Journal Archive

The computational thinking, emphasized by Wing (2006), refers to the basic thinking skills based on the computer science principles that all people in the 21st century should have. There are various definitions for the components of the computational thinking. Wing (2006) distinguishes the computational thinking as the abstraction and automation abilities[9] and the Computer Science Teachers Association (CSTA) has subdivided it into the data collection/analysis/representation, problem decomposition, abstraction, algorithm, automation, simulation, and parallelization[10]. The Korean Ministry of Education replaced, in the definition of CSTA, data representation with structuring, abstraction with modeling, and included coding and simulation in automation. In order to solve the problems of teaching and learning activities caused by too granular components, KERIS proposed a DPAA(P) model consisting of five elements: decomposition(D), pattern recognition(P), abstraction(A), algorithm(A), and programming(P)[11].

In the age when all industries are converged with software and computers, the computational thinking is the key goal of the programming education for non-SW major students.

PBL, which started in medical education, is spreading to various fields as a learner-centered learning model in which learners solve the presented practical problems. Through PBL, students can acquire a variety of information to solve problems, internalize learning motivation, and gain effective problem-solving abilities, self-directed learning, and effective collaborative abilities.

Several researches on the effects of PBL have been conducted. The effects of PBL on the logical thinking and problem solving abilities of the under high school students have been studied[12,13] and the effects of applying PBL to various college subjects have been studied[14,15]. Recently, the studies on applying PBL to programming education have also been carried out[16,17].

Since the SW education is gradually spreading to non-SW majors, it is necessary to study the effects of PBL applied programming education on the academic achievement and the computational thinking abilities for non-SW major students.

The programming class in this study was a liberal arts class that was held during one semester and conducted for 2 hours in a week for a total of 15 weeks. Because the students in the class were unfamiliar with programming, we used the 'Partial-PBL' method, which conducted Scratch training in the first three weeks and then applied PBL to the class, rather than the 'Whole-PBL' method in which the entire semester was conducted as PBL.

Two types of problems were presented to students. Students resolved the problem of the type-1 from the 4^th week to the 7^th week in the 'Mixed-PBL' approach where the professor lectured basic programming techniques for one hour and the students performed PBL for the remaining one hour. The problems of the type-2 were resolved from 9^th week to 14^th week in the 'Pure-PBL' approach where the intervention of the professor was minimized and the students conducted PBL for 2 hours. <Table 1> shows the curriculum of the PBL-based programming class for 15 weeks.

The problems used in the PBL were classified into two types: the problem of the type-1 mainly requires the abilities for designing and implementing the algorithms to solve the problem, and the problem of the type-2 requires the activities for planing creatively SW contents and programming them in Scratch. More specifically said, the problem of the type-1 is to implement a program for analyzing the test result of a class. The program should have the functions of calculating the lowest, highest, median, and average values. To solve the problem, students were guided to investigate and understand the typical sorting algorithms of O(n2) - bubble, insert, and select sorts, and use one of them for problem solving. The problem of the type-2 is that students choose one of game or interactive story, and implement it in Scratch. As there were no tip about the contents of the game and story, students should start from the planning activities such as the market data and consumer’s trend data collection and analysis. <Table 2> summarizes the problems of each type used in the PBL.

For the experiment, we divided a programming course for non-SW major students of a local private university into two classes. One class (called ‘PBL Class’) was composed of 36 students and applied the PBL-based programming curriculum described in the previous chapter III and the other class (called ‘non-PBL Class’) was composed of 34 students and taught with traditional lectures using a Scratch textbook.

We got the students' consent in advance about the difference in the way of teaching depending on the classes and surveyed to see the experiences of the students in the programming languages and computational thinking. <Table 3> summarizes the results. No student had experienced the computational thinking and programming languages ​​other than Scratch, and there was one student trained with Scratch in each class.

This study measured the effects of PBL-based programming education on the academic achievement and computational thinking (see <Table 4>). The academic achievement was again subdivided into the test score evaluation and the learning effectiveness. The test score was measured by carrying out the mid-term and final exams using the same questions in PBL and non-PBL classes. In order to measure the learning effectiveness, a survey consisting of 10 questions was conducted on self-directed learning ability, continuous learning willpower, and mutual cooperation. In order to measure the effects on computational thinking, a survey consisting of 9 questions was performed on the cognitive level of computational thinking and the problem solving ability using computational thinking. The responses of all questionnaires consisted of 5 points of Likerttechnique, and the results of the survey were analyzed using the IBM SPSS Statistics program.

The problems of the mid-term and final exams were classified into the types of ‘concept and grammar’, ‘simple code writing’, and ‘algorithm and programming.’ <Table 5> shows the average scores of the exams of the PBL and non-PBL classes according to the type of problem. There was no significant difference between two classes regarding the 'concept and grammar' and 'simple code writing' types. However, for the 'algorithm and programming' type, the score of the PBL class was statistically significantly higher than that of the non-PBL class. It can be interpreted that the ability of programming algorithms is formed relatively high when performing self-directed learning with clear learning motivation.

<Table 6> shows the questionnaire items used to survey the learning effectiveness and the t-test results of the survey. The questionnaire items were designed based on NASEL, which is the criterion of teaching and learning competency of Korea Education Development Institute. According to the table, the PBL class showed statistically significant higher scores than the non-PBL class in all questionnaire items except for the first item asking the overall satisfaction level. In particular, self-directed learning, problem-solving ability, and motivation for learning have shown notable differences. This result is significant for students who have to live in the 4^th industrial revolution era, which requires voluntary and continuous learning of the software.

In order to examine the effectiveness on the computational thinking, we conducted a questionnaire survey to investigate the cognitive level of computational thinking and the problem-solving ability using computational thinking.

<Table 7> shows the questionnaire items used to survey the effectiveness on the computational thinking and the t-test results of the survey. In the case of the cognitive level of computational thinking, the PBL class showed statistically significantly higher scores in all items than the non-PBL class. In particular, the score of the PBL class for the item of asking if the computational thinking is necessary was significantly higher than that of the non-PBL class. This can be interpreted as a result of experiencing the effects of computational thinking on actual problem solving activities in PBL.

In the case of the ​​problem solving ability using the computational thinking, the scores of the PBL class was statistically significantly higher than those of non-PBL class in all items except one item. The difference between the two classes was small in the item of applying solutions to similar problems. It can be interpreted that the students in the PBL class did not experience similar problem solving before this study. It is noticeable that the score of the PBL class in the item of being confident in applying it to everyday life is significantly higher.