RESEARCH ESSAY due in 48 hours

Learning and Individual Differences 51 (2016) 291–298

Contents lists available at ScienceDirect

Learning and Individual Differences

journal homepage: www.elsevier.com/locate/lindif

Do award-winning experiences benefit students' creative self-efficacy and creativity? The moderated mediation effects of perceived school support for creativity

Shu-Hsuan Chang a, Chih-Lien Wang a, Jing-Chuan Lee b,⁎ a Department of Industrial Education and Technology, National Changhua University of Education, Taiwan b Graduate Institute of Educational Leadership and Evaluation, Southern Taiwan University of Science and Technology, Taiwan

⁎ Corresponding author at: Rm. N104, No.1, Nantai S Taiwan.

E-mail address: [email protected] (J.-C. Lee).

http://dx.doi.org/10.1016/j.lindif.2016.09.011 1041-6080/© 2016 Elsevier Inc. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history: Received 3 December 2015 Received in revised form 12 August 2016 Accepted 16 September 2016

The study aims to investigate the effects of students' perceived school support for creativity and creative self-ef- ficacy on their individual creativity. It further examines the mediation of creative self-efficacy and the moderated mediation of perceived school support for creativity. A survey was conducted over a sample of 720 Design- majored students from universities of science and technology in Taiwan, among which 335 have won awards for their work in creative design (with award-winning experience), and 385 have none (without award-winning experience). The results were analyzed with a path analysis-based moderated mediation analysis to determine whether the hypotheses were supported. Research findings include: (1) students' perceived school support for creativity had significant positive effects on both their creative self-efficacy and individual creativity; (2) per- ceived school support for creativity affected individual creativity via the mediation of creative self-efficacy; the mediation effects were higher on students with award-winning experiences(s) than on those without; (3) per- ceived school support for creativity had the effect of moderated mediation between their creative self-efficacy and individual creativity for students with award-winning experience(s); the effect was absent for students without award-winning experience. Based on the above findings, the implications and suggestions are provided for future research.

© 2016 Elsevier Inc. All rights reserved.

Keywords: Creative self-efficacy Individual creativity Moderated mediation model Perceived school support for creativity

1. Introduction

An environment supportive of creativity can potentially enrich life experiences of an individual and spark his/her individual creativity(IC). In our current 21st century, we have entered the so-called “Creative Age”, when creativity has become the common goal pursued by almost all developed countries in the world and the development of which has marked the general direction where education of every country strives toward (Florida, 2004). As a result of that, the cultivation of student cre- ativity and the development of student potential are both now essential topics in personnel training. The importance of research in creativity has never been more pronounced.

In recent years, scholars have unanimously emphasized the signifi- cant impact the interaction between individual and environment has on creativity (Amabile, 1996; Gong, Huang, & Farh, 2009; Hirst, Van Knippenberg, & Zhou, 2009; Jaussi, Randel, & Dionne, 2007; Oldham & Cummings, 1996; Shalley, Zhou, & Oldham, 2004). Many researchers have noticed that the generation or manifestation of creativity is not

t., Yongkang Dist., Tainan City,

simply a matter pertaining to the creative individual themselves, but is at the same time very much influenced by the environment and cul- ture the individual is in or has come from; environment plays the role of either nurturing/enhancing creativity or inhibiting it (Amabile, 1996; Csikszentmihalyi & Wolfe, 2000). On the other hand, creativity plays a crucial role in the cultural life of people; it is essential to the de- velopment and advancement of human civilization (Hennessey & Amabile, 2010). Amabile and Pillemer (2012) therefore suggested re- searches on creativity be conducted from the perspective of social psy- chology, with particular emphasis on the overall development of an individual's creative behaviors within social context. In view of the fore- going, it is clear that in order to better understand students' IC, it is es- sential to consider the school environment (contextual factor), as well as the individual factors on the part of the students.

Amabile (1996) maintained, in her componential theory of creativi- ty, that a “supportive social environment” directly affects an individual's intrinsic motivation to be creative. This environment can incorporate other extrinsic motivators to influence the creative process of an indi- vidual. According to Bandura's social cognitive theory, self-efficacy is a major mechanism in moderating motivation and behavior (Bandura, 1977). Creative self-efficacy (CSE) was defined by Tierney and Farmer (2002) as one's belief that “one has the ability to produce creative

292 S.-H. Chang et al. / Learning and Individual Differences 51 (2016) 291–298

outcomes” and was considered as an efficacy belief specific to creative performance (Tierney & Farmer, 2004). Research studies have proven that CSE positively predicts creative performance and the generation of creative ideas/products (Gong et al., 2009; Jaussi et al., 2007; Tierney & Farmer, 2002, 2004). Up until now, most studies have focused on the development of group/team creativity and organizational crea- tivity (Amabile & Khaire, 2008; Gaggioli, Mazzoni, Milani, & Riva, 2015), the studies about IC were comparatively few. In the field of edu- cational psychology, researchers investigated mainly the roles of teachers and peers (Karwowski, 2011, 2015; Karwowski, Gralewski, & Szumski, 2015), studies relating to the effect(s) of student's perceived school support for creativity (PSSC) (i.e., a student's subjective percep- tion concerning the extent to which school recognizes, supports, and encourages their creativity) on their IC were also scarce. Currently, the effects of perceived organizational support for creativity on employee creativity have been confirmed in many studies (Choi, 2004, 2012; Shalley et al., 2004; Tierney & Farmer, 2002, 2004), though they mostly focused only on the contextual factors of leader encouragement and peer support. A study becomes theoretically and empirically more sig- nificant if its research structure incorporates mediation and moderation simultaneously (Muller, Judd, & Yzerbyt, 2005). In light of that, this study investigates the mediation of CSE (personal factor) between PSSC (contextual factor) and IC. It further examines whether personal factors are influenced by the moderated mediation of contextual factors.

Many sources contribute to the development of a person's self-effi- cacy, including rewards from past performances, personal life experi- ences, and social influences (Hoy & Miskel, 2001). However, its development is commonly attributed to four major sources: perfor- mance accomplishments (enactive mastery experiences), vicarious ex- periences, verbal persuasion, and physiological/emotional states (Bandura, 1997). Among them, the most influential is enactive experi- ences, one's authentic mastery experiences directly related to the job; past successes raise self-efficacy, while failures lower it. In the motiva- tional process, an individual's self-efficacy appraisal could constitute a concept of self-management; CSE could greatly inspire an individual's intrinsic motivation (Bandura, 1997). The strong belief in one's ability to do well on a given task could enhance one's intrinsic motivation and enable one to work with devotion voluntarily and joyously. In addi- tion, work environment attributes (e.g. reward, support, warmth, iden- tity, etc.) directly or indirectly perceived by working members (Litwin & Stringer, 1968) may also affect their motives and behaviors. Gralewski and Karwowski (2013a, 2013b) have noticed that participations in com- petitions and thematic Olympics were positively associated with stu- dents' creative self-perception. The current study takes into consideration the above claims and investigates whether there is a dif- ference in the strength of mediation of CSE between students with and without award-winning experience(s) (AWE), and whether these effects are affected by the moderated mediation of students' PSSC. The AWE, in the context of our study, refers to students' experience of win- ning award in national and/or international creative design-related competitions.

2. Theoretical background and hypotheses development

2.1. Mediation of creative self-efficacy

Amabile (1983, 1996) stated in her componential theory of creativ- ity that “creativity is best conceptualized… as a behavior resulting from particular constellations of personal characteristics, cognitive ability, and social environment”. For college students, this social/organizational environment would be the college/university they attend. Lubart (2008) has noted the great individual differences in creativity, and that some schools may be better at enhancing student creativity than others. A creative school climate, i.e., a school culture supportive of cre- ativity, can provide the necessary support for the creativity of its stu- dents (Amabile, Conti, Coon, Lazenby, & Herron, 1996; Oldham &

Cummings, 1996). Research has found that the level of trust in the class- room/school (interpersonal component of creative climate) could influ- ence students' CSE (Karwowski, 2011). Organizational encouragement of idea generation, freedom of expression, and etc. has been shown to have the effect of stimulating its members to engage in creative behav- iors (Amabile & Gryskiewicz, 1989). Both creative school climate and teacher's intrinsic motivation have significant direct positive effects on teacher's creative teaching performance (Hsiao, 2011).

When Bandura first proposed the concept of self-efficacy, it was pos- ited as the cognitive mechanism that plays the main role in influencing the psychological functioning of an individual (Bandura, 1977). High self-efficacy was believed to be a prerequisite to creative output or dis- covery of new knowledge (Bandura, 1997). Beghetto (2006) found that higher levels of CSE were linked to students holding more positive be- liefs about their academic abilities in all subject areas. Studies have con- firmed that individual factors and work environment jointly affect an individual's CSE (Chong & Ma, 2010), and that there is a positive rela- tionship between organizational climate and self-efficacy (Liang & Wei, 2013).

The causation of IC is not limited to the intricate interactive results of psychological processes such as individual cognition, motive, attitude, emotion, and etc.; in the process of self-regulation, the self-efficacy of an individual plays an important role of mediation (Bandura, 1977, 1982). Tierney and Farmer (2002, 2004) observed that CSE is predictive of individual creative performance; and is a significant mediator in the relationships between supervisors' expectations of employee creativity and employee creativity, supervisors' creativity-supportive behaviors and employee creativity, and employees' view of creativity expectations and employee creativity. By integrating the effects of environment and the psychology of college students, it is possible to predict students' sci- entific imagination via the mediation of their self-efficacy (Hsu, Liang, & Hsu, 2014). The mediation of self-efficacy has been found to enable the intrinsic motivation of university science majors to significantly affect their imagination (Liang & Chang, 2014). Self-efficacy also mediates the effects of organizational climate on employee creativity (Liang & Wei, 2013). Contextual factors such as the social influences from super- visors and colleagues can affect individual creative performance through the mediation of CSE (Choi, 2004, 2012). Inferring from the aforementioned findings, this study proposes the following hypotheses:

H1. Students' PSSC will positively affect their IC. (X → Y, c).

H2. Students' PSSC will positively affect their CSE. (X → M, a1).

H3. Students' PSSC will affect their IC via the mediation of their CSE. (X、M → Y, c′, b1).

2.2. Moderated mediation of perceived school support for creativity

The interactionists hold that it is impossible to truly understand the emergence of each specific individual behavior by analyzing the behavior singularly from personal or contextual perspective alone; the behavior needs to be examined from the standpoint of interactions between per- sonal and contextual factors (Schneider, 1990). Runco (2007) posited or- ganizational climate as a way to manipulate the effects of social support. It is suggested that climate dimensions are effective predictors of creative performance (Hunter, Bedell, & Mumford, 2007), and that creative people appear particularly reactive to climate variables (Oldham & Cummings, 1996). Hirst and colleagues proposed that team context, in fostering indi- vidual learning, facilitates IC (Hirst et al., 2009). Conversely, unsupportive social climate negatively affects IC (Choi, Anderson, & Veillette, 2009). When appraising self-efficacy, one instinctively assesses the resources and limitations of self and environment in terms of knowledge, informa- tion, experience, support, etc., and then form a belief of one's having high or low coping capacity based on those assessments (Gist & Mitchell, 1992). Each individual, therefore, has their own unique way of cognizing

Perceived school support for

creativity (X)

Individual creativity (Y)

c

Perceived school support for

creativity (X)

Individual creativity (Y)c'

Creative self- efficacy (M)

a1 b1

b2

Fig. 1. Conceptual model.

293S.-H. Chang et al. / Learning and Individual Differences 51 (2016) 291–298

environmental support (PSSC in our case) and situational change, interpreting personal experiences, and behaving accordingly. It has been recognized that contextual factors, as individually/subjectively perceived by the students and reflecting their comprehensive beliefs about the school's organizational behavior, may be measured by self-perception in- struments (Choi, 2004, 2012; Karwowski, 2011, 2015; Karwowski et al., 2015).

Empirical studies have found that: perceived support for creativity moderates the relationship between CSE and self-perceived creativity (Diliello, Houghton, & Dawley, 2011); and organizational support moder- ates the relationship between organizational justice and employee inno- vative behaviors (Hsu & Chiu, 2011). Studies on moderated mediation have shown that: school creative climate affects creative teaching perfor- mance via the mediation of teachers' intrinsic motivation (Hsiao, 2011); and students' PSSC positively moderated the mediated relationship be- tween their savoring capacity and IC via their CSE (Chang, Wang, & Lee, 2015). Accordingly, the current study takes into consideration the effects of both mediation and moderation to examine the conditional indirect ef- fects (moderated mediation) of students' PSSC. Students' CSE, in this con- text, will serve as both the mediator and the moderated variables simultaneously. We thus hypothesized as follows:

H4. Students' PSSC will positively moderate the relationship between their CSE and their IC. (XM → Y, b2).

H5. The moderation of Students' PSSC will affect their IC via the media- tion of their CSE. (a1(b1 + b2X) → Y, c′).

M

X Y

XM

a

c'

b1

b2

eM

eY

Fig. 2. Path diagram of moderated mediation as proposed by Preacher et al. (2007; p. 194). X = independent variable; M = mediator; XM = moderator; Y = dependent variable; a1, b1, b2, and c′ = regression coefficients.

3. Research design

3.1. Participants

The present study selected fourth-year design-majored students from several universities of science and technology in Taiwan as re- search participants. They were chosen to take the survey based on their experience in taking design courses. A stratified purposive sam- pling was used as the sampling method. After obtaining the consent from 13 universities to participate in the study, a total of 850 copies of survey questionnaires were sent to the schools by mail. The resulting 720 valid samples consisted of 67% females and 33% males; 87% pri- vate-university and 13% public-university students; 47% students with AWE and 53% students without.

3.2. Research framework and method

Statistical analyses were conducted using AMOS 22.0 and SPSS 20.0. Confirmatory factor analyses (CFA) were performed to examine the con- vergent validity of our measurement model and the content of the good- ness-of-fit index (GFI). The method of combining both mediator and moderator variables (i.e., moderated mediation) has been suggested by re- searchers Baron and Kenny (1986) and James and Brett (1984). Moderated mediation refers to instances in which mediation on every level (direct, in- direct, or simultaneous) is influenced by another moderator variable (Edwards & Lambert, 2007; Muller et al., 2005; Preacher, Rucker, & Hayes, 2007). As illustrated in Fig. 1, this study employed path diagram Model 1 presented by Preacher and colleagues (Preacher et al., 2007) as the framework of our theoretical model, taking CSE as the mediator vari- able (M), and PSSC as both the independent variable (X) and moderator variable (XM). The paths of this moderated mediation model are shown in Fig. 2. In order to simultaneously examine all paths proposed in the the- oretical model and to avoid type I error and the problems of statistical strength, survey data were analyzed using Andrew Hayes' PROCESS macro Models 4 and 74 to examine the effects of mediation and moderat- ed mediation respectively (Hayes, 2013). Subsequently, we utilized the Sobel test and bootstrapping to obtain the confidence interval (CI) and to determine whether it included 0: if it did, the indirect effect would be

deemed insignificant; if it did not, the effect would be considered signifi- cant. The mediation of CSE in students with and without AWE, and the moderated mediation of PSSC were each examined following this procedure.

3.3. Measures

3.3.1. Measure of creative self-efficacy We used the 4-item, 7-point Likert CSE scale developed by

Tierney and Farmer (2002) to assess students' CSE with questions such as “I feel that I am good at generating novel ideas” and “I have confidence in my ability to solve problems creatively”. In evaluating reliability and validity, the results of the CFA showed that the t-value of all items on the measure were at the 0.001 level of significance; factor loadings were between 0.878–0.912; composite reliability (CR) was 0.943; and average variance extracted (AVE) was 0.805. The fact that factor loadings were N0.7, CR was N0.7, and AVE exceeded the standard value of 0.5 (Hair, Black, Babin, & Anderson, 2009) demonstrated the adequate convergent validity of the mea- sure. In terms of model fit, RMSEA = 0.016, it was ≤0.06 and thus reflected a good fit (Hu & Bentler, 1999); GFI = 0.999, AGFI = 0.992, NFI = 0.999, and CFI = 0.999, all achieved values greater than the standard of 0.950 (Bentler, 1995; Hu & Bentler, 1999); the overall model fit was almost perfect.

Table 2 Descriptive statistics and t-test of difference between AWE and Non-AWE.

Variables Students M SD t Skewness Kurtosis

PSSC (X) AWE 2.802 0.622 −0.189 −0.425 −0.230 Non-AWE 2.811 0.583 −0.243 −0.168 Total 2.807 0.601 −0.338 −0.192

CSE (M) AWE 5.192 1.137 6.137⁎⁎⁎ −0.425 −0.559 Non-AWE 4.664 1.162 −0.165 −0.093 Total 4.910 1.179 −0.273 −0.378

IC (Y) AWE 4.338 0.853 6.159⁎⁎⁎ −0.185 −0.765 Non-AWE 3.941 0.873 0.228 −0.473 Total 4.126 0.886 0.026 −0.722

Note. N = 720; n = 335 (Award-winning); n = 385 (Non-award-winning). M = mean, SD = standard deviation; PSSC = perceived school support for creativity; CSE = creative self-efficacy; IC = individual creativity. ⁎⁎⁎ p b 0.001.

294 S.-H. Chang et al. / Learning and Individual Differences 51 (2016) 291–298

3.3.2. Measure of perceived school support for creativity In line with the purpose of this study, the “organizational encourage-

ment” section of KEYS: Assessing the Climate for Creativity developed by Amabile and colleagues (Amabile et al., 1996) was used as the instru- ment to assess students' PSSC. The measure comprised of eight 4-point Likert scale items with questions such as “People are encouraged to solve problems creatively in this school”, and “People are rewarded for creative work in this school”. The results of the CFA showed that the t- value of all items on the measure were at the 0.001 level of significance; factor loadings were between 0.683–0.794; CR was 0.910; and AVE was 0.556; adequate convergent validity of the measure was thus demon- strated (Hair et al., 2009). The overall model fit was satisfactory (RMSEA = 0.060, NFI = 0.926, and CFI = 0.943, GFI = 0.898, AGFI = 0.877) (Bentler, 1995; Hu & Bentler, 1999). Although GFI and AGFI were not N0.90, they nevertheless fell within the reasonable range of ≥0.80 and thus can still be deemed acceptable (Hu & Bentler, 1999). This study examined its aggregate effect showed that ICC(1) value of PSSC was low, and ICC(2) value was generally lower than the 0.70 criteria. Thus, this study adopted self-perception instrument with well validity and reliability to measure school support for creativity and mea- sured as the Level-1 variable.

3.3.3. Measure of individual creativity A 9-item 6-point Likert IC scale (Tierney, Farmer, & Graen, 1999) was

used to evaluate the IC of students with questions such as “I demon- strated originality in my work”, and “I took risks in terms of producing new ideas in doing my job”. The results of the CFA showed that the t- value of all items on the measure were at the 0.001 level of significance; factor loadings were between 0.770–0.831; CR was 0.944; and AVE was 0.650; all demonstrated the adequate convergent validity of the mea- sure (Hair et al., 2009). The overall model fit was almost perfect (RMSEA = 0.059, GFI = 0.978, AGFI = 0.953, NFI = 0.984, and CFI = 0.988) (Bentler, 1995; Hu & Bentler, 1999).

3.3.4. Discriminant validity between measures To ensure that every aspect (PSSC, CSE, or IC) of the measures could

be effectively distinguished, the discriminant validity between mea- sures was assessed using the technique proposed by Fornell and Lacker (1981). This would mean that the AVE of each variable need to be greater than the square value of the correlation coefficient between them. As shown in Table 1, all AVE values are indeed greater than the square value of the correlation coefficient between variables, and there- by confirming adequate discriminant validity. Furthermore, the correla- tion between the PSSC, CSE and IC scales for all students as well as for students with/without AWE were all statistically significant (p b 0.01), signifying that the data are fit for hypothesis testing.

Table 1 Square root of AVE, correlations between variables, and factor loading.

AWE students n = 335

Non-AWE students n = 385

Total students N = 720

Variables X M Y X M Y X M Y

PSSC (X) 0.757 0.737 0.747 CSE (M) 0.370 0.896 0.157 0.890 0.250 0.897 IC (Y) 0.396 0.776 0.800 0.236 0.741 0.798 0.303 0.768 0.806 FL(Min) 0.701 0.884 0.759 0.662 0.867 0.768 0.683 0.886 0.770 FL(Max) 0.821 0.909 0.830 0.792 0.922 0.831 0.794 0.914 0.831 CR 0.915 0.942 0.941 0.905 0.938 0.941 0.910 0.943 0.944 AVE 0.573 0.803 0.640 0.543 0.792 0.637 0.558 0.805 0.650 Cronbach's α 0.984 0.918 0.930 0.881 0.912 0.929 0.887 0.919 0.933

Note. The on-diagonal values shown in boldface are the square root of AVE; the off-diag- onal values are the correlation coefficient (**p b 0.01); PSSC = perceived school support for creativity; CSE = creative self-efficacy; IC = individual creativity. FL(Min) = minimum factor loading; FL(Max) = maximum factor loading.

4. Results and analysis

4.1. Descriptive statistical analysis

Table 2 illustrates that the mean scores of CSE and IC from students with AWE were significantly higher than those of without AWE. Inter- estingly, the mean score of students without AWE in PSSC was non-sig- nificant difference that of students with AWE.

4.2. Hypotheses testing for mediation

In testing the mediation of CSE proposed in H3, we followed the pro- cedure suggested by Baron and Kenny (1986). Data were analyzed using Hayes's PROCESS macro Models 4 (Hayes, 2013).

The regression coefficients of mediation are displayed in Table 3. For students with AWE, Model 1 shows that their PSSC had a significant positive effect on their IC (Y) (c = 0.539, p b 0.001; R2 = 0.155, p b 0.001); H1 was therefore supported. Model 2 shows that their PSSC had a significant positive effect on their CSE (a1 = 0.661, p b 0.001; R2 = 0.131, p b 0.001), H2 was therefore also supported. Model 3 demonstrates that their CSE had a significant positive effect on their IC (b1 = 0.540, p b 0.001; R2 = 0.605, p b 0.001); and although the regression coefficient of their PSSC on their IC decreased with the mediation of their CSE, it was still at the significance level (c′ = 0.182, p b 0.001), therefore, H3 too was supported. This partial mediation was quantified using the difference-in-coefficients method developed by Judd and Kenny (1981), Bindirect = 0.357 (a1 ∗ b1 = c-c′). The indirect effect in H3 was tested with Sobel, results showed that the z-value of CSE was 6.643 (p b 0.001); and bootstrap 95% CIs [0.507, 0.873] did not include 0, the indirect effect was therefore proven significant. The strength of mediation was at 66.18%.

For students without AWE, Model 1 shows that their PSSC had a sig- nificant positive effect on their IC (c = 0.339, p b 0.001; R2 = 0.051, p b 0.001), H1 was therefore supported. Model 2 shows that their PSSC had a significant positive effect on their CSE (a1 = 0.282, p b 0.01; R2 = 0.020, p b 0.01), H2 was therefore also support. Model 3 demonstrates that their CSE had a significant positive effect on their IC (b1 = 0.539, p b 0.001; R2 = 0.554, p b 0.001); and although the re- gression coefficient of their PSSC on their IC decreased with the media- tion of their CSE, it was still at the significance level (c′ = 0.187, p b 0.001), therefore, H3 too was supported. This partial mediation was quantified using the difference-in-coefficients method developed by Judd and Kenny (1981), Bindirect = 0.152 (a1 ∗ b1 = c-c′). The indirect effect in H3 was tested with Sobel, the results showed that the z-value of CSE was 2.764 (p b 0.01) and bootstrap 95% CIs [0.153, 0.717] failed to include 0, the indirect effect was therefore proven significant. The strength of mediation was at 44.75%.

Based on the above, H3 was supported by the results of both stu- dents with and without AWE: students' PSSC, mediated by their CSE, in- deed affected their IC. However, in comparing the strengths (66.18% N

Table 3 Regression coefficients for the mediation models of creative self-efficacy.

Variable Award-winning students (n = 335) Non-award-winning students (n = 385)

Individual creativity (Y) CSE (M) Individual creativity (Y) CSE (M)

Model 1 Model 3 Model 2 Model 1 Model 3 Model 2

B p B p B p B p B p B p

Constant 2.828 0.000 1.025 0.000 3.341 0.000 2.988 0.000 0.903 0.000 3.873 0.000 PSSC (X) 0.539 0.000 0.182 0.000 0.661 0.000 0.339 0.000 0.187 0.000 0.282 0.006 CSE (M) 0.540 0.000 0.539 0.000 R2 0.155 0.000 0.605 0.000 0.131 0.000 0.051 0.000 0.554 0.000 0.020 0.006 Quantifying indirect effects

c = 0.539 a1 = 0.661 c = 0.339 a1 = 0.282 c′ = 0.182 b1 = 0.540 c′ = 0.187 b1 = 0.539 c-c′ = 0.357 a1 ∗ b1 = 0.357 c-c′ = 0.152 a1 ∗ b1 = 0.152

Testing indirect effects Sobel test (Z) 6.643 0.000 2.764 0.006 Bootstrap 95% (CI) [0.507, 0.873] [0.153, 0.717] Strength of mediation, SM (%) 66.18% 44.75%

Note. n = 720. Unstandardized regression coefficients are reported. Bootstrap sample size = 10,000. PSSC = perceived school support for creativity; CSE = creative self-efficacy. CI = confidence interval.

295S.-H. Chang et al. / Learning and Individual Differences 51 (2016) 291–298

44.75%) and the quantifications (0.357 N 0.152) of mediation, students with AWE were higher in both than the ones without. The estimated re- sults of this mediation model are shown in Fig. 3.

4.3. Hypotheses testing for moderated mediation

In testing the moderated mediation of PSSC proposed in H5, data were analyzed using PROCESS Model 74 of Hayes (2013) to see if they supported the hypotheses. We estimated and examined the conditional indirect effect(s) of this moderated mediation model as follows: (1) PSSC has to significantly predict CSE (X → M, a1); (2) CSE has to signif- icantly predict IC (M → Y, b1); (3) the interaction between PSSC and CSE has to significantly predict IC (XM → Y, b2); (4) the moderation of PSSC has to significantly predict student creativity via the mediation of stu- dents' CSE (a1(b1 + b2X) → Y, c′).

The regression coefficients of moderated mediation are shown in Table 4. For students with AWE, the results illustrated that their PSSC significantly predicted their CSE (a1 = 0.661, t = 7.077, p b 0.001);

Table 4 Regression coefficients of the moderated mediation model of perceived school support for crea

Variables Award-winning students (n = 335)

CSE (M) IC (Y)

B t p B t p

Constant iM 3.341 12.471 0.000 iY 4.309 140.997 0.0 PSSC (X) a1 0.660 7.077 0.000 c′ 0.186 3.724 0.0 CSE (M) – – – b1 0.544 19.825 0.0 X × M – – – b2 0.116 3.060 0.0

R2 = 0.131 R2 = 0.616 F (1, 333) = 50.077 0.000 F (3, 331) = 177.064 0.0

Note. N = 720. Unstandardized regression coefficients are reported. Bootstrap sample size = 1

c'=.182***

CSE (M)

CSE (M)

PSSC (X)

PSSC (X)

IC (Y) IC (Y)

a1=.661*** b1=.540***

SM=66.18%

c=.539*** PSSC

(X)

PSSC (X)

IC (Y) IC (Y)

AWE Students (n = 335)

Fig. 3. Estimated results of the mediation of cr

their CSE significantly predicted their IC (b1 = 0.544, t = 19.825, p b 0.001); and the interaction between their PSSC and their CSE signif- icantly predicted their IC (b2 = 0.116, t = 3.060, p b 0.01). These proved that H4 was supported. Additionally, the moderation of PSSC signifi- cantly predicted IC via the mediation of CSE (c′ = 0.186, p b 0.001); the support for H5 was thereby also confirmed. This demonstrates that the moderation of students' PSSC has a positive effect on their IC via the mediation of their CSE. CSE serves not only as a mediator vari- able, but also a moderated variable.

For students without AWE, their PSSC significantly predicted their CSE (a1 = 0.282, t = 2.792, p b 0.01); their CSE significantly predicted their IC (b1 = 0.533, t = 20.188, p b 0.001); but the interaction between their PSSC and their CSE failed to significantly predict their IC (b2 = 0.049, t = 1.202, p N 0.05); H4 was therefore not supported. This sig- nifies that the moderation of PSSC did not have a positive effect on IC via the mediation of CSE. In this case, CSE served as a mediator variable, but not a moderated variable. The estimated results of this moderated mediation model are shown in Fig. 4.

tivity.

Non-award-winning students (n = 385)

CSE (M) IC (Y)

B t p B t p

00 iM 0.000 0.000 1.000 iY 3.936 131.030 0.000 00 a1 0.282 2.792 0.006 c′ 0.177 3.383 0.001 00 – – – b1 0.533 20.188 0.000 02 – – – b2 0.049 1.202 0.230

R2 = 0.02 R2 = 0.556 00 F (1, 383) = 7.793 0.006 F (3, 381) = 158.907 0.000

0,000. PSSC = perceived school support for creativity.

c'=.187**

CSE (M)

CSE (M)

PSSC (X)

PSSC (X)

IC (Y) IC (Y)

a1=.282** b1=.539***

SM=44.75%

c=.339*** PSSC

(X)

PSSC (X)

IC (Y) IC (Y)

Non-AWE Students (n = 385)

eative self-efficacy. **p b 0.01. ***p b 0.001.

c'=.186***

CSE (M)

PSSC (X)

IC (Y)

a1=.661*** b1=.544***

c'=.177**

CSE (M)

PSSC (X)

IC (Y)

a1=.282** b1=.533*** b2=.116** b2=.049

AWE students (n = 335) Non-AWE students (n = 385)

Fig. 4. Estimated results for the moderated mediation model of perceived school support for creativity. **p b 0.01. ***p b 0.001.

296 S.-H. Chang et al. / Learning and Individual Differences 51 (2016) 291–298

The conditional indirect effects (CIE) of PSSC were examined among students with AWE, using the bootstrapped 95% CI, to see whether the results were different between the times when it (PSSC) was high and when it was low: if the 95% CI of the CIE included 0, the CIE would be deemed insignificant; if it did not include 0, the CIE would be deemed significant. As illustrated in Table 5, at neither time did the 95% CI of the CIE include 0: when PSSC was low (M-1SD; 95%CI [0.214, 0.417]), and when PSSC as high (M + 1SD; 95%CI [0.272, 0.545]); which means that PSSC, regardless of it being high or low, had significant CIE on student creativity (see Fig. 5).

5. Conclusions and recommendations

5.1. Conclusion and discussion

5.1.1. Creative self-efficacy mediates the effect of perceived school support for creativity on individual creativity

The results of our analyses show that students with AWE have higher CSE and CI than those without AWE, but there is no difference between their PSSC. Students' PSSC is found to have a significant posi- tive direct effect on their IC, which is line with the result of earlier stud- ies (Amabile & Gryskiewicz, 1989; Hsiao, 2011); and to have a significant positive direct effect on their CSE, which also concurs with earlier findings (Chong & Ma, 2010; Liang & Wei, 2013). Our results con- firmed the positive indirect effect (partial mediation) of CSE between PSSC and IC, as did the results of earlier studies (Choi, 2004, 2012; Hsu et al., 2014; Liang & Chang, 2014; Liang & Wei, 2013; Tierney & Farmer, 2004); and in accordance with Amabile's assertion of contextu- al factors being able to affect individual outcome variables via the medi- ation of individual factors (Amabile, 1996).

Further examination showed that the strength of the mediation of CSE in students with AWE was higher than that in the ones without. This means that the stronger the PSSC, the greater it's strength in en- hancing the CSE of students with AWE, and thereby further augmenting their IC. It implies that PSSC is more helpful in enhancing IC via CSE in students with AWE. Such finding supports the conclusion of the second study of Karwowski (2014) which demonstrated the strong positive as- sociation between growth creative mindset (believing that creative tal- ents and abilities can be developed through dedication, effort, good teaching, etc.) and creative self-concept (e.g. creative self-efficacy). It reinforces the observation made by Beghetto (2006) that feedbacks on

Table 5 Regression results for the conditional indirect effects of perceived school support for creativity.

Standard deviation (SD) a (b1 + b2X) Bootstrap 95% CI

PSSC (XM) Boot CIE Boot SE LLCI ULCI

Low (M-1SD) −0.622 0.311 0.052 0.214 0.417 Mean (M) 0.000 0.359 0.058 0.246 0.472 High (M + 1SD) 0.622 0.407 0.070 0.272 0.545

Note. n = 335 (award-winning students). Unstandardized regression coefficients are re- ported. Bootstrap sample size = 10,000. PSSC = perceived school support for creativity, CIE = conditional indirect effect, SE = standard error, CI = confidence interval, LL = lower limit, UL = upper limit.

students' creative ability (school support) were positively related to stu- dents' CSE.

5.1.2. Perceived school support for creativity plays the role of moderated mediation on the effect of students' creative self-efficacy on their individual creativity

Moderated mediation of PSSC was found to happen only in students with AWE and not in the ones without. This indicates that the PSSC of students with AWE has a positive moderating effect on the relationship between their CSE and IC. The current finding coincides with the results of earlier research (Diliello et al., 2011; Hsu & Chiu, 2011). We also found that PSSC moderates the mediation of CSE, and that it has a CIE (moderated mediation) on IC. This finding supports the claims made by prior studies (Chang et al., 2015; Hsiao, 2011), and is also in line with the interactionist perspective that individual behavior is deter- mined by the interaction between personal and contextual factors (Schneider, 1990). This has significant implications in that it suggests that (1) by improving the PSSC of students with AWE, it is possible to in- crease their IC through the enhancement of their CSE; (2) the CSE of such students significantly enhances their IC; (3) by raising their PSSC, their IC will heighten accordingly.

As Bandura (1997) pointed out, self-efficacy allows an individual to apply and utilize his/her own ability and necessary resources to success- fully complete task(s) mandated in a given situation, thereby obtaining positive experiences. However, for students without AWE, their PSSC did not manifest the effect of moderated mediation in the current study. We suspect the reason to be their lack of positive experience (award-winning). The encouragement from their PSSC alone might not have been enough to boost their CSE and therefore fell short of hav- ing a significant interactive effect with it, and consequently failed to af- fect their individual creative performance. This claim has yet to be proven by future research to determine whether the number of an individual's positive experience indeed plays a part in affecting their IC.

Fig. 5. Moderated mediation of perceived school support for creativity (PSSC) of students with award-winning experiences.

297S.-H. Chang et al. / Learning and Individual Differences 51 (2016) 291–298

5.2. Theoretical and practical implications

CSE plays the dual roles of mediator and moderated variables in the current study. Reviewing the perspectives of empirical studies, social cognitive theory (Bandura, 1977) marked CSE as an important mediator between PSSC and IC; componential theory of creativity (Amabile, 1996) suggested that IC is affected by one's CSE, PSSC, and the interac- tive effects between the two (personal and contextual factors respec- tively). The results of the present study provided empirical support for the forenamed theories.

This study also found that CSE mediates the relationship between PSSC and IC in both students with and without AWE, and that the strength of mediation was higher in students with AWE than in the ones without. It may be inferred that the confirmation and confidence gained through the experience(s) of winning competitions might have the effect of raising the CSE of AWE students, thereby encouraging them to become more devoted while working, persevere through ob- stacles, be more resilient against stress and fatigue, and be more persis- tent in completing tasks; and therefore result in the higher mediatory effect of their CSE. Based on the findings of our study, we devised several recommendations for schools: 1. establish systems and methods that encourage the creativity of all school members so that a creative climate becomes the norm; 2. Design and offer positive psychology courses that may enhance the CSE of students, especially the ones without AWE; 3. Encourage students to participate in school, national, and/or interna- tional creativity competitions to help them gain confidence and enthu- siasm in working creatively; 4. Sponsor intercollegiate creativity exchange events so that the schools and students can be inspired by and learn from each other's creativity. Consequently, the CSE of students might develop and improve continuously by being immersed in the cre- ative climate of the school environment and through the encourage- ment and stimulation of the bountiful opportunities to compete and perform creatively.

5.3. Limitations and future research

Despite the value in its theoretical and educational implications, and our effort in being meticulous in the process of its execution, the current study nevertheless has the following limitations: (1) The participants were limited to fourth-year design-major students from universities of science and technology in Taiwan, the results might not be generaliz- able to students of other years, majors, or countries; (2) The mea- sures/scales used were all developed by Western researchers that, in spite of their proven reliability and validity, inevitably are restricted by their social and cultural limitations, thereby opening the door to pos- sible measurement errors; (3) IC was gauged by self-report measures, so the responses were susceptible to social desirability bias. We recom- mend that future studies incorporate information/data from various sources, including objective indicators (such as creative works and win- ning records from creative competitions, etc.) or evaluations of creative performances (from teachers and team members, etc.), in their assess- ment of IC. This will enable the researchers to obtain a more compre- hensive evaluation of creativity, and to avoid probable measurement error stemming from social desirability bias.

References

Amabile, T. M. (1983). The social psychology of creativity: A componential conceptualiza- tion. Journal of Personality and Social Psychology, 45(2), 357–376.

Amabile, T. M. (1996). Creativity in context. Boulder, CO: Westview press. Amabile, T. M., & Gryskiewicz, N. D. (1989). The creative environment scales: Work envi-

ronment inventory. Creativity Research Journal, 2(4), 231–253. Amabile, T. M., & Khaire, M. (2008). Creativity and the role of the leader. Harvard Business

Review, 86(10), 100–109. Amabile, T. M., & Pillemer, J. (2012). Perspectives on the social psychology of creativity.

The Journal of Creative Behavior, 46(1), 3–15. Amabile, T. M., Conti, R., Coon, H., Lazenby, J., & Herron, M. (1996). Assessing the work en-

vironment for creativity. Academy of Management Journal, 39(5), 1154–1184.

Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist,

37(2), 122–147. Bandura, A. (1997). Self-efficacy: The exercise of control. New York, NY: Freeman. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social

psychological research: Conceptual, strategic, and statistical consideration. Journal of Personality and Social Psychology, 51(6), 1173–1182.

Beghetto, R. A. (2006). Creative self-efficacy: Correlates in middle and secondary stu- dents. Creativity Research Journal, 18(4), 447–457.

Bentler, P. M. (1995). EQS: Structural equations program manual. Encino, CA: Multivariate Software, Inc.

Chang, S. H., Wang, C. L., & Lee, J. C. (2015). The effects of savoring on individual creativity: The mediatory effects of creative self-efficacy and the moderated mediatory effects of perceived school support for creativity. Chinese Journal of Science Education, 23(4), 397–419.

Choi, J. N. (2004). Individual and contextual predictors of creative performance: The me- diating role of psychological processes. Creativity Research Journal, 16(2–3), 187–199.

Choi, J. N. (2012). Context and creativity: The theory of planned behavior as an alternative mechanism. Social Behavior and Personality, 40(4), 681–692.

Choi, J. N., Anderson, T. A., & Veillette, A. (2009). Contextual inhibitors of employee crea- tivity in organizations the insulating role of creative ability. Group & Organization Management, 34(3), 330–357.

Chong, E., & Ma, X. (2010). The influence of individual factors, supervision and work en- vironment on creative self-efficacy. Creativity and Innovation Management, 19(3), 233–247.

Csikszentmihalyi, M., & Wolfe, R. (2000). New conceptions and research approaches to creativity: Implications of a systems perspective for creativity in education. In K. A. Heller, F. J. Monks, R. J. Sternberg, & R. F. Subotnik (Eds.), International handbook of giftedness and talent (pp. 81–93) (2nd ed ). Oxford: Elsevier.

Diliello, T. C., Houghton, J. D., & Dawley, D. (2011). Narrowing the creativity gap: The moderating effects of perceived support for creativity. The Journal of Psychology, 145(3), 151–172.

Edwards, J. R., & Lambert, L. S. (2007). Methods for integrating moderation and media- tion: A general analytical framework using moderated path analysis. Psychological Methods, 12(1), 1–22.

Florida, R. (2004). Americas looming creativity crisis. Harvard Business Review, 37, 122–134.

Fornell, C., & Lacker, D. F. (1981). Evaluating structural equation models with unobserv- able variables and measurement error. Journal of Marketing Research, 18(1), 39–50.

Gaggioli, A., Mazzoni, E., Milani, L., & Riva, G. (2015). The creative link: Investigating the relationship between social network indices, creative performance and flow in blend- ed teams. Computers in Human Behavior, 42, 157–166.

Gist, M. E., & Mitchell, T. R. (1992). Self-efficacy: A theoretical analysis of its determinants and malleability. Academy of Management Review, 17(2), 183–211.

Gong, Y., Huang, J. C., & Farh, J. L. (2009). Employee learning orientation, transformational leadership, and employee creativity: The mediating role of employee creative self-ef- ficacy. Academy of Management Journal, 52(4), 765–778.

Gralewski, J., & Karwowski, M. (2013a). Polite girls and creative boys? students' gender moderates accuracy of teachers' ratings of creativity. The Journal of Creative Behavior, 47(4), 290–304.

Gralewski, J., & Karwowski, M. (2013b). Polite girls and creative boys? Students' gender moderates accuracy of teachers' ratings of creativity. The Journal of Creative Behavior, 47(4), 290–304.

Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2009). Multivariate data analysis: A global perspective (7th ed.). Upper Saddle River, NJ: Prentice Hall.

Hayes, A. F. (2013). An introduction to mediation, moderation, and conditional process anal- ysis: A regression-based approach. New York: Guilford press.

Hennessey, B. A., & Amabile, T. M. (2010). Creativity. The Annual Review of Psychology, 61, 569–598.

Hirst, G., Van Knippenberg, D., & Zhou, J. (2009). A cross-level perspective on employee creativity: Goal orientation, team learning behavior, and individual creativity. Academy of Management Journal, 52(2), 280–293.

Hoy, W. K., & Miskel, C. G. (2001). Educational administration: Theory, research, and prac- tice. New York: McGraw-Hill.

Hsiao, C. C. (2011). The relationship between schools' climate of creativity, teachers' intrinsic motivation, and teachers' creative teaching performance: A discussion of multilevel moderated mediation. Contemporary Educational Research Quarterly, 19(4), 85–125.

Hsu, S. W., & Chiu, Y. T. (2011). A study of organizational justice, employees' innovative behavior and job performance in the food and beverage department of international tourist hotels-organizational support as a moderator. Journal of Tourism and Leisure Studies, 17(1), 49–74.

Hsu, Y. L., Liang, C. Y., & Hsu, S. K. (2014). Learning model of intrinsic motivation and self- efficacy as a mediator to predict the scientific imagination. Chinese Journal of Science Education, 22(4), 389–412.

Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure anal- ysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55.

Hunter, S. T., Bedell, K. E., & Mumford, M. D. (2007). Climate for creativity: A quantitative review. Creativity Research Journal, 19(1), 69–90.

James, L. R., & Brett, J. M. (1984). Mediators, moderators, and tests for mediation. Journal of Applied Psychology, 69(2), 307–321.

Jaussi, K. S., Randel, A. E., & Dionne, S. D. (2007). I am, I think I can, and I do: The role of personal identity, self-efficacy, and cross-application of experiences in creativity at work. Creativity Research Journal, 19(2–3), 247–258.

Judd, C. M., & Kenny, D. A. (1981). Process analysis estimating mediation in treatment evaluations. Evaluation Review, 5(5), 602–619.

298 S.-H. Chang et al. / Learning and Individual Differences 51 (2016) 291–298

Karwowski, M. (2011). The creative mix? Teacher's creative leadership school creative cli- mate and students' creative self-efficacy. Chowanna, 36, 25–43.

Karwowski, M. (2014). Creative mindsets: Measurement, correlates, consequences. Psychology of Aesthetics, Creativity, and the Arts, 8(1), 62–70.

Karwowski, M. (2015). Peer effect on students' creative self-concept. The Journal of Creative Behavior, 49(3), 211–225.

Karwowski, M., Gralewski, J., & Szumski, G. (2015). Teachers' effect on students' creative self-beliefs is moderated by students' gender. Learning and Individual Differences, 44, 1–8.

Liang, C., & Chang, C. C. (2014). Predicting scientific imagination from the joint influences of intrinsic motivation, self-efficacy, agreeableness, and extraversion. Learning and Individual Differences, 31, 36–42.

Liang, J. -Y., & Wei, C. -H. (2013). The mediating role of self-efficacy in the organizational climate-employee creativity relationship. Journal of National Taipei College of Business, 24, 47–70.

Litwin, G. H., & Stringer, R. A. (1968). Motivation and organizational climate. Boston: Divi- sion of Research, Graduate School of Business Administration, Harvard.

Lubart, T. (2008). Connecting learning, individual differences and creativity. Learning and Individual Differences, 18(4), 361–362.

Muller, D., Judd, C. M., & Yzerbyt, V. Y. (2005). When moderation is mediated and medi- ation is moderated. Journal of Personality and Social Psychology, 89(6), 852–863.

Oldham, G. R., & Cummings, A. (1996). Employee creativity: Personal and contextual fac- tors at work. Academy of Management Journal, 39(3), 607–634.

Preacher, K. J., Rucker, D. D., & Hayes, A. F. (2007). Addressing moderated mediation hy- potheses: Theory, methods, and prescriptions. Multivariate Behavioral Research, 42(1), 185–227.

Runco, M. A. (2007). Creativity: Theories and themes: Research, development, and practice. New York, NY: Academic Press.

Schneider, B. (1990). Organizational climate and culture. San Francisco: Jossey-Bass. Shalley, C. E., Zhou, J., & Oldham, G. R. (2004). The effects of personal and contextual char-

acteristics on creativity: Where should we go from here? Journal of Management, 30(6), 933–958.

Tierney, P., & Farmer, S. M. (2002). Creative self-efficacy: Its potential antecedents and re- lationship to creative performance. Academy of Management Journal, 45(6), 1137–1148.

Tierney, P., & Farmer, S. M. (2004). The Pygmalion process and employee creativity. Journal of Management, 30(3), 413–432.

Tierney, P., Farmer, S. M., & Graen, G. B. (1999). An examination of leadership and em- ployee creativity: The relevance of traits and relationships. Personnel Psychology, 52(3), 591–620.