Understanding decision-making styles aims to help researchers and managers understand the phenomenon of why individuals participate in activities that may result in high-risk and possibly damaging outcomes [1]. One may only look as far as the decisions made by professionals in the financial sectors are concerned that resulted in the 2008 global financial crisis, or the British exit from the European Union and seemingly self-destructive decision that leaders make in relation to climate change, to understand why understanding decision-making styles has become a significant aspect of leadership. Within professional sectors, including the education sector, making decisions that will have a significant impact on the lives of people is so commonplace that Lunenburg [2] suggests that it is “a way of life”, that saturates all levels of organisation (from leadership, management, administrators to the grassroots level). Even though research indicates that there is a correlation between the democratisation of decision-making autonomy and positive relationships within the organisation [3], there remains a lack of understanding how decision-making styles influence the dimensions of decision-making and how this has an impact on the development of policy and practices, especially in professions where a sense of purpose and emotion plays a significant role, such as the education sector.

We assumed that there are 10 dimensions of the decision-making process. These dimensions can be listed as cautiousness, openness to experience, decision difficulty, agency, emotion neutrality, goal orientation, intuitive awareness, plan orientation, pro-activity, and rationality. We conducted an online survey designed by Google forms. The survey was posted to online forums. As the survey was conducted online, there was no pressure or motivation to state biased answers. The only motivation for the survey respondents was to help this academic research by their contribution. The survey was completely anonymous with no personal identification possibility. We collected information on the participants' age group, educational attainment, gender, current location, and current occupation. In order to measure the decision-making attributes, we followed a quantitative approach. Specifically, we asked survey participants to rank how they make decisions for each dimension. The answer scale was from 1 (not like me) to 10 (very much like me). Here is how each dimension is measured:

• Cautiousness: I am very cautious when it comes to new experiences.
• Openness to Experience: I feel comfortable following my intuitions when I'm making decisions.
• Agency: I prefer to have complete freedom over making my own decisions.
• Decision Difficulty: Making decisions is difficult for me, I prefer to follow somebody else’s advice.
• Emotion Neutrality: My emotions are neutral and objective when I make a decision.
• Goal Orientation: I am goal orientated and I’m willing to change the plan to achieve the goal.
• Intuitive Awareness: I depend on my intuitive awareness to make decisions.
• Plan Orientation: I am a logical and plan-orientated decision-maker and I don’t like changing the plan.
• Pro-Activity: I prefer being pro-active when making decisions.
• Rationality: I prefer to think about a decision I have to make in a rational way.

Google survey analytics was not enough to perform detailed statistical reporting on the data. Therefore, we used different data analysis and visualization tools to explain our data. Initial data organization was performed with Excel 2016. Minitab software is used to create relevant statistical tables. Finally, Tableau 2018.3 is utilized to create cross tabulation tables and visual analytics.

In terms of demographic information, the age distribution shows normal behaviour. Out of 356 respondents, 103 of them stated that they are around 36-45 years old. Next comes 46-55 interval, followed with 26-35 interval each of which has 86 and 79 respondents, respectively. In terms of gender, our data is similar to that of Karwowski [56]. 204 of all responses were of females; of which 89 were of those working in an education sector. 152 of all responses were of males, of which, 50 were of those working in an education sector. 134 responses were from Africa, followed by Europe (97), America (39), Middle-East [49] and Australasia [37]. Table 1 below shows cross-tabulation of respondents based on different demographic factors (Table 1).

It is worth noting about (Table 1) that while we have data on individual occupations, the sectoral data was highly diversified including administrative positions, agriculture, business and financial operations, computer design, education and training, engineering, entertainment, fashion, government, healthcare, hospitality, legal professions, marketing, media, military, retired, service, student, and even unemployed. However, the number of educational sector workers constitutes almost 40% of the data. Therefore, we decided to combine all other occupational sectors into the “Others” category. That also helped us to test if educational sector workers are more likely to have different personality categories. Similarly, in the original data, the educational attainment had categories such as doctorate degree, post-graduate degree, university degree, secondary school, and technical degrees. In order to see if having a university degree makes a difference, educational attainment has also been categorized. Those with a doctorate degree, post-graduate degree, or at least a university degree have been categorized as University+ whereas those with a secondary school or technical degrees categorized as others.

In the model, we assumed 10 dimensions in the decision-making process. Based on these dimensional measurements, the individuals can be categorized into 4 styles of decision-making. These decision-making styles are avoidant, designer/auditor, flexible, and fluent. A quick statistical analysis of the decision-making dimensions is shown in Table 2.

It is of interest to see whether each dimension is a unique one or redundant. Thus, if some measurements are highly correlated with each other, then these measurements would be redundant. Table 3 shows us the correlation between measured dimensions.

The correlation matrix in Table 3 suggests that each dimension is a unique one. There is almost no correlation between dimensions. The only correlation that is higher than 0.5 is the correlation between the openness (Q8) and Intuitive Awareness (Q2) which is equal to 0.52. However, these dimensions are obviously complementary with each other and can be used to classify the decision-making styles without redundancy. The decision-making dimensions are used to categorize each respondent into a specific personality type. As most answers were clustered at values above the arithmetic means, we first transformed the variables by normalization. The normalized values work as follows:

(1)

Here each Z_i value refers to the normalized score of decision-making dimension. X_i is the original score (from 1 to 10), mean (X) is the arithmetic mean, and Std. Dev. (X) is the standard deviation. This transformation enables us to see where each decision-maker stands according to other survey respondents.

Using 10 dimensions of decision-making, each individual is classified into a specific personality type. The decision-making styles are avoidant, designer/auditor, flexible, and fluent decision-making styles. We tested the decision-making classifications using both linear and quadratic discriminant analysis. According to the linear analysis, out of 356 respondents, the model estimated 303 of them correctly. The quadratic discriminant analysis performed much better. The quadratic model estimated 333 classifications correctly. This translates into a success rate of 93.5%. A summary of classification results is shown in Table 4.

According to the model, 150 of the respondents had fluid type of character, 146 had flexible type character, 31 could be called avoidant, and 29 could be called as designer/auditor style.

The quadratic classification method has correctly identified both avoidant and design/auditor decision-making styles correctly (100% success rate). Out of 146 flexible decision-making styles, 132 were classified correctly; 7 were misclassified as fluid, 6 were misclassified as avoidant, 1 was misclassified as designer/auditor type (90.4% success rate). Out of 150 fluid decision-making styles, 141 were classified correctly; 3 were misclassified as avoidant, 3 were misclassified as design/auditor, and 3 were misclassified as flexible type (94% success rate). These results suggest that personality classification based on decision-making dimensions was very well according to quadratic discrimination analysis.

Based on the above decision-making styles, it is also of interest to see whether associated individual characteristics such as age, educational attainment, gender, geographical location, and occupational sector affect the decision-making styles. Therefore, we tested each of these factors using Chi-square tests for association. For duality purposes, anyone below 35 is classified as a young professional, whereas above 35 is another category. Educational attainment measures whether the individual has a university degree or not. Gender is simply classified as male or female. Geographic location measures whether the individual is located in the western world or not. Occupational sector measures whether the individual is working in an educational institution or in another sector.

The default (null) hypothesis for the effect of each individual characteristic is as follows:

H₀: The decision-making style is independent of individual characteristic.

The alternative hypothesis is defined as follows:

H_a: The decision-making style is affected from individual characteristics.

These tests are conducted using Minitab analysis for an association. The computer output also included details on the true and expected number of respondents under each category. However, in Table 5, we report only the simplified results based on analysis:

As can be seen in Table 5, the chi-square tests for association do not suggest any significant relation between decision-making styles and personal socio-demographic characteristics. These findings imply that factors such as age, educational attainment, gender, geography, and occupation do not affect decision-making styles. However, if we consider the significance value of 90% instead of 95%, then gender can be considered as an effective factor in personality. Therefore, it is worth investigating the effect of gender on personality type. Table 6 shows the results for detailed chi-square test for association between gender and personality styles.

The fractional numbers in Table 6 correspond to the expected number of respondents based on probability distribution whereas the integers correspond to the actual number of respondents. For independence, the actual numbers should be as close as possible to the expected numbers. For Designer/Auditor and Fluid decision-making styles, these numbers are very close. However, 24 females and 7 males are classified as avoidant type whereas these numbers are expected to be 17.76 and 13.24, respectively. Thus, female respondents are more likely to have avoidant personality compared to their male counterparts. Similarly, 76 females and 70 males are classified as flexible type, whereas these numbers are expected to be 83.66 and 62.34, respectively. Thus, male respondents are more likely to have flexible personality compared to their female counterparts. These results suggested that gender might be a significant factor in personality classification. Therefore, we decided to test each individual decision-making factor to see if gender differences are significant. The individual test results for each dimension of decision-making are shown in Table 7.

Based on Table 7, one can conclude that there are minor gender differences in decision making process. Female decision makers are slightly more cautious, pro-active, rational, intuitively aware, slightly less comfortable, goal oriented, and plan oriented compared to male decision makers. They also experience higher decision difficulty accompanied with more decision freedom when compared with male counterparts. However, none of these factors are statistically significant. The only significant divergent factor is emotion neutrality. Female decision-makers are less emotion neutral than male decision-makers. Our results in Fig. (1) suggest that this finding is amplified for educational sector workers (teachers, university lecturers, lab instructors, teaching assistants).

The visual in Fig. (1) shows the confidence interval for the mean values. Females are less emotion-neutral compared to male decision-makers. While this factor is insignificant for those who work in other sectors, the gender difference is magnified within the education and training sector. The average emotion-neutrality score for females working in the education sector is 5.2, whereas males have an average emotion-neutrality score of 6. This implies a significant difference of 0.8 points between male and female workers in the education sector. Thus, females working in the education sector are more likely to make emotional decisions.

This result also partially explains the association between gender and decision-making style. As females are more likely to make emotional decisions, they are also more likely to be considered as avoidant, and less likely to be considered as fluent decision-making style. Similarly, the avoidant behaviour is less observed in male decision-makers and they are more likely to fall into fluent decision-making style.

Fig. (1). Mean Emotion-Neutrality Scores by Gender and Occupation.

The results suggest that female participants have a significantly stronger leaning towards making emotion-dependent decisions; a finding that is specifically strong in the education sector. Hutson-Comeaux and Kelly [57] assert that women show more emotional expression, which supports these findings. However, there is no indication in the data that the prevalence of emotion should be mistaken as an antithesis of rationality, which is supported by the findings of Gaudine and Thorne [58] as well as Pulsford, [59]. A stronger emotion-dependent approach to making decisions does not lead to irrationality, conversely, it should rather be considered as an inimitable characteristic of the decision-making process.

It is important to reiterate that emotions are an innate part of the decision-making process [60]. Research suggests that emotional behaviour may as a matter fact contribute to less deleterious decision-making [61]. As the inclusion of emotional intelligence training has become a primary aspect of the corporate ethos [62], the acceptance of emotion as a dimension of decision-making has grown. Specifically, in the educational sector, school leaders who show emotion-neutral decision-making may face opposition from millennial parents who are more versed in the importance of a better student-teacher relationship. This assertion may be of significance when one considers that female teachers are more likely to encourage innovative educational ideas in comparison to their male counterparts. Emotion and gender is another important, if not controversial field of inquiry [63].

Research on the stereotyping of women in the workplace confirms that there remain stereotypical claims that women are more emotional than their male counterparts [64]. Negative stereotyping is a contributing factor to why female workers remain under-represented in management, politics and senior leadership [65-67]. Colloquially known as the glass ceiling, this stereotyping may impede the career progression of female workers [68-71]. In addition, research indicates that scientific fields such as in medical science, being a female doctor can affect a patient’s perceived performance by the medical practitioner [72, 73]. However, there is also research that suggests that where gender diversity of company boards is present, such diversity is not indicative of superior financial performance in these companies [74]. The relationship between gender and race is also of interest [75]. In some studies, the term ‘race’ is replaced by ethnicity [76].

The role of gender differences in emotion dependence may also be accredited to the parochial cultures or even to the ethnicity of the participants in gender-based research [77-79]. Some studies claim that the prejudice against women is prevalent when decision-making takes place in mixed gender groups [80]. The relationship between the gender-role composition of the group and explicit gender role division in group-based decision-making has a significant impact on how emotional dependence in women is perceived [81]. The prevalence of male bias in such research findings is concerning and leads one to question the validity of equal representation in the participant samples of such research.

However, as cracks in the glass ceiling begin to appear, typecasts perceptions of women in the workplace are changing and the gender-gap is shrinking as more women find their way into private and public sector leadership positions as well as taking precedence as active contributors in the economy [82, 83]. Active participation of women workers in the labour market might lead to an organic reduction in gender-differentiated roles within the economy [84]. As an example, entrepreneurs are more likely to encourage family-oriented company policies [85]. Similarly, businesses that encourage women in leadership positions have an increased likelihood to nurture female-friendly policies [86].

In this article, we revealed results of an international survey conducted on a global scale. The survey asked respondents to quantify how they make daily decisions. We specifically gathered information on their levels of cautiousness, openness to experience, decision difficulty, agency, emotion neutrality, goal orientation, intuitive awareness, plan orientation, pro-activity, and rationality when making decisions. Despite the survey was conducted anonymously, we gathered socio-demographic information such as age group, educational attainment, gender, current location, and current occupation that provided a deeper understanding of the phenomenon of decision-making. The participants were categorized into four different decision-making styles, based on how they responded to the decision-making dimension questions. These decision-making styles have been listed as avoidant, designer/auditor, flexible, and fluid decision-makers. Of the total participants, 150 were classified as fluid, 146 were classified as flexible, 31 were found to be avoidant, and 29 had a designer/auditor decision-making style. We tested whether personal socio-demographic characteristics such as age, educational attainment, gender, geography, and occupational sector are effective in the construction of decision-making styles. We found that except the gender dimension, these factors are not positively correlated with any decision-making style. Females showed a greater propensity towards having an avoidant decision-making style and less likely to be considered as having a fluent decision-making style. This phenomenon warranted further investigation and we also tested whether there are also substantial differences in decision-making dimensions for women in relation to men. We found that, in terms of emotion neutrality dimension, significant differences exist between female and male respondents; a fact more visible in the education sector.