With science, technology, engineering, arts, and mathematics (STEAM) education in full swing worldwide, several science competitions based on STEAM have sprung up. One is Maker-Highlight-STEM&A, a competition sponsored by the Department of Education of the Taipei City Government, which involves STEAM learning and creativity through hands-on skills. In Maker- Highlight-STEM&A project making, knowledge sharing is crucial for seeking approaches to solve problems. Participants must know how to collaborate with each other and have a positive collective attitude to knowledge sharing to achieve the team’s goals. In particular, in the competition, the project topic is given on the spot, which encourages participants to share their knowledge and discover new ideas while creating project functions and improving their project quality, thereby helping them win the competition. Extending Kahneman’s (2003) dual process system, which states that automatic and reflective responses exist in knowledge sharing, we divided reflective responses into rational and emotional responses based on the cognition- affection model. To analyze the role of the three types of knowledge sharing, this study explored cooperative attitude as their antecedent and continuous intention as their predictor. In the highly competitive environment of Maker-Highlight-STEM&A, knowledge sharing in teamwork is crucial to win the competition; however, few studies have examined the correlation among cooperative attitude, knowledge sharing, and continuous intention to share knowledge. Because self-identity can foster attitudes toward a behavior in relation to the intention to perform that behavior, the present study considered the self-identity of participants in three types of knowledge sharing to predict their continuous intention to share knowledge in the competition.

Group interactions can affect people’s motivation to share their knowledge. Thus, this research built a research model extended from trait activation theory and social sharing motivation theory. With cooperative attitude as an independent variable, knowledge sharing as an intervening variable, and continuous sharing intention as a dependent variable, we proposed the following six research hypotheses:

H1: Cooperative attitude is positively associated with autonomous response.

H2: Cooperative attitude is positively associated with rational reflection.

H3: Cooperative attitude is positively associated with emotional reflection.

H4: Autonomous response is positively associated with continuous sharing intention.

H5: Rational reflection is positively associated with continuous intention to share knowledge.

H6: Emotional reflection is negatively associated with continuous sharing intention.

The purposive sampling technique was used in this research wherein all participants in the 2018 Maker-Highlight-STEM&A were invited to fill out a questionnaire survey. A total of 219 questionnaires were collected, of which 50 invalid questionnaires (with incomplete data) were excluded, leaving 169 effective samples for an effective rate of 77.2% and including 102 male (60.4%) and 67 female (39.6%) respondents.

By performing the first-order confirmatory factor analysis, questionnaire items with lower factor loadings were deleted. Then, the reliability and validity of the questionnaire constructs were tested to ensure they met the thresholds of statistical standards. Model fit and path analyses were performed to verify the research model.

When performing the first-order confirmatory factor analysis to evaluate the internal validity of items in each construct, items with residual values over 0.5 should be removed from the original questionnaire (Hair et al., 2018). In the original questionnaire, there were 11 items for cooperative attitude; five each for autonomous response, rational reflection, and emotional reflection; and five for the continuous sharing intention construct. Of the remaining items, those for cooperative attitude were reduced from 11 to six; those for autonomous response, rational reflection, and emotional reflection were reduced from five to four each; and those for continuous sharing intention were reduced from five to four. Exemplary items included “If my teammates have trouble shooting, I will help him or her right away” for measuring the cooperative attitude construct, with CR = .89 and Cronbach’s α = .90; “If any idea comes to my mind, I speak out right away” for measuring the automatic response construct, with CR = .84 and Cronbach’s α = .83; “If any idea comes to my mind, I determine whether it is rational before speaking out” for measuring the rational reflection response construct, with CR = .89 and Cronbach’s α = .89; “If any idea comes to my mind, I think about whether it will be rejected before I speak out” for measuring emotional reflection, with CR = .86 and Cronbach’s α = .86; and “I would like to continuously share knowledge in this type of contest in the future” for measuring continuous sharing intention, with CR = .94 and Cronbach’s α = .94. The CR and Cronbach’s α values were above the threshold suggested by Hair et al. (2018), indicating that the composite reliability and internal consistency of the questionnaire were acceptable.

In this research, having a cooperative attitude was the independent variable, the three different types of knowledge sharing were mediating variables, and continuous sharing intention was the dependent variable. Data were analyzed and tested using structural equation modeling with AMOS 20.0. The results of the data analysis revealed that cooperative attitude was positively associated with autonomous response, rational reflection, and emotional reflection. Moreover, autonomous response was negatively associated with continuous sharing intention; rational reflection was positively associated with continuous sharing intention; and emotional reflection was negatively associated with continuous sharing intention.

Throughout the competition, there were numerous benefits for participants apart from the purpose of the competition itself, including teamwork and knowledge sharing. Despite the fact that it is important to know how to effectively improve students’ knowledge sharing through such events, it has seldom been mentioned in research related to science competitions; hence, we conducted an empirical study targeting the 2018 science competition. The research results revealed that all three knowledge sharing types, namely autonomous response, rational reflection, and emotional reflection, played a positive predictive role in cooperative attitude. Emotional response and automatic reaction played a negative antecedent role in continuous intention to share knowledge. In line with the Chinese saying “thinking three times before taking action,” we should focus on knowledge sharing from rational reflection and deactivate knowledge sharing from emotional reflection and automatic response to improve knowledge transmission effects for participants to experience a more meaningful STEAM process.

Remarkably little research has been conducted on learners’ knowledge sharing behavior in school and even less on such behavior in competitions. This research revealed that a competition involving teamwork, the behavioral model of knowledge sharing, and the intention of continuous sharing can be expanded to understand cooperative learning in any other STEAM- related contests. In addition, this study proposed three different types of knowledge sharing (i.e., autonomous response, rational reflection, and emotional reflection) extended from Kahneman’s dual-process theory (intuitive feedback and reflective feedback) to review continuous sharing intention. These three knowledge sharing types could be a reference for future research teams to improve knowledge-sharing effects in other cooperative settings.

Onsite team competitions promote cooperative opportunities for participants to develop their attitude to collaborative work and knowledge sharing in rational reflection. Cooperative attitude development is essential for meeting the requirement of new 108 curricula. In line with this, other types of science or STEAM contests can integrate onsite activities to ensure that participants must work collaboratively to win the contest. In particular, National Science Fair or Science Olympic contests can be implemented with trouble shooting onsite, in which effective knowledge sharing can be embedded along with project making.