From STEM to STEAM: How to Monitor Creativity

Abstract

Creativity is a broad and complex construct, difficult to define and to quantify, assumed to introduce new impulses into science education (STEM), and leading to better acceptance of science by adolescents. Therefore, increasing efforts are being undertaken to integrate traditional creativity (Arts), in modifying STEM to STEAM. Consequently, a valid way of empirically quantifying of creativity of adolescents is needed. In this study, part of a European initiative (CREATIONS), an 8-item Likert-scale questionnaire quantifying individual creativity was administered to a sample of 2,713 students, aged 11–19 (M ± SD = 15.71 ± 2.24; 54.7% females), revealing two subscales: one, labelled Act, covering conscious and trainable cognitive processes; the second, named Flow, contained items describing elements of flow experiences, a mental state of creativity. Analyses indicated that there were no gender differences and that younger students’ creativity scores were higher than those of older students. Recommendations for implementation in STEAM lessons are discussed.

Creativity is a broad and complex construct, difficult to define and to quantify. Unlike other complex variables such as personality or attitudes, research into creativity quantifying is still disputed. Although personality has been measured since the 1940s (Parish, Eysenck, & Eysenck, 1965), and attitudes in general since the 1930s (Likert, 1932; Thurstone, 1928), and enviromental attitudes in specific (Blaikie, 1991; Bogner & Wiseman, 1999; Dunlap & Van Liere, 1978). Although Spearman (1904) and Hargreaves (1927) started the first move, it took much longer to appropriately capture its construct psychometrically. Anderson (1964) defined, “Creativity is crossing out mistakes” (in Torrance, 1995, p. 147). The definition of creativity as a process of sensitizing for problems, closing gaps in knowledge, searching for solutions and communicating results (Torrance, 1966) is based on the recognition that “sensitivity to problems seems to be essential in getting the creative thinking process in motion” (Guilford, in Torrance, 1995, p. 69).

Discussions of creativity remain ambiguous or even contradictory (e.g., Amabile, 2012; Runco, 2004; Sternberg & Lubart, 1996). A comprehensive definition could focus on the ability to generate novel and useful ideas (Mumford, 2003). Csikszentmihalyi (1996) subsumed creativity within a triangular system of individual, domain, and appreciative environments. Guilford (1950) derived his understanding of creativity from a structural model of intelligence, defining creativity as any new, unprecedented, and effective method of solving problems. Similarly, creativity seems to be a combination of talent, knowledge, ability, intrinsic motivation, and personality traits, additionally supported by environmental conditions (Holm-Hadulla, 2010). From a psychological point of view, fluid thinking and association, as well as the ability to change perspectives, are important. Readiness for an open mind and preparedness for meaningful processing of ideas are also to be considered. Several methods may foster creative thinking, for example concept mapping, mind mapping, or brainstorming (Gordon, 1961; Novak & Gowin, 1984; Osborn in Merrotsy, 2017). Of particular interest is the effect of problem-solving tasks on creativity (Ma, 2009), maybe because the need to solve problems might have been the evolutionary trigger for the development of (human) creativity (Csikszentmihalyi, 1996).

Neuronal research has regarded creativity either as the result of right brain hemisphere processes or of interhemispheric interactions (Kowatari et al., 2009). Meanwhile, more complex neuronal locations have been discussed (e.g., Carlsson, Wendt, & Risberg, 2000; Chávez-Eakle, Graff-Guerrero, García-Reyna, Vaugier, & Cruz-Fuentes, 2007; Shah et al., 2013). As human creativity may have its foundation in the combination of previously unconnected mental representations, creative thinking is regarded as a matter of combining neural patterns (Thagard & Stewart, 2011). Creativity seems rooted in normative cognitive processes and neuronal representations apparently measurable by neuronal imaging techniques. Several studies have demonstrated a new and promising way to investigate aspects of creative thinking using neuroscientific methods (Thagard & Stewart, 2011; Tyan, Liao, Shen, Lin, & Weng, 2017). Particular brain areas have been identified with creativity. Aberg, Doell, and Schwartz (2016), for instance, have described the right brain hemisphere as specifically contributing to creativity and dopamine-related activities as a determinant of human creativity.

Recent evidence has demonstrated neurophysiologically determined gender differences regarding rewarding (Volf & Tarasova, 2013). Tyan et al. (2017) reported gender differences in communication between brain areas: Brains of male teenagers seem to involve better intrahemispheric communication, brains of female teenagers for interhemispheric communication. The male network organization is more local, more segregated, and more similar to small-world networks. These findings are in line with the differences in divergent and convergent thinking and might indicate a neuronal difference in this gender-specific ability (Shen, Liu, Shi, & Yuan, 2015).

Gender differences in creativity studies still seem ambiguous. Besançon and Lubart (2008) reported diverse studies supporting the hypothesis that women are considered more creative (e.g., Ülger & Morsünbül, 2016); others claim the same for men (e.g., Shin, Jung, Choe, & Han, 2002) or even deny that there are differences (e.g., Besançon & Lubart, 2008). Knowledge about differences in neurological structures might provide a deeper insight into the nature of creativity, the emergence and importance of creativity, motivation, and personality development (Lau & Cheung, 2010). Human creativity is regarded as providing an important evolutionary advantage (Csikszentmihalyi, 1975). Cultural aspects, too, may provoke gender differences in the development of creativity (Sternberg, 2006; Torrance, 1966), which brings education into play (Vosylis, Kaniusonyte, & Raiziene, 2016; Wechsler, 2006).

Creative processes often involve a special state of consciousness called flow, characterized by complete absorption in an activity (Csikszentmihalyi, 2000). In this mental state, a person performing an activity becomes fully immersed in a feeling of energized focus, full involvement, and enjoyment. Flow is perceived as resulting in high intrinsic motivation scores, particularly at young ages. Unfortunately, the ability to flow tends to vanish, probably replaced by knowledge-based, logical sense-making patterns (Csikszentmihalyi, 2000). A one-sided education based on verbalizing and on testable knowledge may even accelerate this shift.

Behaviors that promote or inhibit creative potential stem from social backgrounds, working climates, and education experiences. Perfectionism may cause strict target orientation and problem-solving rituals to prevent mistakes, maybe leading to fear, disappointment, failure, and mistakes (Grant, Grant, & Gallate, 2012). A helpful error management culture, with no blame for failure, may help to avoid such problems. Other factors can be both supporters and blockers of creativity: Social distance inhibits creativity (Sosa, 2011). Strong team spirit might promote self-censorship or prevent presenting new ideas to avoid being different. Another cultural disadvantage might originate in the separation of work and play: Playful testing can encourage the development of something new (Csikszentmihalyi, 1975). Guilford (1950), a pioneer of research into creativity, focused on two aspects: (a) How can the creative promise of children be discovered? (b) How can the development of creative personalities be promoted? As this study monitors individual creativity, a self-reported questionnaire was applied to answer three questions: (a) Are the factors of the questionnaire consistent with earlier published ones? (b) Are there gender differences? (c) Are there age-dependent differences in creativity?

Method

Participants and survey

A sample of 2,713 students (aged 11–19 years, M ± SD: 15.71 ± 2.24; 54.7% girls) from six EU countries (United Kingdom, Greece, Sweden, Malta, Italy, Germany) completed the questionnaire, originally consisting of 10 items (Cognitive Processes Associated with Creativity, CPAC; Miller & Dumford, 2016). Responses were assessed using a 4-point Likert scale from ranging from never (1) to very often (4). All items are listed in Table 1. A master version was translated into the national languages and retranslated for control purposes. Compared to the original questionnaire the you in personal statements was replaced by an I did…, as the focus was on self-reported personal behavior.

TABLE 1 Exploratory factor analysis with CPAC after rotation (valid N = 2113)

		Components
Pattern Matrix^a		Pilot Factor Analysis 10		Final Factor Analysis 8
		1	2	1	2
CPAC 04	I joined together dissimilar concepts to create a novel idea.	,749		,726
CPAC 05	I incorporated a previously used solution in a new way.	,729		,717
CPAC 06	I made a connection between a current problem or task and a related situation.	,680		,682
CPAC 07	I imagined a potential solution to explore its usefulness.	,621		,656
CPAC 03	I looked at a problem or task from a different angle to find a solution.	,621		,640
CPAC 01	I tried to generate as many ideas as possible when approaching a task.	,496	,266	/	/
CPAC 02	I asked other people to help generate potential solutions to a problem.	,217		/	/
CPAC 09	I lost track of time when intensely working.		,823		,792
CPAC 10	I felt that work was automatic and effortless during an enjoyable task.		,723		,742
CPAC 08	I was fully immersed in my work on a problem or task.		,646		,704

Extraction method: Main component analysis.

Rotation method: oblimin with Kaiser normalization.

a. The rotation is converged into 5 iterations.

Factor 1 eigenvalue explains 37,07% variance;

Factor 2 eigenvalue explained 13,75% variance

Total eigenvalue explained 50,82%

Table shows factor loadings >±0.2, Cronbach’s alpha = 0.767.

Only complete questionnaires were included in the analysis. An exploratory factor analysis with subsequent oblique rotation yielded two factors. To assess the adequacy of the sample, the Kaiser–Meyer–Olkin (KMO) test (Kaiser, 1970) and Bartlett’s test of sphericity were applied. Two of the 10 original items were dropped due to cross loadings or low loadings. A subsequent factor analysis, the KMO test = 0.869 and Cronbach’s alpha = .77 were adequate. Similarly, the Kaiser–Guttman criterion suggested two factors (Kaiser, 1960).

Based on the central limit theorem, a normal distribution of the data was assumed and differences between factors using a paired sample t-test were used. Gender differences were examined using an independent sample t-test. As the objectives were not directional, a two-tailed level of significance was chosen the for Spearman correlation between subscales of CPAC, with a corresponding 95% confidence interval.

Results

After dropping of two items, exploratory principal axes factor analysis with oblique rotation yielded two factors on the basis of eigenvalues > 1.0, explaining 50,82% of the total variance. The scree-plot and the component plot in rotated space (Figure 1) supported a two-factor solution. All items loaded higher than 0.4, the KMO value of 0.869 indicated distinct and reliable factors (Kaiser, 1970). The values of the measure of adequacy of the anti-image matrix were above .80. The Bartlett test showed correlations between items as significantly different from zero (p < .001). The pattern matrices of all items and of the reduced set as well are shown in Table 1. The Cronbachs’ alpha when an item was deleted ranged between .736 and .755. The biggest improvement caused item02 with a Cronbach alpha = .781, supporting dropping it. Item01 was dropped due to cross-loading.

FIGURE 1 Component plot in rotated space with remaining CPAC 8 items indicating 2 factors (Act = Item03-07, Flow = Item08-10).

A subsequent repeated factor analysis confirmed the two-factor model with 8 items based on Kaiser–Guttman criterion (Kaiser, 1970). The Cronbach’s alpha of 0.77 indicated an acceptable overall reliability (Lienert & Raatz, 1998).

The exploratory factor analysis yielded two oblique factors (KMO measure of sampling adequacy = 0.869 and Bartlett’s test of sphericity Chi² = 3832,946, df = 45, p < 0.0001). The component plot in rotated space (Figure 1) showed the two factors of the CPAC. The two factors explained 50,9%.

Although the analysis supported Miller and Dumford (2016), new labels were used: For component 1, primarily named deliberate, the label act was used, as all the cognitive processes mentioned are conscious and active. For the second component, primarily labelled intuitive processes, flow was used, because all items were typical elements of a flow experience (Csikszentmihalyi, 1996, 2015). The latter was even used for the subscale of the long version questionnaire (Miller, 2014).

There were no correlations with gender (Table 2; Figure 2). Age, however, correlated negatively with self-reported creativity with a correlation coefficient of –.14; (p < 001, N = 2142; Figure 3). This correlation nevertheless seems substantially meaningless, only significant because of a very large sample size, as the two variables have had less than 2% common variance. Consequently, in Table 2 only the correlation of .38 was regarded with substantive importance—and even .38 means there is only 14% common variance.

TABLE 2 Correlation of subscales with gender and age. Spearman-Rho

Spearman-Rho		Deliberate	Flow	Gender	Age
Deliberate	Correlation Coefficient	1.000	.380**	−.026	−.128**
	Sig. (2-tailed)	.	<.001	.231	<.001
	N	2136	2136	2136	2136
Flow	Correlation Coefficient		1.000	.021	−.133**
	Sig. (2-tailed)		.	.329	<.001
	N		2136	2136	2136
Gender	Correlation Coefficient			1.000	−.001
	Sig. (2-tailed)			.	.952
	N			2706	2706
Age	Correlation Coefficient				1.000
	Sig. (2-tailed)				.
	N				2706

** The correlation was significant (2-sided) at the 0.01 level.

FIGURE 2 Gender differences in the two subscales of the CPAC: There is no gender effect.

* Act is calculated without item01 and item02

FIGURE 3 Age differences according to the two subscales of the CPAC: significant decline of perceived creativity with age, particularly with regard to subscale FLOW.

Discussion

Two factors were sufficient to summarize creativity, as originally measured by the CPAC. Miller and Dumford (2016) labelled their factors deliberate and intuitive, but for component 1 the label act was used, as the mentioned cognitive processes were conscious and active operations that could be trained and taught. For component 2, the label flow was used because all items were typical elements of a flow experience. Even Miller (2014) named this set of items flow in the long version of the questionnaire. The Flow factor might help to measure an important element of students’ experiences at school.

Creativity is understood as a process of sensitization for problems or gaps in knowledge, searching for solutions and even contribute to a Deeper Learning (Chow, 2010). So far, art education has seen itself as a lawyer of creativity at schools (Brodbeck, 1999). Creative people report discovery processes as a most enjoyable experience, bridging creativity and science as closely related fields (Csikszentmihalyi, 2015). In contrast, subjects associated with negative perceptions learning difficulties especially when STEM is focusing on such as STEM might be a result of subject issues and self-efficacy differences were undervalued (Conradty & Bogner, 2016; Epstein & Fischer, 2017; Schumm & Bogner, 2016). STEM curricula may benefit from the integration of arts and/or creative aspects, thus encouraging creative solutions (Henriksen, 2014). STEAM (STEM & Arts) may help to develop critical thinking also in real-world problems and what is more, to help to make learning science easier. STEAM may improve its image and support motivation. Creativity is not a domain exclusive to art, but a mental ability required for all spheres of life and (perhaps especially) for scientific research. Studying its effects on creative thinking skills should be of special interest in science education, even across multiple disciplines within higher education (Miller & Dumford, 2016). Art might be an archetype of creativity but it is only one product of creativity among many: “Creativity is neither reserved to the genius [as assumed in arts], nor can creativity be produced technically” (Brodbeck, 1999). STEAM might help to transfer enthusiasm and support individual self-efficacy from art to STEM and in this way to close a creativity gap (Runco, 2017). Some researchers request an overlap of creativity, thinking, and intelligence. Several methods, developed to foster, teach, or train creativity, point to the learnability of creativity (Bono De, 1990; Vester, 2007). Educators could promote both creativity and intrinsic motivation, by teaching students to solve problems that have no well-defined solutions. Popular approaches are synectics, to develop creative capacity (Gordon, 1961); brainstorming, “using the brain to storm a problem” (Osborn, 1963, cited in Merrotsy 2017), reducing self-censorship, mind mapping/concept mapping, used in combination with brainstorming (Novak & Cañas, 2006; Novak & Gowin, 1984), and the six thinking hats, to use and coordinate parallel thinking processes more effectively (Bono De, 1999). All these methods are respected variations based on the five phases model of creativity: preparation, incubation, eureka effect, evaluation, elaboration (Csikszentmihalyi, 1996).

Brodbeck (1999) regarded creativity techniques as ineffective, because otherwise in his argumentation they would be used by profit-maximizing companies more frequently as an economic resource. He rejects both the irrational belief in the simple givenness of genius, and the technocratic conceptions of feasibility of creativity but emphasized creativity as a mind-set. The potential for creativity is directed at an early stage by rules—which can be meaningful (Csikszentmihalyi, 1975). Adolescents forget to play with thoughts, and forget the pleasure of seeing, discovering, and listening (Brodbeck, 1999). As a result, adults accumulate a lack of creative thinking abilities. Learning environments following the theories of constructivism expanded with more open learning environments, where a learner can follow his/her own conceptions may be a guideline for curricula integrating creativity. The degree of autonomy and support were understood to be climate dimensions that are effective predictors of creative performance (Hunter, Bedell, & Mumford, 2007). Creativity can be aroused and even trained to preserve the ability for creative (flow) through exercise to age (Bono De, 1990; Ma, 2009).

Studies of the factors that make some people more creative than others have postulated the four Ps—process, product, person, and press—with press considered as environmental press also known as place as the socialcultural environment in which creativity flourishes (Rhodes, 1961). Place and person as an educational setting may promote the process of creativity with an appropriate degree of autonomy, access to resources, and the role of the teacher as a gatekeeping tutor. Many school systems are accused of contributing to the premature deterioration of creativity by concentrating too much on knowledge acquisition (Csikszentmihalyi, 1988; Robinson & Azzam, 2009). Anderson (1964) felt that “creativity is crossing out mistakes” (cited in Torrance, 1995, p. 147). This underlines that creativity cannot only help with problem solving, but also with identifying problems where others have failed to do so (Robinson & Azzam, 2009). For example, creativity was successfully applied to teach additional skills, such has peacebuilding skills with brain writing, a creative technique similar to brainstorming (Olugbenga, 2016). Both creativity and the focused skill were developed and especially underprivileged pupils benefited. Current curricula emphasize the importance of promoting students’ “desire” (motivation) beyond to traditional transfer of “knowledge” (cognition) and “ability” (competence orientation; e.g., Lehrplan PLUS, ISB 2017).

It is well documented that creativity and intrinsic motivation are mutually dependent (Csikszentmihalyi, 1996; Jesus De, Rus, Lens, & Imaginário, 2013; Runco, 2014). Amabile and Tighe (1993) described creativity as driven by intrinsic motivation. Even at a neuronal level, a link with motivation was found with dopamine-related activities as a determinant of human creativity (Aberg et al., 2016). This is a promising starting point for research into fostering school motivation (Aberg et al., 2016), as creativity may support intrinsic motivation in (science) classes. Furthermore, recent evidence has found neurophysiological gender differences (Abraham, 2016; Tyan et al., 2017). From a scientific point of view, this gap might be seen as chance to analyze and understand processes of creativity. As sociodemographic factors play a role in gender differences in creative thinking (Matud, Rodríguez, & Grande, 2007), the importance of supporting both genders appropriately must be recognized. Csikszentmihalyi (1996) even postulated that not gender, but a traditional gender differentiating (discriminating) education of boys and girls determines how they develop. This cultural discrimination might explain the varying results of international studies about gender differences in creativity.

With a deeper understanding of what creativity is and its implications for culture and economy in the 21st century, methods of teaching and educating children need to change. A promising approach is the promotion of motivation and creativity. Especially in the development of the STEM subjects to STEAM, promising results have already been reported. Especially with respect to flow, the presented questionnaire might suffice for educational settings; and act comprises teachable cognitive processes to promote creativity. The instrument may successfully measure both the learning effect of creativity skills and emotional factors that contribute to enjoyment in learning. According to Batey and Hughes (2017) cognitive ability measures rarely relate to self-perceptions of creativity, but with openness to experience and vary relative to the type of self-perception (trait, process, product), the domain of the self-perception (e.g., arts vs. science), or culture. Thus, the implemented modified questionnaire with act and flow as constructs might provide an appropriate psychometric tool to monitor educational settings promoting creativity. As creativity offers clear benefits for individuals and society as a whole (Runco, 2004), today the consensus between creativity experts is that the most urgent next step is to integrate the numerous findings into a coherent theoretical framework. The fact that creative thinking is based on normative cognitive processes emphasizes the need for a close collaboration between cognition domain experts and educational instructors to build a common neurocognitive framework of creative thinking (Kröger, 2015). The importance of creativity in learning, productivity, and mental health cannot be ignored.

ACKNOWLEDGMENTS

This work was supported by the European HORIZON-2020 framework labelled CREATIONS: Developing an Engaging Science Classroom (Grant Agreement No.665917). We would like to thank all students and teachers who supported this study.

Additional information

Funding

This work was supported by the European HORIZON-2020 framework labelled CREATIONS: Developing an Engaging Science Classroom; [Grant Agreement No.665917].