Poster #20 - Multilevel Models of Individual and Classroom Effects of a STEAM Program on Children’s Computational Thinking

• In Event: Poster Session 17

Abstract

Recent studies have applied integrated science, technology, engineering, art, and mathematics (STEAM) education to young children and have examined the program’s effects. As one of the effects of the STEAM program, computational thinking (CT) in early childhood refers to the ability to abstract, formulate, and solve problems (Bers, 2018). Even though STEAM education is implemented, it is not easy to measure the effect on young children. In addition, little is known about whether the effect is the result of the application of the program or whether individual or class-level characteristics affect the effect of the application of the classroom-based STEAM program. Therefore, the primary aim was to assess whether there is variability in the CT progress between children from the same classroom and between those from different classrooms after conducting a classroom-based STEAM program using robotic kits and whether individual and classroom characteristics explain the variability of CT change over time.
The research participants were 421 five- and six-year-old children (mean age = 73 months; boys n = 219, girls n = 231) and 31 teachers from 31 classes of 21 different kindergartens and childcare centers in the cities of Seoul and in Gyeonggi Province in South Korea. Participants engaged in a classroom-based STEAM education program using robotic kits connected with the Nuri curriculum for a 5-week period. Before and after conducting the STEAM program, children’s CT, vocabulary, numeracy skills, and executive function were tested. Teachers’ teaching efficacy, attitudes, and pedagogical content knowledge in math and science were also assessed. After the STEAM program was applied to classrooms, teachers were asked about the progress of the program, children’s interests, and degree of active participation. Teachers’ responses were used as classroom characteristics.
Descriptive statistics and bivariate correlations were computed using SPSS 27. Given the nesting of children from different classrooms in the data with repeated measures, multilevel models with 3 levels were estimated by using HLM 7.03.
The result of the null model shows that there is significant variance in children’s CT over time within both children and classrooms. Next, the mean of the CT change in classrooms is statistically significant with an estimated parameter, β100 = 8.637713. In addition, there are significant variances in CT scores between children in the same classroom and between those in different classrooms. The model with level-2 variables (EF, numeracy, and vocabulary) shows that children’s prior vocabulary and numeracy development levels were significantly related to their CT change. Finally, level-3 variables were entered into the model to specify a significant relationship between classrooms’ characteristics and children’s CT scores. Teachers’ knowledge about math and science and the progress of the STEAM program explained children’s CT scores.
The results from this study suggest that children’s prior vocabulary and numeracy development and classroom characteristics such as teachers’ knowledge about math and science affect the effects of classroom-based STEAM program application. Our results accentuate the importance of training to improve teachers’ math and science knowledge for young children and active attitudes toward STEAM program application.

Author

• Jihyun Sung, Sungkyunkwan University
  Presenting Author