results of this paper can be generalised to the economies in the Gulf Cooperation Council (thereafter GCC) countries since they have the same culture and worldview. These countries are classified as collectivist according to the Hofstede individualism indicator. Second, these economies used to be oil-dependent; however, these countries are moving away from oil dependency especially due to the fall in oil prices. Hence, the financial services sector is one of the sectors that is growing in the GCC region and it is important for investors as well as for fund managers to learn more about the behaviour of their prospects. Third, the UAE Economy has grown vastly over the past 20 years and part of this growth has been accomplished by growth in the investment in financial assets. For example, the total value of the stock market as a percentage of GDP increased from 0.78% in 2003 to 26% in 2009. However, due to the global financial crisis, this rapid growth reversed to reach 5.5% in 2013. It is clear that the economy did not react immediately to the crisis and when the reaction occurred the economy did not bounce back that fast. Therefore, it is imperative for policymakers to acknowledge the fact that there are things that influence the behaviour of investors, whether they are individual or institutional. Mallek & Albaity, Cogent Economics & Finance (2019), 7: 1567965 https://doi.org/10.1080/23322039.2019.1567965 Page 3 of 14 2. Data and method 2.1. Participants In this study, a questionnaire that included statements addressing the three CRT questions—as well as questions targeting behavioural biases such as risk preference, time preference, risk literacy, and overconfidence was distributed. The total number of questionnaires distributed was 1000 however only 880 questionnaires were returned. Out of these only 770 questionnaires were usable in the analysis (see Appendix Afor the questions). The 110 questionnaires that were eliminated were due to them either being incomplete or because some of the respondents had seen some of the questions before. The participants were undergraduate business students from public and private universities across the UAE. All students participated voluntarily and gave their consent before engaging in the study. To ensure further anonymity, no personal information other than gender and nationality was included in filling out the questionnaire. The questionnaire was conducted in English since this is the medium of instruction at all UAE universities. The analysis consisted of running the t-test, ANOVA and cross tabulation tests to examine the proposed hypothesis of this study. Prior studies on consumer behaviour research have found a significant accuracy of student responses when compared to other consumers. Many articles in behavioural finance, economics, and psychology have used students as their sample (Elliott, Hodge, Kennedy, & Pronk, 2007; Finney & Finney, 2010; Inman, 2007; Noussair, Trautmann, & Van de Kuilen, 2013; Statman, 2008; Weber, Blais, & Betz, 2002; Wood & Zaichkowsky, 2004). Ariely and Jones (2012) indicated that student participation was acceptable and reliable based on different reasons. First, young adults act very similarly compared to the working older regarding their core actions. Second, the endowment effect influences both students and experienced people. Finally, students and experienced individuals operate their brains and decision-making techniques under the same restrictions. 2.2. Measures 2.2.1. Cognitive reflection test Cognitive ability is assessed by means of a three-item cognitive reflection test (CRT), as introduced by Frederick (2005). To this end, each of the three items of the CRT has an apparently impulsive (but wrong) answer that quickly springs to mind, and a reflective (but correct) answer that is easy to understand when explained. However, arriving at the right answer may necessitate overriding the initial, impulsive answer. Therefore, the test is intended to measure an individual’s ability to be involved in cognitive reflection, and consequently to resist conveying the spontaneous but incorrect answer. A wide array of research in the areas of judgment and decision-making, as well as memory and reasoning, has demonstrated that the consequences of cognitive processes often thoroughly deviate from what is normatively considered to be rational behaviour. It was in the late 1990s that scholars became aware of the substantial variability across participants on each of the cognitive bias tasks (Cavojová, 2016;Schulze&Newell,2016;Toplaketal.,2011). Thus, individual differences play a key role in the deviation between the outcomes of cognitive processes and those of normative models that have resulted in the growing body of correlational research of cognitive biases. 2.2.1. Risk preference While as expected the utility theory considers individuals to be risk-averse, the prospect theory claims that people will be more eager to take risks to avert losses than to attain gains; that individuals will shift from risk aversion to risk-seeking behaviour when the parity of a gamble alters from positive to negative (Kahneman & Tversky, 1979). With respect to cognitive abilities, subjects with high CRT scores are significantly more eager to gamble compared to their counterparts (Frederick, 2005; Noori, 2016). The experiments of Dohmen, Falk, Huffman, and Sunde (2010) showed that high scoring CRT subjects were significantly more willing to take risks and that the correlation between risk aversion and cognitive ability existed for both males and females, and for young and old, though the relationship was weaker for females and younger individuals. Oechssler Mallek & Albaity, Cogent Economics & Finance (2019), 7: 1567965 https://doi.org/10.1080/23322039.2019.1567965 Page 4 of 14 et al. (2009) found that the high scoring CRT group tended to choose options that were risk neutral. The risk preference question, in this study, measures whether the individual is comfortable with safe investments or with risky investments. 2.2.3. Time preference Several studies have supported the idea that smart people exhibit more patience (Oechssler et al., 2009). Benjamin et al. (2013) found that a group with a higher SAT math score was willing to choose a larger and later reward over a smaller earlier one. A comparable result was found when using the CRT test instead of the SAT math test (Albaity et al., 2014; Frederick, 2005). Dohmen et al. (2010) found that the high-scoring CRT group was significantly more patient and that the correlation between cognitive ability and patience existed for males rather than for females, and for young and old, though the relationship was weaker for younger individuals. The time preference question, in this study, measures whether individuals are willing to wait for a higher return compared with taking a lower return now. 2.2.4. Overconfidence The overconfidence bias is a common behaviour in which people overvalue their capabilities to accomplish a particular task effectively (Brenner, Koehler, Liberman, & Tversky, 1996). With respect to cognitive abilities: while the more analytical decision makers are relatively more successful in evaluating the right number of correct answers, intuitive decision makers seem to be overconfident (Hoppe & Kusterer, 2011; Noori, 2016). In this study, the participants were asked five general knowledge questions. After responding to each of the items, the participants were requested to assess their degree of confidence to guess the number of correct answers. For the overconfidence question, we classified the answers into three groups by comparing the number of effective correct answers to the expected correct answers out of the five statements. If the respondent said five but effectively had only two correct answers, then they were considered overconfident, and vice versa. 2.2.5. Risk literacy The Berlin Numeracy Test is a new psychometrically robust instrument designed to quickly assess statistical numeracy and risk literacy (Cokely, Galesic, Schulz, Ghazal, & Garcia-Retamero, 2012). It has been shown that numeracy is important in a range of everyday decisions and that low numeracy raises susceptibility to a variety of cognitive biases and fallacies, even when common intelligence is controlled (Peters et al., 2006). Four questions were adopted from Cokely et al. (2012) and Ghazal, Cokely, and Garcia-Retamero (2014) to measure the level of risk literacy. To create the risk literacy variables, the questions with correct answers were re-coded for the correct answers, therefore, creating 4 individual variables. 3. Results and discussion The results of the study are reported in this section. We start with Table 1which describes the demographic profile of the respondents. The majority of the respondents were females, non-local, between 18–22 years old, of whom at least one parent had a college degree. This profile was expected since most of the students were undergraduates. Table 2below reports the mean scores of the wrong and right answers for the three CRT questions. The responses to the three CRT questions are divided into two groups, either the Table 1. Demographic profile of the respondents Characteristic Percentage Female 60% 18-22 76% Non-UAE 65% At least one parent have college degree 69% Mallek & Albaity, Cogent Economics & Finance (2019), 7: 1567965 https://doi.org/10.1080/23322039.2019.1567965 Page 5 of 14 respondents scored high ability with 2 or 3 correct answers or low ability with 1 or 0 correct answers. One sample t-test was performed to examine whether the mean score of the correct answer for each question was significantly different from zero. The results suggested that all of the mean scores were significantly different from zero. Based on these results, it was clear that the number of correct answers decreased, the more computationally challenging that the questions became. For example, the mean score of the correct answer for question 1 was 0.34, while the mean score of the third question was 0.20. In our sample, 52% of the subjects answered none of the questions correctly, 25% knew the correct answer to one question, 15% to two questions, and 8% answered all three questions correctly. On average, the subjects answered less than one (0.80) of the CRT questions correctly. Table 3shows the mean difference using the t-test for the CRT questions across gender, nationality, and the interaction between gender and nationality. The results for gender differences indicated that male subjects scored higher than female subjects in the three CRT questions. This result is similar to Albaity et al. (2015) and Noori (2016), who found that males, on average, scored higher than females. Some have hypothesised that the gender gap in CRT performance may be ascribed to enhanced anxiety toward solving mathematical problems (Primi et al., 2016). In terms of nationality, non-UAE nationals scored higher than UAE citizens in all of the questions. Most of the studies conducted on CRT have focused on demographics such as gender, race, and Table 2. The percentage of correct answers, mean scores and standard deviation for CRT questions Questions % Correct answer % Wrong answer Mean p-value Std. Dev. Bat & ball 34.3 65.7 0.34 0.84 0.000* Widgets 25.1 74.9 0.25 0.34 0.000* Lily pads 20.2 79.8 0.20 0.40 0.000* Note: *,**and *** indicate significance at 1%, 5% and 10% respectively. Table 3. Gender, nationality and their interaction mean difference test for CRT questions Characteristic Bat & ball (Q1) Widgets (Q2) Lily pads (Q3) Male 0.51 0.35 0.33 Female 0.25 0.20 0.13 t-value 7.25* 4.48* 6.48* UAE 0.22 0.20 0.09 Non-UAE 0.40 0.27 0.25 t-value −5.03* −2.20** −5.77* Female*Non-UAE 0.29 0.21 0.16 Female*UAE 0.38 0.28 0.23 t-value 3.00* 2.28** 2.65* Male*Non-UAE 0.60 0.38 0.42 Male*UAE 0.26 0.21 0.13 t-value −8.71* −4.29* −7.52* Note: *,**and *** indicate significance at 1%, 5% and 10% respectively. Mallek & Albaity, Cogent Economics & Finance (2019), 7: 1567965 https://doi.org/10.1080/23322039.2019.1567965 Page 6 of 14 age, while few have investigated the effect of nationality on CRT questions (Albaity et al., 2014). Nationality is important, since 89% of the population of the UAE is made up of foreigners, and there are a variety of schools to accommodate specific nationalities. While there are (to our knowledge) no studies linking nationality to CRT questions, we believe that studies carried out on the link between race and CRT scores are relevant. For example, Albaity et al. (2014) found that Malaysian Malay nationals and Malaysian Indian nationals scored lower than Malaysian Chinese nationals in CRT questions. This could potentially be attributed to the different educational backgrounds of Malay, Indian, and Chinese citizens (who attend different schools), in spite of their shared Malaysian citizenship. In the same vein, our results indicated that the difference could have arisen from the fact that the majority of locals attend public schools, while the majority of other nationalities attend international schools. These schools have different syllabi and the outcome would therefore differ. Table 4below shows the comparative results between genders for the risk and time preferences, risk literacy, overconfidence and the CRT questions. The results suggest the existence of a difference between males and females in terms of risk preference, risk literacy and overconfidence. In all cases, the difference is towards males. For example, males appear to be more risk-loving than females (Halko, Kaustia, & Alanko, 2012). In addition, among the four questions of risk literacy, three were found to be significantly different between males and females. The mean score suggested that males were on average more risk literate than females. The same trend appeared when we aggregated the risk literacy questions (Primi et al., 2018). Similarly, females scored a lower mean in overconfidence than males. This is in line with previous findings that indicated that males usually exhibit higher overconfidence than females (Barber & Odean, 2001; Bengtsson, Persson, & Willenhag, 2005). Lastly, the CRT questions showed that males scored higher than females in all of the questions. These results corroborate the results found by Campitelli and Gerrans (2014), Pennycook et al. (2015), Primi et al. (2016). Some have postulated that the gender gap in CRT performance may be ascribed to intensified anxiety toward solving mathematical problems (Primi et al., 2016). Carrying out the same analysis on nationality revealed a similar result to the gender result with the exception that the overconfidence effect disappeared when nationality was considered (Table 5). Local respondents were inclined to be risk takers, were less risk literate, were less overconfident, and had lower cognitive abilities than non-UAE respondents. In order to avoid sampling bias, we carried out an in-depth analysis of gender and nationality. We multiplied the gender by the nationality to obtain more detailed results. Table 4. Behavioural biases and cognitive reflection test by gender Behavioural biases and CRT Gender Male (Mean) Female (Mean) Mean difference p-value Risk Preference 0.35 0.24 0.00* Time Preference 0.64 0.59 0.16 Risk Literacy Q 1 0.14 0.08 0.00* Risk Literacy Q 2 0.19 0.10 0.00* Risk Literacy Q 3 0.13 0.10 0.15 Risk Literacy Q 4 0.12 0.06 0.00* Risk Literacy total 0.59 0.33 0.00* Overconfidence 0.48 0.37 0.00*** CRT 0.37 0.16 0.00* Note: *,**and *** indicate significant at 1%, 5% and 10% respectively. Mallek & Albaity, Cogent Economics & Finance (2019), 7: 1567965 https://doi.org/10.1080/23322039.2019.1567965 Page 7 of 14 © 2019 The Author(s). Thisopen access article is distributed under a Creative Commons Attribution (CC-BY) 4.0 license. You are free to: Share —copy and redistribute the material in any medium or format. Adapt —remix, transform, and build upon the material for any purpose, even commercially. The licensor cannot revoke these freedoms as long as you follow the license terms. Under the following terms: Attribution —You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use. No additional restrictions You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits. Cogent Economics & Finance (ISSN: 2332-2039) is published by Cogent OA, part of Taylor & Francis Group. Publishing with Cogent OA ensures: •Immediate, universal access to your article on publication •High visibility and discoverability via the Cogent OA website as well as Taylor & Francis Online •Download and citation statistics for your article •Rapid online publication •Input from, and dialog with, expert editors and editorial boards •Retention of full copyright of your article •Guaranteed legacy preservation of your article •Discounts and waivers for authors in developing regions Submit your manuscript to a Cogent OA journal at www.CogentOA.com Mallek & Albaity, Cogent Economics & Finance (2019), 7: 1567965 https://doi.org/10.1080/23322039.2019.1567965 Page 14 of 14