創意傾向量表編製暨信效度考驗之研究：以花蓮縣國小高年級學童為例

Development and Validation of a Creativity Tendency Scale: A Study of Upper-Grade Elementary School Students in Hualien County

隨著21世紀進入創造力時代，創造力成為學童應對問題的關鍵能力，受到教育政策的高度關注，以往研究多聚焦於認知層面，本研究則探討情意層面的「創意傾向」，包括挑戰性、想像力、好奇心和冒險性，這些向度被視為影響創造行為的重要因素。本研究旨在編製花蓮縣國小高年級學童適用的創意傾向量表，並探討學童在創意傾向的性別與年級差異。研究分兩階段進行：第一階段進行專家審查，以281位學童進行預試以實施探索性因素分析；第二階段以944位學童為正式樣本，進行驗證性因素分析及內部一致性信度考驗，最後針對性別與年級背景變項進行差異分析以考驗量表之建構效度，並其在創意傾向的差異情形。研究結果顯示，創意傾向量表為二階四因素模式，總計20題，量表具有不錯的整體適配度、內在結構適配、聚斂效度、區別效度、建構效度及內部一致性信度。性別分析發現，男學童的挑戰性、好奇心與冒險性得分顯著高於女學童，而女學童的想像力則顯著高於男學童。年級分析顯示，六年級學童在四個向度的平均得分略高於五年級學童，但未達顯著差異，顯示創意傾向在高年級的發展相對穩定。本研究結果可作為教師設計差異化教學活動的依據，幫助學童均衡發展不同創意向度，以提升創造力培養的成效。

Creativity has gained increasing prominence as a critical educational objective worldwide in recent years. The OECD’s PISA 2022 assessment officially included creative thinking as a core domain, highlighting its importance for personal adaptability, civic participation, and innovation. In Taiwan, national initiatives such as the 2003 White Paper on Creativity Education and the 12-Year Basic Education Curriculum have likewise emphasized cultivating students’imaginative thinking and problem-solving abilities. Creativity tendency refers to the affective and motivational traits that underlie creative behavior, including an individual’s attitudes, interests, and emotional dispositions toward innovation. Williams (1969) proposed a widely adopted cognitive-affective interaction model, identifying four key dimensions of creativity tendency: Challenge, imagination, curiosity, and adventure. Creativity has been increasingly recognized as a core educational objective in global and national policy agendas in recent years. The OECD’s PISA 2022 assessment officially included creative thinking as a key test domain, defining it as the ability to generate meaningful and novel ideas. This inclusion underscores the importance of creativity for individual adaptability, civic engagement, and innovation. Taiwan has also prioritized creativity education through initiatives such as the 2003 White Paper on Curriculum Guidelines of 12-Year Basic Education: General Guidelines (2021), both of which emphasize the development of students’imaginative thinking and problem-solving skills. Creativity tendency is considered a critical affective trait that influences creative performance. It encompasses emotional and motivational factors such as interest, persistence, and willingness to explore. Williams proposed a cognitive-affective interaction model of creativity, identifying four core dimensions of creativity tendency: Challenge, imagination, curiosity, and adventure. This framework has become the theoretical basis for widely used measurement instruments. However, most existing tools, such as the Revised Williams Creativity Tendency Scale, were designed for older students or adapted from Western contexts, often without cultural localization or developmental adjustments. As a result, their validity for younger students in specific cultural regions remains limited. Hualien County, located in eastern Taiwan, features a unique combination of natural ecology, indigenous cultures, and a slower pace of urbanization. According to the ecological systems model of creativity development, creativity is shaped by interactions between individuals and their surrounding systems from family and school to social values and environmental context. Therefore, students in Hualien may express creative tendencies differently from those in metropolitan areas, highlighting the need for a culturally and developmentally appropriate assessment tool. A review of existing creativity tendency instruments shows that tools widely used in Taiwan, such as the Revised Williams Scale, may contain outdated language, abstract phrasing, or contextually irrelevant items for today’s students. Although several versions have been revised, item clarity, length, and response sensitivity issues persist. This study aims to develop and validate a Creativity Tendency Scale (CTS) for upper-grade elementary students in Hualien County, adapted to their linguistic, cognitive, and cultural contexts based on Williams’four-dimensional model. It also examines gender and grade-level differences to provide validity evidence and inform educational applications.
This study employed a two-phase process for developing and validating the Creativity Tendency Scale. In Phase I, a preliminary pool of 24 items was constructed based on Williams’four-dimensional framework, encompassing challenge, imagination, curiosity, and adventure. The items were formulated in language appropriate for upper-grade elementary school students and rated on a four-point Likert scale (1 = strongly disagree to 4 = strongly agree). To ensure content validity, the draft scale was reviewed by seven experts in creativity education, educational measurement, gender studies, and elementary pedagogy. After incorporating expert feedback, a pilot test was conducted with 281 fifth- and sixth-grade students in Hualien County. Exploratory Factor Analysis (EFA) was used to examine the preliminary factor structure. In Phase II, a revised scale was administered to a stratified random sample of 944 students drawn from 15 schools across both remote and non-remote areas of Hualien County. Confirmatory Factor Analysis (CFA) was conducted to evaluate the structural validity of the scale and to test whether a higher-order construct of creativity tendency could represent the four hypothesized first-order factors. Reliability was assessed using Cronbach’s alpha coefficients and composite reliability (CR), while construct validity was evaluated through average variance extracted (AVE) and discriminant validity analyses. These procedures served not only to verify the scale's internal structure but also to confirm its ability to distinguish conceptually related but empirically distinct subconstructs. Furthermore, a one-way Multivariate Analysis of Variance (MANOVA) was conducted to examine gender and grade-level differences in creativity tendencies. These demographic comparisons explored group differences and provided evidence for construct validity, demonstrating the instrument’s capacity to detect variance across theoretically relevant populations. Bonferroni adjustments were applied to control Type I error in post hoc comparisons.
In the pilot phase (N = 281), Exploratory Factor Analysis (EFA) supported a four-factor structure aligned with theoretical expectations: Challenge, imagination, curiosity, and adventure. The Kaiser-Meyer-Olkin value was .93, and Bartlett’s test of sphericity was significant (χ²= 3694.14, p < .001), confirming the suitability of the data for factor analysis. One item with a factor loading below .50 was removed, resulting in a 23-item scale. Factor loadings for the retained items ranged from .50 to .88, with the four factors explaining a cumulative 62.62% total variance. In the formal validation phase (N = 944), a first-order Confirmatory Factor Analysis (CFA) was conducted using the 23-item model. Initial fit indices suggested the need for model refinement. Three items (CQ4, CQ17, CQ23) were removed based on high modification indices and semantic overlap. The refined 20-item model was then subjected to second-order CFA. The results supported a four-factor model in which each dimension loaded significantly onto a higher-order latent construct, creativity tendency. Model fit indices were acceptable: GFI = .93, CFI = .94, TLI = .93, RMSEA = .06, and SRMR = .05. Standardized factor loadings of items onto their respective first-order constructs ranged from .58 to .77. The four first-order factors loaded strongly onto the higher-order construct, with coefficients ranging from .69 to .93. Composite reliability (CR) values for all subscales exceeded .83, and average variance extracted (AVE) values ranged from .49 to .53, indicating acceptable convergent validity. Discriminant validity was confirmed by comparing the square roots of AVE values to inter-construct correlations. Internal consistency was high: Cronbach’sαvalues ranged from .82 to .84 across subscales and reached .93 for the full scale. Multivariate analysis (MANOVA) revealed significant gender differences (Wilks’Λ= .96, p < .001,η²= .04). Boys scored higher in challenge, curiosity, and adventure, while girls scored higher in imagination. All univariate F-tests were significant (p < .013), though effect sizes were small (η²≈.007–.008). No significant grade-level differences were found, suggesting stability in creativity tendencies across fifth and sixth grades.
This study developed and validated a culturally and developmentally appropriate instrument to assess creativity tendency among upper-grade elementary students in Hualien County. Grounded in Williams’cognitive-affective framework, the Creativity Tendency Scale (CTS) demonstrated a clear four-factor structure, high internal consistency, and satisfactory convergent and discriminant validity. A second-order confirmatory factor analysis (CFA) further supported the conceptualization of the four sub-dimensions under a unified higher-order construct representing overall creativity tendency. Findings related to gender differences were consistent with previous studies, which have suggested that boys may exhibit greater adventurousness and curiosity, whereas girls may score higher on imagination-related traits. Although the observed effect sizes were relatively modest, the results nonetheless highlight the potential value of implementing gender-sensitive educational approaches to support balanced creative development. For example, instructional strategies might be considered to enhance risk-taking and exploratory behaviors among girls, while fostering imaginative expression among boys. The absence of significant grade-level differences may suggest that affective dimensions of creativity begin to stabilize during late childhood. This finding lends further support to the importance of introducing creativity-enhancing interventions at an early stage and maintaining consistent efforts across grade levels to cultivate students’creative dispositions. Importantly, this study underscores the relevance of culturally contextualized assessment tools. The CTS was carefully developed and validated to reflect the unique sociocultural and environmental characteristics of students in Hualien County. In doing so, it addresses limitations associated with applying generalized or externally developed instruments that may not fully capture region-specific expressions of creativity. As such, the CTS provides a meaningful foundation for local pedagogical practices and future academic inquiries into creativity education. It is hoped that subsequent research may build upon this work by applying the CTS in other regions to examine its cross-cultural applicability or by exploring longitudinal trajectories of creativity tendency. Adopting a mixed-methods approach that integrates quantitative data with qualitative insights—such as interviews or student reflections—may also offer a more nuanced understanding of how creative dispositions emerge and manifest in real-world contexts. In conclusion, the Creativity Tendency Scale represents a psychometrically sound, theoretically grounded, and contextual relevant instrument that can contribute to the assessment and understanding of creativity in education. It may assist educators and researchers in identifying students’creative strengths and developmental needs, thereby supporting more informed and responsive educational strategies, particularly in culturally diverse or underrepresented populations.