sofia SJOPOA. Journal of Learning and Teaching in Digital Age, 2023, 8(1), 113-123 ceuonaton bs enginar ong. twen/pavjlida a ISSN: 2458-8350 (online) = Research Paper Developing Computational Thinking Scale for Primary School Students and Examining Students' Thinking Levels According to Different Variables Ergiin Yuldinm”, Gelebi Uluyot” NORCID ID; doe. 4001-9774-0547), Cant Univers, a = *Conesponding author ARTICLE INFO ABSTRACT ‘Receed: [6 Sepiewbe 027 Tn recent years, computalional tiiking has been considered as one of the 21st century skills that all Revised 17 Oster 2022 students should have. Rescarchers emphasize the importance of determining and developing students! Accepted 19 Oster 2022 computational thinking levels from the earliest possible age. However, no measurement toal has been Found in the litertut tat sms to reveal the computational thinking levels of primary school stadent In this study, it was aimed to develop the computational thinking scale for primary school students Keywords: Eermord inking and to examine the computational thinking levels of primary school stents according to different Savrabvelopnene = variables (grade level daily computer se time). Inthe first stage ofthe study a scale with appropiate Priary schoo! ‘psychometric properties was developed to measure computational thinking. Inthe scale development Grade level phase ofthe research, exploratory sequential mixed methods research design was used Inthe other Dally computer wage ine phase ofthe sty, it was investigated whether the computational thisking levels of primary school Stadents differed according to the grade evel and daly computer usage ime without any intervention, For this reason, the research was caried out in aovordance with the general survey model, which is cone ofthe descriptive research types. For the fist stage, the study group ofthe research consisted of 287 stents studying inthe Ist, 2nd, 3d and dt grads of primary schools in Ankara Golbasi district in the second term of the 2021-2022 academic year. In the process of examining the students computational thinking levels according othe variables of grade level nd daily computer usage time, the study group consisted of a total of 96 students attending the primary education classes of a private School in Ankara in the second term of the 2021-2022 academic year. In this context, the one- dimensional computational thinking scale consisting of 17 items was applied to 287 primary school Students and the obtained data were subjested to validity and reliability analysis. According to the explanatory factor analysis, the seale explains 46% of the total variance, When the reslls of the explanatory factor analysis ee examined, it is sen thatthe fctr loads of 17 items in the sale vary between 56 and 86, The internal consstncy coeficient ofthe scale was found to be Cronbach Alpha 92. The developed scale was applied to primary school students inthe next stage. Asa result ofthe ‘study, it as found shat the generally computational thinking level of primary schoo! students differed significantly according to the grade level, On the other hand, it was observed that the students tational thinking levels difered significantly according to the time spent in Groat of the computer daly, and the mean ofthe students’ computational thinking seale increased asthe daily computer use time increased 8 SEES INTRODUCTION Today, the needs of societies, science and technology are changing rapidly. This situation undoubtedly changes the characteristics ‘expected from individuals, Learning and teaching approaches, theories and strategies also differ in order to raise individuals with desired and needed characteristics (MEB, 2018). Individuals need new skills different from traditional methods at the poiat of accessing and using information, revealing new information and sharing the information produced (Polat, 2006), For this purpose, the “21 Century Learning Framework” has been determined by the “21* Century Skills” joint working group, which includes high level companies and associations such as Microsoft, Lego, American Association of School Librarians, Pearson, National Education Association, Intel, Dell, Apple (DODEA, 2014). In ine with this determined framework, the 21° century skills that individuals are expected to have been put forward in detail (P21, 2009). These skills are defined as three main themes: leaming and innovation sails, life and career skis, knowledge, media and technology skills, and different skill groups under each theme (Kalemkus & Bulut Ozek, 2021). At this point, computational thinking emerges as one of the 21® century skills that every student should have (Grover & Pea, 2013; Gulbabar, Ker, & Kaleliogls, 2019), ‘Computational thinking is defined as meatal processes used in areas such as understanding human bchavior, designing systems, solving problems efficiently and effectively, being aware of information processing capacity, and designing aulonomous processes (Wing, 2006). According to Barr and Stephenson (2011) computational thinking is explained as the reconstruction of data through abstractions such as models and simulations, use of the data and producing appropriate solutions to existing problems by considering the limits of computing. Computational thinking is also expressed as skills such as processing, building and ‘transforming 1 1 2023, Jounal of Learning and Teaching in Digital Age, (1), 113-128E Yildirim, & ¢. Uluyol information, technology literacy and ereat oducts a5 result of all these provesses (Sendurur, 2018; Yoleu, 2018). Computational thinking is seen as a prerequisite for future professions and when evaluated in terms of 21* century competencies, itis considered as one of the basic skis such as reading, writing and ealculation (Ambrosio, Almeida, Macedo, & Franco, 2014). Computational thinking develops skills such as problem solving, critical and logical thinking and creativity in individuals. Iralso prepares them forthe global race and blends success in school life with suceess in real lfe (ISTE, 2011; Korkmaz, ‘Orden, Oluk, & Sariogla, 2015). For this reason, in today’s information and technology age, computational thinking is considered as attitudes and skill that can benefit not only computer engineers, but also all individuals from every profession, and it is ‘emphasized that these skills should be gained to every student as early as possible (Wing, 2006). As a result, computational thinking defines skills such as processing, constructing and transforming information, technology literacy and creating ereative and innovative products asa result of all these processes (Sendurur, 2018; Yolcu, 2018). Computational thinking is seen asa prerequisite for future professions and when evaluated in terms of 21st century competencies, itis considered us one of the base skills such as reading, writing and calculation (Ambrosio, Almeida, Macedo, & Franco, 2014). For this reason, developed countries carry out studies to inchide more computational thinking in their K12 curricula and develop programs in which computational thinking takes a large place Research context Education and training programs in Turkey are updated in accordance with the needs of the age. The special objectives of the Information Technologies and Software course curriculum updated in 2018 include the development of students’ problem-solving and computational thinking skills (MEB, 2018). This situation clearly reveals the necessity of increasing the importance given to ‘computational thinking, which is seen as a 21* century skill in educational environments, Because the computational thinking ‘process includes many concepts and processes such as critical and creative thinking, abstraction, algorithm design, automation, daca collection, data analysis, data presentation, parsing, pattem recognition, patter generulization, collaboration and modeling (Gulbahar, Kert, & Kalelioglu, 2019; ISTE, 2011. In this respect, itis tought tha itis important to actively integrate computational thinking into education systems. However, researchers emphasize that there are not enough resources and activities to cary ‘computational thinking into the classroom environment (Barr & Stephenson, 2011). Ip addition, i is another problem that studies ‘in the relevant literature tend to develop this skill hefore determining the computational thinking levels of individuals (Rerikan, 2018; Gulbahar, Kert, & Kalelioglu, 2019). However, it is very important to determine the level of computational thinking that already exists in individuals, as well as to acquire computational thinking skills, or to evaluate to what extent individuals have acquired computational thinking skills as a result of the designed applications. Therefor, the process of determining the level of computational thinking in individuals, whose scape is extremely wide, should also be structured comprehensively, and ‘computational thinking competencies should be carefully evaluated by considering knowledge, aitudes and skils (Snow, Katz, Elliost Tew, & Feldman, 2012). Student development files, multiple choice test, computational thinking pattern graph, project and performance evaluation, computational thinking seale and rubrics can be used to evaluate computational thinking (Gouws, Bradshaw, & Wentworth, 2013; Grover, Pea, & Cooper, 2015), When the literature is examined, it i seen that there are limited studies on the measurement and evaluation of students’ computational thinking level. Again, i ie noteworthy that a very Timited ‘number of measurement and evaluation tools have been developed for the relevant subject in our country. When the measurement tools developed in our country are examined; here is a computational thinking skills sete for university students (Korkmaz, Cakit, & Caden, 2017) and a computational thinking skills test for secondary school students (Ozmen, 2016). I is also seen that the ‘computational thinking skill measurement tool (Yildiz, 2021) and the self-efficacy perception scale (Gulbaher, Kert, & Kalelioglu, 2019) for computational thinking skills have been developed for secondary schools. tis emphasized by researchers that there are ‘not enough measurement tols to measure the level of computational thinking, especially in our country (Demir & Seferoghu, 2017) In addition, when the literature was reviewed, no measurement tool aimed at revealing the computational thinking levels of primary school students was found. Again, no other study has been found in the literature examining the computational thinking levels of primary schoo! students according lo different variables. AL this point, this study aims to provide a valid and reliable computational thinking scale for primary school students to the relevant literature. Another aim of the study is to examine the computational thinking levels of primary school students, who are the sample group in which there are limited studies on computational thinking levels, according o various variables (grade level, daily computer use time) within the scope ofthis researc. ‘Therefore, there are two main problem statements in this research. The first problem statement of the study was “What are the psychometric properties of the Computational Thinking Scale (CAN) that will be developed for primary school students?” has been determined, The other problem statement of the research is “Do primary school students’ computational thinking levels vary according to different variables (grade level, daily computer use time)?" is in the fort, METHOD Research model ‘This study, which aimed to develop the computational thinking seale and fo examine the computational thinking levels of primary schoo! students according to different variables, was carried out in two stages. In the first stage ofthe study, a scale with appropriate psychometric properties was developed to measure computational thinking. In the scale development phase of the research, exploratory sequential mixed methods research design was used. Inthe other phase of the study, it was investigated whether the ‘computational thinking levels af primary school students differed according tothe grade level and daily computer usage time without 4 © 2023, Jounal of Learning and Teaching in Digital Age, 8(1), 113-123,Developing Computational Thinking Scale for Primary School Students ‘ny intervention. For this reason, the research was carried out in accordance with the general survey model, which is one of the descriptive research types, Descriptive research is defined as research in which the current situation ofthe subject is examined and the relationship between the variables is revealed without changing (Buyukozturk, Akgun, Karadeniz, Demirel, & Kilic, 2016; Yildirim & Simsek, 2018). Survey models are studies in which research data are collected from a sample group representing the universe, rather than the entire universe, in order to reveal the views or characteristics adopted by large populations (Fraenkel & Wallen, 2009). The general screening model is defined as a model in which the researcher has no effec on the independent variable and includes studies conducted inthe entire universe or in a smaller group to be taken fom the universe in order to reach a general {judgment about the universe (Karasar, 2017), Working group ‘The study group of this research was determined by convenient sampling (Cohen, Manion & Morrison, 2007). In convenient sampling, the researcher or researchers conducting the study select the participants who will form the study group from volunteer individuals who are suitable forthe research and easy to reach (Gravetter & Forzano, 2012), During the seale development process, Which isthe fist stage of the reseatch, the study group of the research consisted of 287 students studying in the 1*, 2% 3 and 4 ‘grades of primary schools in Golbasi district of Ankara inthe sccond term of the 2021-2022 academic year. It is stated by Bryman and Cramer (2001) thatthe sample size reached in the seale development studies should be atleast five times the number of items in the scale. Since the draft scale consists of 26 items, itis thought thatthe sample size of 287 students is sufficient In the process of examining the students’ computational thinking levels according o the variables of grade level and daily computer "usage time, the study group consisted ofa total of 96 students attending the primary education classes ofa private school in Ankara in the second term ofthe 2021-2022 academic year. Of the 96 students inthe study group, 24 are I" grade, 24 2 grade, 24 3° grade and 24 4® grade. The frequency distribution ofthe students inthe study group according tothe grade level and daily computer usage time is given in the table below (Table 1). ‘Table 1. Frequency distribution of the study group according tothe variables of grade level and daily computer usage time Variable £ % Grads level Tgrade 2 25 2 grade 4 25 Seigrade 24 25 grade 24 25 Tess than 2 Fours 2 BH Daily computer usage time 2-4 hours 34 35 46 hours 21 2 More than 6 hours 19 20 Data collection tool ‘The data collection tool used in the research isthe Computational Thinking Scale developed by the researchers. The development process ofthe scale and its psychometric properties are given below in detail Before starting the research, it was aimed to investigate whether the computational thinking levels of primary school students change according to different variables. When the relevant literature is reviewed, there are measurement tools that aim to determine the computational thinking levels of university (Korkmaz, Cakir, & Ozden, 2017) and secondary school students (Gulbabar, Kert, & Kalelioglu, 2019; Ozmen, 2016; Yildiz, 2021) in our country. However, it has been observed that there is no measurement 1001 to determine the computational thinking levels of primary school students. Therefore, with this study, it was decided to develop a seale ‘that ean reveal the computational thinking levels of primary school students, In the first stage of the scele development process, information about computational thinking and the sub-dimensions of ‘computational thinking was obtained by examining various sources (Barr & Stephenson, 2011; Grover & Pea, 2013; Kaleliogl, Gulbahar, & Kukul, 2016; Selby & Woollard, 2013; Wing , 2006; 2008; 2011). Based on the relevant literature and especially considering the classification of Wing (2008; 2011), who has many studies on computational thinking, the sub-dimensions of ‘computational thinking were determined as abstraction, algorithm, automation, decomposition, generalization/evaluation. Then, an item pool of 33 items was created to include all these suib-dimensions and ensure content validity. While preparing the items in the item pool, attention was paid to write the items in aecordance withthe age group. In addition, all the items are written in a way that includes a single situation and is also clear and understandable, The scale was prepared in 3-point Likert type as “agree”, "no idea” and “disagree”. The scale was prepared in a triple Likert type because itis stated that scales with fewer choices are suitable for less ‘educated or younger respondents (Koklu, 1995). Empirical studies reveal that participants do not perceive the difference between ‘options equally in Likert-type questions (Hart, 1996), This situation causes a change in the number of participation levels and therefore affects the validity and reliability ofthe scale (Ozkan & Bindak, 2021). Expert opinions were sought to ensure the content validity of the 33-item, point Likert-type scale, The prepared seale was ‘examined in terms of content validity by two facully members working in the departments of computer and instructional technologies education and one faculty members working in the division of classroom instruction education. In addition, the seale ‘was examined by 2 classroom teachers in terms of suitability for student level, language and intelligibility, and by an assessment nis 1© 2023, Journal of Learning and Teaching in Digital Age, S(1), 113-123,E Yildirim, & ¢. Uluyol and evaluation expert in terms of compliance with assessment and evaluation criteria. Inline with the suggestions from the experts, it was decided to remove some items from the seale because they were not suitable for the assessment and evaluation criteria and the level of primary school students. Again, the scope of some items in the scale was narrowed down and the items were corrected tobe simpler and more understandable. After these regulations, a draft scale consisting of 26 items was crested, Sine the age group of the participants is small and especially the students who are just learning to read, such as the frst graders, are also included in the participant group, the draft scale with 26 items was thought to be appropriate In the pilot study of the research, the draft seale was applied to a total of 20 primary school students. As a result ofthe feedback ‘obtained by the students while answering the scale, the items that were difficult to understand were revised again, As a result ofthis preliminary application, the application time ofthe seale was determined as 25 minutes. After this stage, exploratory factor analysis ‘was started ‘The draft scale, which was prepared in line with the opinions obtained from the experts and the preliminary application made to primary school students, was applied to a total of 287 students studying in primary schools in Ankara province Golbasi district uring this practice, including 135 female and 152 male students; 69 students in the 1% grade, 71 students in the 2°! grade, 73 students in the 3° grade and 74 students in the 4® grade were reached. The data obtained from the application were anelyzed inthe SPSS 21 package program and the findings related to the scale were reached. {As aresul ofthe analysis ofthe dat, the Kaiser-Meyer-Olkin (KMO) value ofthe éraft scale consisting of 26 items was calculated as .8. In addition, a significant difference was found in the Bartlet Sphericity test result ofthe scale (p < 05). As a result ofthe analysis, the significant difference in the Barilet Sphericity test and the KMO value greater than 70 revealed that the data were suitable Tor facior analysis (Leech, Barrett, & Morgan, 2005; Tavsaneil, 2010), ‘After the scele was found to be suitable for factor analysis, factor analysis was applied tothe scale and after the analysis ofthe main ‘components of the scale, 7 factors with eigenvalues greater than 1.00 emerged. However, the fact thet the eigenvalue ofthe frst factor was significantly higher than the eigenvalues ofthe other factors indicated that the scale had a single-factor structure, At this point, in order to determine the factor number ofthe scale, Catel’s"screo" test was performed (Kline, 1994) and a graph as below ‘was obtained (Figure 1). sere Piet Eigenvaue ‘component Number Figure 1, Computational Thinking Scale line chart The sharp decline points in the graph determine the factor number ofthe scale (Singh, 2007). When the Seree plot graph in Figure | is examined, itis seen tha there is only one sharp decline point, This showed that the scale measures a single factor structure and it was decided to have a single factor scale. The results obtained regarding the variance value ofthe single factor scale are given in the table below (Table 2), ‘Table 2. Findings related toa factor as a result of factor analysis Factor Eigenvalue. ‘Percentage of Varianse Percentage of Total Variance 1 5.14 46.49 46.49 When Table 2 is examined, the eigenvalue ofthe single ctor in the sale was found to be 8.14, and the total variance percentage of this factor was found to be 46.49. The single facor inthe scale explains 46% of the total variance. According to Kline (1994), the acceptable rate is 41%, The fact that the value obtained from the relevant scale is above 41% allows the scale to be used as & scale consisting ofa single factor. In scale development studies by researchers, itis considered sufficient thatthe variance explained, 16 2023, Jowrnal of Learning and Teaching in Digital Age, 8(1), 113-123,Developing Computational Thinking Scale for Primary School Students especially in social sciences, is between 40% and 60%, The height ofthe explained variance is accepted as an indicator of how well the relevant structure is measured (Buyukozturk, 2012; Scherer, Wiebe, Luther, & Adams, 1988) Factor analysis was also used to examine the factor structures measured by the scale items. At this stage, its stated that the factor load values ofthe items in the see should be above .30 and the difference between two high factor loads should be a lest 10 in ‘order to ensure the construct validity of the scale (Bryman & Cramer, 2001; Buyukozturk, 2012), At this point, experiments were ‘carried out by removing scale items from the scale in order to ensure thatthe factor loads were in accordance with the referenced values, lems with factor Toad less than 30 and under more than one factor atthe same time were excluded from the seale (Md, ‘MS, M6, M8, M9, MIO, M12, M17, M19), As a result of the analyses, a total of 17 items remained in the scale, As a result of removing the relevant items ffom the seale, the factor loads of the items in the scale were found to be between .56 and .86. The factor load values of the remaining 17 items inthe scale are given in Table 3 ‘Table 3. Factor loading values of the items in the scale Tem no Factor load M3) 86 M26 83 M23 St MIs 9 M2 9 M24 n Mos 69 MIL 65 M7 65 Mis 6 M3 6 M20 59 M21 59 MIs 38 M22 38 MI 57 MIs 56 In the next step, to determine the reliability ofthe scale, the internal consistency coefficient Cronbach alpha was calculated and found to be 92. A Cronbach alpha coefficient greater than 70 i considered sufficient, and greater than .90 is indicated as excellent. ‘The seale was found to be highly reliable (Kilic, 2016.) After the validity and reliability analyzes, a 17-item scale was reached. The distribution ofthe items in the seale according to the computational thinking sub-dimensions is given in Table 4 Table 4. Distribution of items according to computational thinking sub-dimensions ‘Sub-dimension Sub-dimonsion ‘Abstraction Algorithm ‘Automation Decomposition Generalization! Evaluation 6212 ‘The lowest score chat can be obtained ftom the scale is 0 and the highest score is SI. In the seale, reverse items were determined as (M2, MS, MB, M10, M12 and MI4 items. The response time ofthe scale is approximately 15 minutes. Data Collection Process ‘Within the scope of this research, a valid and reliable computational thinking scale consisting of 17 items and one dimension was developed. After the development ofthe scale, an answer was sought for another problem statement of the esearch. At this stage the scale was applied to primary school learners and it was examined whether the computational thinking levels of primary school students changed according tothe grade level and daily computer usage time. At this stage, the scale was applied toa total of 96 primary school students attending the primary school classes of a private school in order to iavestigate whether the computational ‘thinking levels of primary school students differ according to the grade level and daily computer usage time variables. Data Analysis “Microsoft Excel spreadsheet program and SPSS 21 statistical analysis program were used inthe analysis of the data ofthe research ‘The Shapiro Wilk test was used to determine whether students responses tothe seale were normally distributed. One-Way ANOVA, ‘one ofthe parametric analysis techniques, was used to determine whether there was a statistically significant difference between the uw 1 2028, Jounal of Learning and Teaching in Digital Age, 8(1, 113-123E Yildirim, & ¢. Uluyol ‘mean scotes ofthe students in different groups. In case of significant difference as a result of the analysis, Bonferroni test, one of ‘the multiple comparison tests, was used to determine between which groups the difference was, The significance level was accepted as 05 in all analyzes, FINDINGS AND INTERPRETATION Before proceeding to the analysis of the data, the statistical method to analyze the data collected through the Computational Thinking Scale was examined. In order forthe data collected during the study to be analyzed with parametric tests, all of them should show ‘anormal distribution. For this reason, frst of all, the data should be analyzed by choosing the appropriate normal distribution test, and it should be decided whether the normality assumption is provided and parametric or non-parametric tests should be used inthe analysis ofthe data (Ginasemi & Zahediasl, 2012; Sim & Wright, 2002). The Shapiro Wilk test was applied to all data obtained from the Computational Thinking Scale, which was also used as a data collection too inthis study, The Shapiro Wilk test is generally preferred in cases where the number of participants is less than $0 (n<50) (Ghasemi & Zahediasl, 2012). The Shapiro Wilk test results of the study group's Computational Thinking Scale scores according to grade level, daily computer use time and gender Variables are given in Table S. ‘Table $. Shapiro Wilk Test results regarding grade level and daily computer use time variables Variable Group, Satistios a T grade 7 cry Grade level 2 prade 4 24 3 grade 37 24 4 grade 35 24 Daily computer Tess than 2 hours 5 2 usage time 2-4 hours oT 34 4-6 hours 98 21 More than 6 hours 26 19 ‘When the Shapiro Wilk test resus in Table I are examined, itis seen that the significance level of the Computational Thinking Seale scores is greater than p > 0S according to the variables of class level and daily computer usage time of the study group. The values obtained show thatthe scale scores show a normal distribution according tothe variables of grade level and daily computer use time, Parametric tests were used inthe analysis ofthe data obtained from the Computing Thinking Seale as a result of the normal distribution of the data obtained from the scale in terms of all variables. After the data showed normal distribution, One-Way ANOVA was applied to the data in order to determine whether there was a statistically significant difference between the scale ‘mean scores of the students according tothe grade level, The obtained result are given in Table 6 ‘Table 6 ANOVA results of Computing Thinking Scale scores according to grade level variable ‘Variable ‘Sum of Sq rT Mean Sq F 2 Grade level Between groups 3860,12 3 1286,71 19590 _,00* Within groups 601,22 92 634 Total 4461.33 95 paus ‘When Table 6 showing the results of the ANOVA test according to the grade level variable of the seale scores is examined, itis sven that thre is statistically significant difference between the grade levels ofthe students and their computational thinking scores [Fo,s7= 19.90, p =.05] . Post-hoc test was needed in order to find out from which grade level or levels this statistical difference ‘between grade levels originates. At this point, the Bonferroni Test was preferred because itis a multiple comparison test that docs rot require the principle of equal sample size (Miller, 1969), The Bonferroni Test results regarding the multiple comparison of scores are given in Table 7. ‘Table 7. Multiple comparison test results according to grade level variable (Degroup group Mean difference Stderor —p ‘95% Coniidence interval oy Lower bound Upper bound gn 2 grade 526 2 ‘008 “72 “331 38 grade 642 5 ‘oot 845 “438 4% grade 17.42 74 19.4 1543, Tgrade 3° 5,26 2 721 grade “HS 77 Ba 4M prude 12,15 in 1020 Te grade 642, 5 85 2% grade 1.15 4 315 4% grade -11,00 5 “897 us (© 2023, Jounal of Learning and Teaching in Digital Age, (1), 113-123Developing Computational Thinking Scale for Primary School Students Pgade 1" grade 17a Te oF Ta TAT 2% grade 12,15 2 ‘oot 1020 van 3% grade 11,00 5 ‘oor 897 13,03 pa0s ‘The results of the Bonferroni Test, which was used to investigate the source of the significant difference between the grade levels ‘of the students and the Computational Thinking Scale point averages, ae given in Table 7. When the table is examined, iis seen that there is significant difference between the mean scores ofthe 1* grades and the mean scores of the 2, 3% and 4® grades in favor of the 2, 3" and 4" grades (p <.05), There was no significant difference between the mean scores of the 2* graders and the ‘mean scores of the 3% graders (p ».05). Again, looking a the data in the table, e significant difference was found between all grade levels and 4" graders in favor of 4" graders (p <.05). In the study, it was also aimed to examine whether the Computational Thinking Seale scores ofthe students differ according tothe daily computer use time, One-Way ANOVA was applied to the data in order to determine whether there isa statistically significant difference between the scale mean scores according to the daily computer usage time variable of the students. The obtained results are given in Table 8. ‘Table 8. ANOVA results of Computing Thinking Scale scores according to daily computer use time variable Varial ‘Sum of Sq af Mean Sa F 2 Daily computer usage Between groups 1353.19 3 451,07 eae ——«0 time Within groups 608,14 92 661 Total 1961.3 95 pas ‘The ANOVA test results according to the daily computer use time variable of the Computing Thinking Seale scores are given in Table 8. A statistically significant difference was found betw in front of the computer during the day and the ‘computational thinking scores of the students in the study group [Fn 68.24, p < 05]. The Bonferroni Test was applied to the data in order to determine between which hour intervals the significant difference found in the daily computer usage time of the students was. The results of the multiple comparison ofthe scores are given in Table 9, ‘Table 9. Multiple comparison test results according to daily computer usage time variable @Deroup—@) group ‘Mean Std P 195% Conlidence mierval difference error Lower bound Upper bound as) Las than 2 2-Shows 47 70 “00° “61 “EST hours 4.6 hours 1,66 78 :00* 9.77 558 More than 6 hous 10,97 ‘31 ‘00* “134 8.80 a saat Tess than 2 hours 4,71 10 00" 281 661 4-6 hows 2.95 7 :00* “487 1,02 More than 6 hows 6.26 4 00" “427 Fi saat Tess than 2 hours 7,66 8 00" 37 2-4 hours 295 7 ‘0* 437 More than 6hours 3,3 1 100* 12 ethan 6 Less than 2 hours 10,97 ST “008 13,14 2-4 hours 626 4 ‘o0* 825 4-6 hows 331 81 00" S31 ‘When the data in Table 9 were examined, i was observed that there were significant differences in daily computer usage times ‘between all ime zones (p<.05). Computational thinking mean scores of students who spend les than 2 hots a day tthe computer and 2-4 hours, 46 hous and more than 6 hou infront ofthe computer show a significant difference in favor of increasing the ime Spent in front of the computer (p<.05), Again, there i a signfiant difference between the mean scores ofthe students who spend 2-4 hours in font ofthe computer and those who spend less than 2 hours infront ofthe compute in favor ofthe students who spend 2-4 hours in font of the computer (p <5). The mean scores of students who spend 2-4 hours onthe computer and those who spe 46 hours or more than 6 hours a dey difer in favor of students who spend 2-4 hours (p-<.05). Similarly, the average score of the students who spent 46 hours onthe computer and the averages ofthe students who sper more than 6 hour in rot of the computer differ in favor ofthe students who spent more than hours (p< 05). Finally, there isa significant difference between the mean Scores ofthe studens who spend more than 6 hous a day atthe computer and the mea: scores of the students who spend less than 2hours, 24 hours and 4-6 hours in favor of the students who spend more than 6 hours on the computer (p =.) DISCUSSION AND CONCLUSION ‘With this researc, it was primarily aimed to develop a valid and reliable Computational Thinking Seale for primary school students. ‘When the measurement tools developed in our country on the subject are examined: itis seen that there is @ computational thinking 9 1 2028, Jounal of Learning and Teaching n Digital Age, 8(1), 13-123E Yildirim, & C. Uluyol skills scale for university students (Korkmaz, Cakit, & Ozden, 2017), In addition, there are computational thinking skills test (Ozmen, 2016), computational thinking skills measurement tool (Yildiz, 2021) and self-efficacy perception scale for computational thinking skills (Gulbabar, Ker, & Kalelioglu, 2019) for secondary school students in the literature. However, as it was emphasized ‘before, no measurement foo! aimed at revealing the computational thinking levels of primary school students was found. However, it's stare by the researchers that there are problems in the acquisition of computational thinking skills by the students since there ‘sno measurement tool to evaluate computational thinking skis, Therefore tis emphasized that evaluation tools that ean meastre ‘computational thinking skills should be developed without losing much time (Werner etal. 2012). In this context a valid and reliable ‘measurement tool for primary school students has been brought to the literature with this study, ‘The developed Computational Thinking Scale was then applied to primary school students and it was investigated whether stud ‘computational thinking levels differ according to the grade level variable. As a result of the analyzes made, it was found thatthe ‘computational thinking levels of the students difTered according tothe grade level variable In general, it was found that asthe grade level increased, the computational thinking levels ofthe students also increased, When the literature on the subject was examined, few studies were found that investigated the computational thinking level of learners, especialy in terms of different variables. It hhas been seen that the results of the limited number of studies also support the result obtained from this study, Examining the ‘computational thinking skis of secondary school students in terms of various demographic characteristis, Korucu etal. (2017), as 4 result oftheir studies, found that there is a difference between the students’ computational thinking levels in terms of the grade level variable. The researchers stated thatthe computational thinking skill levels ofthe participants differed significantly according to their grade levels, Again, Seiter and Foreman (2013) aimed to determine the differences between the computational thinking skills of students of different ages in their studies and stated that it is necessary to create a rescarch-based and age-appropriate curriculum for primary school students as a result ofthe study. In another study, Catana Kuleli (2018) examined the computational thinking skills of pre-service teachers according to grade level and found that there was a significant difference between the 4th sade and Ist grade pre-service teachers in favor ofthe th grade level. Contrary to the result obtained from this study, Korkmaz et al, (2015) concluded in their study that while the increase in the grade level is expected to inerease along with the computational thinking skills, this skill gradually decreases withthe increase in the grade level. As a result of the study, it was revealed thatthe level of computational thinking of individuals who are active in business lie is high. In the study, it was also aimed to examine whether the students computational shinking levels differ aecording to the daily computer use time. The findings revealed thatthe mean score ofthe scale differed depending on the daily computer usage time, As students! daily computer use time inereases, their computational thinking mean scores also increase. In the literature on the subject, i has ‘been seen that there are a very limited number of studies investigating the effect of daily computer use time on computational thinking. The results obtained from these studies show parallelism with the results of this study. For example, Saritepeci (2017) found that participants with easy computer access had higher computational thinking scores than those without computer access Again, in the research conducted by Yildiz Durak and Saritepect (2018), it was found that computational thinking skills were predicted by the variables of information technology use experience and daily intemet use time. In another study conducted by ‘Qualls and Sherrell (2010), it was suggested that computational thinking is 2 problem-solving approach that consolidates logical skill with computer concepts. According tothe researchers, computational thinking should be included in primary and secondary ‘education programs, and pre-service teachers who will teach in these fields should be familiarized with computational thinking skills, Contrary tothe results oblained fom this study, Korucu etal. (2017) suggested in their study that there was no significant difference in computational thinking skills in terms ofthe variables of woekly internet usage time and mobile device usage abilities, however, their computational thinking skils differed in terms of having mobile technology. In another study conducted by Oluk and Korkmaz (2016), it was revealed that there was no significant difference in the level of computational thinking skills ofthe students according to the variables of computer usage time, but there was a significant relationship between the students’ Seratch programming skills and their computational thinking skill. ‘Recent reports published ia Europe reveal that children should be familiar with computer science concepts from the beginning of their education and gain computational thinking skills (Ozbey & Kucukoglu, 2018), For this reason, computational thinking skills should be taught to students from the earliest ages possible (Wing, 2006). In other studies, to be conducted on the subject, activities ‘that will enable young age groups to acquire computational thinking skills can be developed, Again, it can be examined whether ‘computational thinking skills vary in terms of different variables that were not studied inthis study Limitations ‘This study has several limitations that should be addressed. First of all, only primary school students constitute the participant group of this study. Secondly, it i focused only on students’ computational thinking levels according to different variables, and in this respect, itis quite difficult to reveal the causa relationships between the relationships. Finally, the data collection tool ofthe study is only the computational thinking seale. DECLARATION Funding: No funding source was used for this article. Ethics and Consent: Gazi University Ethies Committee discussed the research permit at its meeting dated 19.04.2022, numbered (08 and decided that there was no ethical problem with the 2022-621 research code. 120 (© 2023, Journal of Learning and Teaching in Digital Age, (1), 113-123Developing Computational Thinking Scale for Primary School Students ‘Computational Thinking Scale Final Form: The final version ofthe developed scale in Turkish is presented in APPENDIX-1 REFERENCES Ambrosio, A. P. Almeida, L. 8, Macedo J, & Franco, A. H. R. (2014), Exploring core cognitive skis of computationsl thinking. Proceedings from Psychology of Programming Interest Group Annual Conference, 25-34 Bar, V, & Stephenson, C. (2011), Bringing computational thinking to K-12: What is involved and what isthe role ofthe computer science education community? ACM Inroads, (1), 48- $4, hitp/dx doi org/10.1145/1929887 1929905. Berikan, B. (2018), Bilgiislemsel diiiame becerisine vineliktasarlanan ‘ver selerivle problem cézme' drenme deneviminin bicimlendiricideerlendirmesi [Formative evaluation of problem solving with datasets" learning experience designed for computational thinking skils). Doktora Teri. Gazi Universtes, Etim Bilimleri Ensttisd, Ankara, Bryman, A. & Cramer, D. (2001). Quantitative data analysis with spss release 10 for windows. USA and Canada: Routledge is an imprint ofthe Taylor & Francis Group. Buyukorturk, S. 2012), Sosyalbilmler ign veri analizi el kitabs [Manual of data analysis for social sciences]. Ankara: Pej ‘Akad Buyukozturk, S., Akgun, O. E., Karadeniz, 8. Demirel, P., & Kili, B, 2016), Bilimsel arastrma yéntemlert [Scientific research methods}. Ankara: Pezem Akademi Cohen, L, Manion, L., & Morrison, K. (2007). Research methods in education. London: Routledge/Taylor & Francis Group. Catana Kuleli, $. (2018), Ogreimen adaylarinn cevrimicl dgrenmeye hacirbulunusluk diceyleri ve bilgi islemse! disinme becerilerinin degertendivibmesi [Evaluation of pre-service teachers' readiness for online learning and computational ‘thinking skills]. Yiksck Lisans Tezi. Dizce Universitesi, Ssyal Bilimlee Ensitis, Dizee Demir, 0. & Seferoglu, S. S, (2017). Yeni kavramlar, farkl kullanemlar: bilg-slemsel distinmeyle igi bir degerlendirme [New concepts, different uses: An evaluation of computational thinking. H. F. Odabasi, B. Akkoyunlu ve A. Isman (Ed), Egitim ‘eknolojiler okumalsn, 2017 (ss. 468- 484) igi. Ankara: TOJET-Sakarya Universite DODEA- Department of Defense Education Activity, (2014). The 21* century principal. Retrieved January 13, 2021, from hntpsu/eontent.dodea,edwteach Learnprofessional_development/21/docs'prncipalsiprincipal paper draft.pat Fraenkel, JR. & Wallen, N.E. 2009). How to design and evaluate research in education, New York: McGraw-Hill Ghasemi, A. & Zahedias, 8. (2012). Normality tests for statistical analysis: @ guide for non-statisticians. International Journal of Endocrinology and Metabolism, 10, 486-489. bips:doiog/10,5812%2Fijem. 3505 Gouws, L. A., Bradshaw, K., & Wentworth, P. (2013). Computational thinking in educational activities: An evaluation of the ciducatonal game light-bot. Paper presented at the 18th ACM Conference on Innovation and Technology in Computer Seience Education, ltaly Gravette, F. J. & Forzano, LB, 2012), Research methods for the behavioral sciences. Belmont, CA: Wadsworth. Grover, S. & Pea, R. 2013). Computational thinking in K-12: A review of the state ofthe field, Educational Researcher, 42(1), 3843 htp/dx.doiorg/10.3102/0013189X12463051 Grover, 8, Pea, R, & Cooper, S. 2015). “Systems of assessments" for deeper learning of computational thinking in K-12. Paper presented atthe Annual Mecting of the American Educational Research Association, Chicago, United States of Amerika Gulbshar, ¥. Ker, §. B., & Kaleloglu F. (2019), Bilg glemsel digtinme becerisine ynelik dz yeterlik alist Slgegi: Gegerlik ve svenirik gaismas [Self-efficacy perception scale for computational thinking skills: Validity and reliability study. Tirk Bilgisayar ve Matematik Egitimi Dergisi 10{1), 1-29. hips: doiong/ 0.16949 turkbilmat 3850 Hart, M. C. (1996). Improving the discrimination of SERVQUAL by using magnitude scaling. G. K. Kanji (84), Totel Quality Management in Action, London: Chapman and Hall ISTE- Intemational Society for Technology in Education, (2011). Computational thinking in K-12 education leadership toolkit Retrieved February 12, 2022 from http: ww. jsteong/docs/et documents‘ctleadershipt-toolkit péf?sfvr=n=4 Kaleliogl, F., Gulbahar, Y., & Kukul, V. 2016). A framework for computational thinking based on a systematic research review Baltic Journal of Modern Computing, 43), 583-596, Kolemkus, F. & Bulut Ozek, M. (2021). 21. yiiyil becerileri konusunda arastrma efilimleri: 2000-2020 [Research tends in 21st century skills: 2000-2020], Manas Sosyal Arastirmalar Dergisi,10(2), 878-800. btps!/Go.org/10.33206imjss. 774848 Kerasar, N. (2017). Bilimsel aragnrma yintemi [Scientific research method]. Ankara: Nobel Yaymnek. Kilic, "S. @016).Cronbachs alpha reliability coefficient. Journal of Mood Disorders 6(1)1 butpslidi.org/10.5855imood.20160307122825, Kline, P1994). An easy guide to factor analysis. London: Routed Koklo, N. (1997). Tutumlarm lgélmesi ve likert tipi algeklerde kullanilan sepenckler [Measuring atitudes and options used in likert-type scales). Ankara Universitesi Etim Bilimleri Faktesi Dergisi, 28(2), 81-98. Korkmaz, 0., Cakir, R., & Ozden, M. Y. (2017). A validity and reliability stady of the computational thinking scales (CTS), Computers in Fluman Behavior, 72, 558-569. htps:/ido.org/10.1016),chb 2017.01.00 Korkmaz, O., Cakit, R, & Ozden, M. Y., Oluk, A, & Sariogly, S., 2015). Biteylern bilgisayarca dgiinme becerlernin farklt eiskenler agisindan incelenmesi (Examination of individuals’ computational thinking skills in terms of different variables]. Ondokuz Mayis Cniversitesi Egitim Fakiitesi Dergisi, 34(2), 6887, Leech, N. L., Bare, KC, & Morgan, G. A. (2005). SPSS for Intermediate Statistics: Use and Interpretation. NI: Lawrence Enlbaum Associates, Inc. MEB- Milli Egitim Bakanlig), 2018). Bilisim teknolojleri ve yazsh derst diretim programs (lkokul 1, 2, 3 ve 4. smyfar) [aformation technologies and software curriculum (Primary School Ist, 2nd. 3rd and 4th grades). Ankara: Mili Egitien Bakanhgs Yaymlan. pI © 2025, Joumal of Learning and Teaching in Digital Age, 8(1), 113-123E Yildirim, & ¢. Uluyol Miller, RG. (1969), Simultaneous statistical inference, New York: McGraw-Hill. ‘luk, A. & Korkmaz, 0. (2016). Comparing students’ seratch skills with their computational thinking skills in terms of diff erent variables. £ J. Modern Education and Computer Science, 11, 1-1. bitpl/dx.doi.org/10.581S/ijmees.2016.11.01 Oxbey, N. & Kugukogla, A, (2018), Ingiltere, Avustralya ve Tirkiye'nin biligim teknolojileri oretim_ programlarmin karyastinimast [The comparison of curriculum regarding information technologies in England, Australia and Turkey], International Journal of Innovative Approaches in Education, 23), 76-109. hipstidoi.org/ 0.29329 ijiape 2018.177.2 ‘Orkan, R.& Bindak, R. (2021). Likert tpi digeklerde kat dizeyi sayisindaki degiskligin psikometrik 6zeliklerinin incelenmnest ‘Examining the psychometric properties of the change in the number of participation levels in Likert-type scales). Nicel Bilimler Dergisi, 32), 150-172, hlips:/doi.org/10.51541/nicel. 1028839 ‘Ozmen, B. (2016). Oraokul dgrencilerine ydnelk bile islemsel disiinme becerleri testinin gelistiilmesi: gegerlik ve givenitlik ¢abymas: [Development of computational thinking skills test for secondary school students: validity and reliability study Conference: 4th International Instructional Technologies & Teacher Education Symposium. Elazi® P2I- Parinershipfor 21 Century Skills, (2009). P21 framework definitions. Retrieved Februsry 02, 2022 from biip/www.p2 I ore/storage/documents/P21Framework Defiitions pd. Polat, C. (2006). Bilgi sagonda tniverste egtimi igin bir agama: Bilgi okuryazarngu Sretimi [An opening for university education {in the information age: Information literacy teaching]. A. U. Tirkivat Arastarmalar: Ensttsii Dergisi, 30, 249-266, Qualls, J. A & Sherell, L. B. (2010), Why computational thinking should be integrated into the curriculum, Jounal of Computing Sciences in Colleges, 25(5), 66-71 Saritepeci, M. (2017). Ortadgretim dizeyinde bilg-isemsel diinme beccrisinin gesiti degiskenler agisindan incelenmesi ‘Examination of computational thinking skills at secondary education level in terms of various variables). Fifth International Instructional Technologies Teacher Education Symposium (ITTES), kzmir, Turkiye. Scherer, R. F., Wiebe, F. A., Luther, D.C., & Adams, J S. (1988). Dimensionalty of eoping: Factor stability using the Ways of Coping Questionnaire, Psychological Reports, 62(3), 763170, hitp:/doi.org/10.2466ipr0,1988.62.3,763 Sciter, L., & Foreman, B. (2013). Modeling the learning progressions of computational thinking of primary grade students. Proceedings ofthe Ninth Annual International ACM Conference on International Computing Education Research, 59-66. Selby, C. & Woollard, J. 2013). Computational thinking: the developing definition. Special Interest Group on Computer Science Education (SIGCSE), Atlanta GA. Sim, J. & Wright, C. (2002), Research in health care: concepts, designs and methods. United Kingdom, Cheltenham: Nelson Thornes Singh, K. (2007). Quantitative social research methods. Los Angeles: Sage Publications. Snow, E., Katz, I, Elliot-Tew, A., & Feldman, J. (2012). Assessing computational thinking. NSF-CE21 Community Meeting, ‘Washington, United States of America, Sendurur, P. (2012). Identification of factors affecting integration of information and communication technologies in basic education schools grades from 4 through 8, Doktora Tezi. Orta Dogu Teknik Universitesi, en Bilimleri Ensttisd, Ankara, ‘Tavsancil, E. (2010). Tutumlarin dleiimesi ve SPSS ile veri analizi (Measuring attitudes and data analysis with SPSS]. Ankara: Nobel Yayin Dagit Wemer, L., Denner, J. Campe, 8, & Kawamoto, D, C. (2012), The fairy performance assessment; measuring computational thinking in middle school, Proceedings of the 43rd ACM Technical Symposium on Computer Science Education, 215-220. Wing, J. M. (2006). Computational thinking, Communications ofthe ACM, 49(3), 33-35, https:/doi.org/10. 1145/1 118178.1118215 Wing, J. M. (2008). Computational thinking and thinking about computing, Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences, -366(1881), 3717-3725, ‘utps:/do.org/10.1098/sta.2008.0118 Wing, J. M. 2011), Research notebook: computational thinking -what and why’? The Link Magazine, 20-23, Retrieved January 10, 2022 from htpss/www.cs.cmu edulink/research-notcbook-computational-thinking-whal-andwhy Yildiz Durak, H. & Saritepeci, M, (2018). Analysis of te relation between computational thinking skills and various variables with the structural equation model. Computers & Education, 116, 191-202. hups:/doi.org/10.10161}.compedu.2017.09.004 Yildiz, M. (2021). Bilgiijlemsel disinme becerisinin sireg temelli dlciimesi ve degerlendiriimesi [Process-based measurement ‘and evaluation of computational thinking skills]. Doktora Teri. Trabzon Universtesi, Lisansisti Egitim Enstitis, Trabzon Yoleu, V. (2018). Progranlama egitiminde robotik kullanmimn akademik hasan, bilg-islemsel disiinme becerisi ve dgrenme ransferineetksi [The effect of using robotics in programming education on academic achievement, computational thinking, skills and learning transfer]. Yksek Lisans Tezi. Sileyman Demirel Universitesi, Pitim Bilimleri Ensttis, Isparta 1 © 2023, Journal of Learning and Teaching in Digital Age, 8(1), 113-123Developing Computational Thinking Scale for Primary School Students APPENDIX-1: Computational Thinking Scale Final Fon BILGI iSLEMSEL DUSONME OLCEGL Itadeler = Kathyorum Katimyorum 1. Bir probleme kargilagigimda ([2-Problem gGzerken arkadaslanmla gérev paylasim: yapamam. salim yolunu bilgisayarda ararum. | 3. Problemi her zaman dof tanimlarim, [4 Cozi ign yapilmast gerekenler listelerim: S. Cozi igin gerckli adimlan takip edemem. [6 Benzer problemleri gizerken aym géziim youu denerim. 7 Cozim yollan igerisinden en pratik olan: socerim, [.Problem gGzerken tiim ¢Ozim yollarim: pozden 9, afta sony ddevlerimalnlere bilerim, (Tio. Gdeverins bilgisayarda yazmaktazoelananan 1. Problemin ¢éziimi igin uygun admlar: olusturarum. 12, Problemi gézalikien sonra salamasim yapamam. 13. Dir iglemi hesap makinesinde yapmays tercih ederim 14, Problem gSzerken en uygun gOzim yolunu bulamam, 15. Problem gzme basamaklarindaki hatalan tespitedezim, 16, Problemin hepsini tek seferde defil parga parya gozerim, 17. Problem igerisinde verlenler ve istenenleri bulurum, 1 2023, Jounal of Learning and Teaching in Digital Age, (1), 113-123