Impact of Coding on Preschoolers' Math Abilities

Framework

The theoretical premise of our research stems from cognitive studies of the human mind. We reached to Piaget’s foundational concept of children developing schema by negotiating cognitive conflicts during active and engaging learning experiences (Ginsburg et al., 1988). We integrated constructionism, Papert’s coinage for learning by doing, where children work as designers in a technological setting, deeply involved in projects that capture and grip their attention (Papert, 1980). We wove into this framework Minsky’s (1986) promotion of learning from failures and enjoying discomforts. These theories of the mind encapsulate our framework for this study on coding.

Early studies on coding included abstract computer programming using various forms of the language Logo. Even the more child-friendly versions of the 1990s required children to type in commands to execute action on a screen. Early 2000 saw the advent of lego or other object-based coding such as Mindstorms. Still these were primarily for elementary age children. Only in the recent past have we seen truly preschool-oriented robotic toys.

Much of the research on the benefits of coding and its impact on children’s development are from the 1990s and all of them are focused on elementary age children and on computer programming only. These studies have found positive effects on creativity, critical thinking, socio-cognitive interactions and mathematical thinking (Burton & Donovan, 1987; Clements, 1990; Nastasi, et al., 1990; Clements et al., 1993). Recent studies have examined the viability of introducing robotic toys in the curriculum and also explored effectiveness of the toys themselves(Sullivan et al., 2013; Hutchison et al., 2016). While these studies have described the promising benefits of coding, no study has directly examined the impact of specific robotic toys on preschooler’s math abilities.

It is our goal to address this gap with our study. Results of our study will support the burgeoning field of coding in early childhood education by identifying effective classroom usage and offering guidelines for refining the coding toys and the curriculum.

Methods

This study was conducted at an accredited campus-based preschool. The preschool follows a constructivist play-based curricular model. It has four classrooms, each led by a head teacher and a teacher associate, both of whom hold a master’s or a bachelor’s degree in early childhood education. For this pilot study, select group of 6 preschoolers were invited to participate. They were all 4 years of age, 3 girls and 3 boys, and from the same classroom. Selection of participants was based on availability and age. A sample size of six was determined to be adequate for this pilot study to validate the methodology.

The study was structured as a pre-and post-assessment with coding activities as intervention. Coding intervention consisted of 5 learning segments, each 20 to 25 minutes in duration. Pairs of preschoolers worked together during each learning segment. The learning segments were spaced out across five weeks. Beebot, a robotic bee-like toy, was used in this study. It has left, right, forward, backward, go and clear icons on its back. Each press of the first four icons either turned the beebot 90 degrees left or right, or moved it forward or backward 6 inches. Go executed the codes and clear erased the codes.

In the first learning segment of the coding intervention, preschoolers were introduced to the beebot and allowed free play time followed by guided demonstrations. In the second learning segment, they connected the icons to the beebot’s movements and created sequences of codes. Next, they purposefully coded it to move from a preset point A to point B on a gridded mat. Then they coded it to move around obstacles to get fromA to B. In the final segment, preschoolers were shown a path that they translated into to a sequence of codes. During each segment, the research assistant prompted the children to verbalize their coding with questions such as: How did you make the robot move there? What do you think went wrong? How can you fix it? Which step should you redo? Is there another way?

Data included pre-and post-assessments of children’s Math abilities usingTEAM, Test of Early Assessment of Math (Clements et al., 2016).Specifically, TEAM was used to capture children’s abilities to compare number quantities, exercise ordinality, finish or copy complex patterns, determine units of patterns, compare lengths, identify/build shapes, and identify turn directions.

Data also included children’s coding behaviors during the intervention, recorded as tallies using the OCB, Observation of Coding Behaviors instrument (Lisenbee, 2019). Behaviors that were observed included verbal communications (asking/answering questions, commenting, communicating results), verbal and non-verbal dispositions (motivation, confidence, enthusiasm, collaboration), and coding behaviors (sequencing, debugging, using patterns/loops, decomposition, creating a story, abstract thinking).

A trained research assistant tested each child individually on TEAM, conducted the interventions and recorded coding behaviors with the OCB.

Data analysis included analysis of variance to determine significant growth in Math abilities and a frequency analysis of coding behaviors by child and by session.

Results

This study addressed three important questions, as articulated earlier. First, we looked at the results related to the impact of coding on children’s math abilities. Pre-and post-TEAM scores were utilized to determine overall gains in math abilities. A total of 13 items (1 point for each correct answer) were assessed. At the pre-assessment, the average score across all six children was 7 and this increased statistically significantly to 9.56 during post-assessment. Analysis of the individual items indicated that the children showed gains in comparing quantities, comparing lengths, copying and finishing patterns, and building shapes.We measured their abilities to compare quantities on two items. In one, the children determined if 3 or 4 was bigger and in the other, they determined if 9 or 11 was bigger. While there was no difference in their scores in the first (smaller quantity) item, children showed significantly increased ability in the second (larger quantity) at the post-assessment. Similarly, at the post-assessment, children were able to compare 2 and 4 lengths better than in the pre-assessment. The three items with no gains or a decrease in scores were determining turns, identifying squares and identifying the unit in a pattern. This is based on overall average scores and not on individual scores (which we will discuss at the conference) where specific children showed remarkable growth. Based on these positive TEAM scores, we determined that coding had a positive impact on children’s math abilities.

Second, we looked at the coding behaviors of the children. We first looked at the children’s behaviors during each learning segment and then studied their overall behaviors. During the first learning segment, children exhibited enthusiasm, task persistence, problem-solving and creativity. Despite this being their first time, all children eagerly pressed the icons to observe the beebot move and stayed focused on the task for the full duration. During the second learning segment, children exhibited several coding behaviors with their growing understanding that each press of the button equated to one movement by the robot. It took the children at least another learning segment and for some children, all the way to the end of the study, to equate the turn icon to a turn only and not a turn and move forward. In learning segment 3, children exhibited more verbal communications, such as questions, comments or descriptions. They also engaged in count aloud behaviors, checking their one-to-one correspondence. During learning segment 4, several metacognitive behaviors surfaced with children thoughtfully sitting back, whispering ‘I don’t know how.’ This is the learning segment where an obstacle was introduced. Eventually, children could navigate around the obstacle if it was in front of the beebot, but not if it was further down the path. During the fifth learning segment, children had to break down a path to its code. Interestingly, the pairs (who had collaborated until then) worked individually to identify and debug codes. More details of the coding behaviors of individual children and theirMath scores will be shared at the conference.

Our third purpose was to determine the viability of the study and the instruments. TEAM was effective in measuring children’s math abilities, specifically the skills related to coding. OCB was effective in capturing children’s behaviors during the intervention. However, it was difficult for one person to observe and code. After the first learning segment, we moved to time-delayed coding, with two minutes of observation followed by one minute of coding. Even this proved to be difficult since the children continued to engage the research assistant in active conversation or problem-solving. The five learning segments that comprised the intervention were effective in both guiding the children to code on their own and to problem-solve.

Importance

The study carries several implications. First, coding does positively impact preschoolers’ math abilities. This was evident in the statistically significant growth we noticed even within our small sample size. Research has established this positive growth in upper grades, but that preschoolers also evince the same positive trend is an important consideration for math education and for coding curricula. Second, a few specific math abilities showed a decrease in scores at the post-assessment. Interestingly, these were advanced skills such as determining turns and identifying units in patterns. Prior research has indicated that as children engage in these complex skills, their preliminary awareness (or knowledge) appears to dip down but later rises up. This U-curve of learning is well documented in literature (Karmiloff-Smith, 1991) and it would be important to conduct follow-up studies to ascertain if indeed that is true here. If so, the implications for the long-term effects of coding would be phenomenal.

Other major implications relate to the coding behaviors. Children exhibited more positive dispositions (than either communication or coding behaviors) at the start of the study. Towards the middle, we noticed more communications with children asking questions, commenting and even talking aloud to themselves. As they got more involved in their coding, they evinced more coding behaviors, with some changes in their collaboration patterns. They also showed more awareness of what they did not know how to execute, an important sign of self-regulation. A future study would need to scrutinize these coding behaviors across a larger sample of preschoolers. It could be that coding elicits behaviors that are supportive of executive functioning.

Finally, this pilot study with its 5 learning segments of coding intervention, measured by OCB andTEAM proved to be a viable model for studying coding in preschools. However, we noted several changes fora future study. We would increase our sample size to include entire classrooms of preschoolers of varying age. By focusing on six 4-year old preschoolers from one classroom, we may have inadvertently included a stronger sample of children. We would also want to move to an experimental/control group model in order to tease out the impact of coding vs. natural growth. We would also video-record the intervention to document children’s coding behaviors with more accurate and detailed descriptions.

References

Bers, M., & Portsmore, M. (2005). Teaching Partnerships: Early Childhood and Engineering Students Teaching Math and Science Through Robotics. Journal ofScienceEducation andTechnology, 14(1), 59-73.

Burton, J., K., & Donovan, C., W. (1987). The effect of Logo programming instruction on first-grade student locus of control orientation. Computers in the Schools, 4(2), 66-70.

Clements, D. H. (1990). Metacomponential development in a Logo programming environment. Journal of Educational Psychology, 82, 141-149.

Clements, D. H., Sarama, J., & Wolfe, C. B. (2008/2016). REMA—Research-based Early Mathematics Assessment Denver, CO: Kennedy Institute, University of Denver.

Clements, D. H., & Meredith, J. S. (1993). Research on Logo: Effects and efficacy. Journal of Computing in Childhood Education, 4(4), 263-290.

Elkin, M., Sullivan, A., & Bers, M. U. (2016). Programming with the KIBO Robotics Kit in Preschool Classrooms. Computers in the Schools, 33(3), 169-186.

Ginsburg, H. P., & Opper, S. (1988). Piaget's theory of intellectual development. Englewood Cliffs, NJ: Prentice Hall.

Hutchison, A., Nadolny, L., & Estapa, A. (2016). Using Coding Apps to Support Literacy Instruction and Develop Coding Literacy. The Reading Teacher, 69(5), 493-503.

Karmiloff-Smith, A. (1991). What every cognitive psychologist should know about the mind of a child. In W. Kessen, A. Ortary, & F. Gaik (Eds.), Essays in honor of George Mandler (pp. 277-288). New Jersey: Erlbaum.

Kazakoff, E. R., & Bers, M. U. (2014). Put Your Robot in, Put Your Robot out: Sequencing through Programming Robots in Early Childhood. Journal of Educational Computing]Research, 50(4), 553-573.

Lisenbee, P. (2019) Computational thinking behaviors of K-2 students during coding. Unpublished manuscript.

Minsky, M. (1986). The society of mind. New York, NY: Simon and Schuster.

Nastasi, B. K., Clements, D. H., & Battista, M. T. (1990). Social-cognitive interactions, motivation, and cognitive growth in Logo programming and CAI problem-solving environments. Journal of Educational Psychology, 82, 150-158.

Papert, S. (1980). Mindstorms: Children, computers, and powerful ideas. New York: Basic Books.

Seidman, R. H. (1981, April). The effects of learning a computer programming language on the logical reasoning of school children.Paper presented at the American Educational Research Association, Los Angeles, CA.

Sullivan, A., Kazakoff, E. R., & Bers, M. U. (2013). The Wheels on the Bot Go Round and Round: Robotics Curriculum in Pre-Kindergarten. Journal of Information Technology Education: Innovations in Practice, 12, 203.