Chae.Lee.2019.2ijee

International Journal of Engineering Education Vol. 35, No. 2, pp. 617–622, 2019 0949-149X/91 $3.00+0.00 Printed in Great Britain # 2019 TEMPUS Publications. Student-Centered Learning and Higher-Order Thinking Skills in Engineering Students* SOO EUN CHAE1 and MI SUK LEE2 1 Associate professor, Jukheon gil 7, Gangneung-Wonju National University, South Korea. E-mail: schae@gwnu.ac.kr 2 Lecturer, Jukheon gil 7, Gangneung-Wonju National University, South Korea. E-mail: ssuk0508@gwnu.ac.kr The current study aimed to investigate the use of higher-order thinking (HOT) skills by engineering students and to classify student profile groups according to the underlying constructs of HOT. We recruited 260 engineering students from six universities in South Korea. The data were analyzed in terms of the existing latent profiles and the chi-square test between the profile groups and their experience of different types of instruction. The latent profile analysis revealed that the use of HOT skills could be classified into four groups (i.e., a lower-order thinking group, a creative and argumentative group, an analytical and caring group, and a HOT group). A chi-square test between the four categorizations of HOT skill uses and instruction methods was not statistically significant. A majority of the students were classified in the HOT group. However, of the six constructs, the creativity skill was the least used, as opposed to skills that fell under other constructs. Therefore, supplementary instruction to fill this gap is suggested. Keywords: engineering college students; higher-order thinking skills; latent profile analysis 1. Introduction related to students’ use of HOT skills. Based on an in-depth observation of students in scientific inquiry Recently, much research has been conducted on the classes, Marshall and Horton [12] reported that the effectiveness of various classes that aim to cultivate level of students’ intellectual ability (or higher higher-order thinking (HOT) skills to improve the thinking skills) was negatively associated with the academic achievement of science and engineering time spent exploring problems. In other words, students. Such student-centered classes include students with less developed intellectual skills active learning [1], problem-based learning [2], and spent more time exploring than managing or solving inquiry-based learning [3, 4]. The advantage of this problems. Therefore, it is concluded that one must student-centered approach is that students are likely provide suitable steps for each student according to to increase their level of instructional involvement their cognitive levels rather than simply providing and thus to solve problems as well as ultimately to them with problem-based instructions. develop HOT skills [5] such as application, analysis, Another issue in developing engineering students’ synthesis, and evaluation, which were classified as HOT skills is how the underlying constructs of HOT high-level activities in Bloom’s taxonomy [6]. will be manifested to students. It is unclear how Most scholars who have studied HOT skills have these underlying constructs actually appear to stu- emphasized scientific thinking processes such as dents once the HOT skills are defined in considera- questioning and inquiry [7, 8]. This trend is due to tion of the students’ characteristics. Will the the paradigm of traditional educational psychol- underlying constructs appear similar, especially ogy, which interprets thought from a cognitive among engineering students, for whom authentic point of view. In recent years, however, the view problem solving is important? Or will they have a that positive aspects such as ‘‘care’’ must be certain pattern and be divided into different groups? acknowledged as a subculture of thinking, espe- If several different groups of HOT skill patterns are cially of high thinking ability, has emerged. revealed, what is the relationship between the Lipman [9, 10], a representative scholar who groupings of these tendencies and the types of makes such claims, regarded HOT skills as a com- instruction (e.g., student-centered, instructor-cen- bination of critical thinking, creative thinking, and tered) in which they appear? This study was con- caring thinking. According to this point of view, a ducted to answer these questions. person with HOT skills tends to consider others The current study is expected to benefit research- when examining problematic situations to find a ers and practitioners in engineering education by basis for solving the problem, to synthesize various demonstrating ways to infer the evaluation account- points of view and to create alternative plans [11]. ability of student characteristics. For instance, these However, it is difficult to conclude that inquiry- results can be used for university-level evaluation in based learning and student-centered classes are courses at universities. The use of HOT skills by * Accepted 12 December 2018. 617 618 Soo Eun Chae and Mi Suk Lee engineering students is also a subject of major of the relevant items based on a 5-point Likert scale interest for those involved in course assignment (1 = very unlikely, 5 = very likely). The reliability of and curriculum development at the college level. the HOTUS was reported as 0.879 [10]. Any technical university-level course that requires To understand the relevance of HOT to class- HOT skills can also take advantage of this study. room practices, the participants were asked about Engineering-related studies have highlighted the the teaching methods they had received in the past importance of HOT; however, little research has and divided them into three categories: professor- been conducted to determine the relationship centered, instructor + student style, and student- between HOT and the class types of engineering centered. students. Given the importance of HOT in engineer- ing, this study should be of interest to any engineer 2.3 Latent profile analysis interested in developing an educational curriculum. The LPA function in M plus 7.11 [13] was used to obtain the HOT profile of the participating college 2. Methods students. The LPA is a person-oriented approach that distinguishes groups through personal charac- The latent profile analysis (LPA) and chi-square teristics and attributes [14]. For the final model tests were conducted to determine how students in selection, the model was evaluated based on classi- engineering majors use HOT skills and how each fication quality, information index, and model group relates to the type of instruction they comparison test. First, the quality of the classifica- received. tion was confirmed through the entropy value, and the estimation equation was as follows [15]: 2.1 Information source The subjects of this study were a randomly selected i k Pik lnPik Ek ¼ 1 ð1Þ group of engineering college students from six NlnðkÞ universities in South Korea. The data were collected from 266 college students in engineering, and the where Pik is the posterior probability of a person analysis used data from 260 samples after six unreli- who belongs to group k, N is the sample size, and K able respondents’ data were excluded. Of the is the number of latent classes. Ek is between 0 and 1, respondents, 79 (30.4%) were female, 101 (38.8%) and the probability of belonging to one latent class were in the first year, 30 (11.5%) were in the second is close to 1; as the probability of belonging to year, and 72 (27.7%) were in the third year. another latent class becomes closer to 0, the Pik value increases. A value of approximately 0.8 or 2.2 Higher-order thinking skills scale for university more is considered a good classification value [16]. students The Akaike information criterion (AIC) [17], To measure the HOT skills of college students, we Bayesian information criterion (BIC) [18] and used the higher-order thinking skills scale for sample-size adjusted BIC (SABIC) [19] were used Korean university students (HOTUS) created by as information indices. For the model comparison Lee [10]. The HOTUS consists of the subcategories test, the Lo-Mendell-Rubin adjusted likelihood of creative thinking, critical thinking, and thought- ratio test (LMRLRT) [20] and parametric boot- ful thinking. Each subcategory consists of five items strapped likelihood ratio test (BLRT) [21] were related to creativity, analysis, argumentation, dia- used. lectic, and caring that explain HOT skills. Creativity A chi-square test was conducted to determine the is a cognitive ability to generate new and useful ideas relationship between the latent profiles and the when facing problem situations. Analysis is a cog- instruction method that the students received in nitive ability used to contemplate a problem situa- the past. This process was performed by checking tion in detail. Argumentation is a cognitive ability the independency between two variables through used to provide adequate grounds for justifying SPSS version 22.0 [22]. conclusions about problem situations. Dialectic is a cognitive ability to provide, synthesize and devise 3. Results new forms of alternatives based on various perspec- tives regarding problem situations. Caring is a 3.1 Latent profile analysis cognitive ability to seek reasonable approaches or Table 1 summarizes the descriptive statistics of the thoughts about problem situations based on the subvariables of HOT skills employed in the LPA. As interest and empathy of others. The HOTUS was seen in the table, the five subvariables of HOT skills composed of a total of 25 items for creativity (4), showed a statistically significant correlation of an analysis (4), argumentation (5), dialectic (5), and approximately medium level, but between some caring (7). Each subcategory score was a mean score matching pairs, such as analysis and creativity and Student-Centered Learning and Higher-Order Thinking Skills in Engineering Students 619 Table 1. Correlations among Subvariables of Higher-Order Thinking Skills 1 2 3 4 5 1. Analysis 1 2. Creativity 0.42** 1 3. Argument 0.22** 0.11 1 4. Dialectic 0.53** 0.38** 0.26** 1 5. Caring 0.46** 0.38** 0.26** 0.48** 1 Mean 3.70 3.69 2.93 3.50 3.37 S.D. 0.66 0.58 0.674 0.61 0.588 Notes. *p < 0.05, **p < 0.01. Table 2. Model Fit in Latent Profile Analysis AIC BIC SABIC Entropy LMRLRT BLRT 2 Profile 14942.9 15213.6 14972.6 0.903 0 0 3 Profiles 14759.0 15122.2 14798.8 0.895 0.4403 0 4 Profiles 14599.8 15055.6 14649.8 0.903 0.2794 0 5 Profiles 14497.6 15045.9 14557.7 0.924 0.4344 0 Table 3. Dependency test LPA  Instruction method Instruction Instructor-centered Instructor + students Student-centered Total Group 1. Lower-order thinking 12 (92.3%) 1 (7.7%) 0 (0%) 13 2. Analytical and caring 58 (86.6%) 4 (6.0%) 5 (7.5%) 67 3. Higher-order thinking 96 (87.3%) 6 (5.5%) 8 (7.3%) 110 4. Creative and argumentative 57 (81.4%) 9 (12.9%) 4 (5.7%) 70 Total 223 (85.8%) 20 (7.7%) 17 (6.5%) 260 Note. The percentage represents students in each latent profile group out of the students in each instruction method. caring and creativity, the correlation was very low Figure 1 outlines the four latent profiles (Groups or hardly existent. 1 through 4) produced as above mentioned. Those Table 2 shows the results of the LPA the HOT groups can be named as followings based on their skills of engineering students. In the LPA, the characteristics: Group 1 is ‘‘lower-order thinking number of profiles exposed by the data is deter- group’’, Group 2 is ‘‘analytical and caring group’’, mined based on the ease of interpretation and the fit Group 3 is ‘‘HOT group’’, Group 4 is ‘‘creative and of the model. The fit of the model was checked by argumentative group’’. increasing the number of profiles from two to five. The characteristics of each type of profile are as The comparison results of model fits are well follows: First, Group 1 showed low scores in most differentiated when the entropy index is 0.8 or indicators, including analysis, argumentation, dia- higher [13]. The classification was almost accurate, lectic and caring, compared with other groups. This with the entropy values of the models compared in group can be termed a ‘lower-order thinking group.’ this study all showing an entropy index of 0.8 or Thirteen students belonged to this group, account- more. The AIC, BIC, and SABIC values all ing for 5% of all the students. Group 2 had high decreased as the number of latent profiles increased. scores in analysis, argumentation and caring and The LMRLRT was significant for up to four relatively low scores in creativity and dialectic. models, and the BLRT was significant for all Given the high construct correlation between crea- models, so it is difficult to determine which model tivity and argumentation among the lower HOTUS is the best fit for these criteria. A researcher’s scores, this group can be understood as having weak judgment becomes important when multiple indices creative and divergent thinking but exceptionally show different results [23], and in such cases, it is strong analytical and caring abilities. Therefore, it necessary to comprehensively consider various fit can be called an ‘analytical and caring group.’ indices and theoretical interpretability. A four- Group 2 had 67 students, comprising 26% of the group classification was employed in this study, total number of students. Group 3, the so-called comprehensively taking into account the ease of ‘HOT group’ whose scores were high across all interpretation, fit index and other aspects. subconstructs, contained 110 students, accounting 620 Soo Eun Chae and Mi Suk Lee Fig. 1. Means of item scores per Latent Profile of Higher-Order-Thinking skills in engineering students. for 42% of the total number of students. Group 4 student-centered approach than other groups. earned relatively high scores, ranking second in However, the changing trends of student-centered creativity, argumentation and dialectic among all approach across groups were not statistically sig- groups, but its caring score was relatively low. This nificant, which meant it was difficult to distinguish group, which can be called a ‘creative and argumen- group differences based on the student-centered tative group,’ contained 70 students, or 27% of the instruction method. total number of students. 3.2 Independency between the LPA and the 4. Discussions instruction method The purpose of this study was to investigate how A chi-square independency test was conducted to engineering school students use HOT skills and how see whether the LPA classification is related to the their current use of HOT skills is related to the types group members’ experience of instruction methods. of teaching they have received in the past. The results indicated no statistical significance (2 = The engineering school students showed four 4.775, DF = 6, p = 0.576). Therefore, one cannot groups with HOT skills: a considerable number of reject the hypothesis that the latent profile group is participants were classified as the ‘‘HOT group’’ independent of the instruction method. In other (Group 2), which frequently uses all of the HOT words, the two variables can be said to be dependent skills. In addition, a ‘‘creative and argumentative on each other. The student numbers for each group group’’ (Group 4), an ‘‘analytical and caring group’’ are presented in Table 3. The independency between (Group 3), and a ‘‘lower-order thinking group’’ the two variables appeared, and the main effect for (Group 1) frequently appeared among the engineer- each variable should be meaningful. Most of the ing students. However, most students, including engineering students who participated in this study those classified in the ‘‘HOT group,’’ used relatively (N = 223, 85.3%) received instructor-centered less of the creativity skill. classes. Of the classified groups, Group 1 (a lower- Most of the engineering students who partici- order thinking group) was prone to experience an pated in this study reported that they had received instructor-centered approach. All the students in instructor-centered classes. A very small number of Group 1 except one reported that they had received students had received student-centered classes, and instructor-centered classes. The likelihood of receiv- many of them were in the analytical and caring ing instructor-centered classes decreased from group (Group 2) or the HOT group (Group 3). Group 1 to Group 3, Group 2, and Group 4. This may reflects the recent concerns of engineering Interestingly, an analytic and caring group (Group educators to highlight development of students’ 2) reported a greater likelihood of experiencing a sustainable thinking. However, the type of instruc- Student-Centered Learning and Higher-Order Thinking Skills in Engineering Students 621 tion experienced did not show a significant relation- 3. A. J. Swart, Evaluation of Final Examination Papers in Engineering: A Case Study Using Bloom’s Taxonomy, ship with the HOT group category. IEEE Transactions on Education, 53(2), pp. 1–8. The relevance of the use of HOT skills and the 4. G. V. Madhuri, V. S. S. N. Kantamreddi and L. N. S. types of classroom instruction that the engineering Prakash Goteti, Promoting higher order thinking skills using inquiry-based learning, European Journal of Engineer- students had received was not significantly promi- ing Education, 32(2), pp. 117–123. nent in this study. On one hand, the reason for this 5. W. Pan and J. Allison, Exploring project based and problem deviation may be related to the fact that many of the based learning in environmental building education by integrating critical thinking, International Journal of Engi- participating students were classified in the HOT neering Education, 26(3), pp. 547–553, 2010. group that used advanced thinking skills. 6. B. Bloom, Bloom’s Taxonomy, Learning, 2010, (1956). These results are somewhat encouraging; how- https://doi.org/10.1016/S0022-3913(12)00047-9. 7. A. Lewis and D. Smith, Defining Higher Order Thinking ever, it is necessary to provide courses that empha- Theory Pract., 32(3), pp. 131–137, 1993. size creative thinking skills, considering that most 8. P. A. Alexander, D. L. Dinsmore, E. Fox, E. M. Grossnickle, engineering students in the current study used less S. M. Loughlin, L. Maggioni, M. M. Parkinson and F. I. Winters, Higher Order Thinking and /knowledge: Domain- creative thinking skills than other HOT skills. general and domain-specific trends and future directions, Indeed, previous research has pointed out lack of Assess. High. Order Think. Ski. pp. 47–88, 2011. student-centered instruction [24] despite much of 9. M. Lipman, Thinking in education, Second edition (Cam- bridge University Press, Cambridge: U.K. the engineering education curricula requires the 10. M. Lipman, Moral education higher order thinking and pedagogical paradigm-shift [25]. On the other philosophy for children, Early Child Development and Care, hand, it is possible that simply enforcing student- 107(1), pp. 61–70. 11. M.-S. Lee, Development of the Higher-Order Thinking skill centered instruction may not necessarily be helpful scale for Korean University Students, Unpublished Doctoral in promoting HOT skills for engineering students. Disseration, Gangneung-Wonju National University, South This tendency has already been noted in the work of Korea, 2016. 12. J. C. Marshall and R. M. Horton, The relationship of Marshall and Horton [12]. Theoretical studies to teacher-facilitated, inquiry-based instruction to student further confirm this tendency should be conducted higher-order thinking, Sch. Sci. Math., 111(3), pp. 93–101, in the future. 2011. 13. B. O. Muthén and K. Muthén, Mplus User’s Guide (Sixth Edition), 2007. 5. Conclusions 14. L. R. Bergman and D. Magnusson, A person-oriented approach in research on developmental psychopathology, The engineering students’ thinking styles have four Development and Psychopathology, 9(2), pp. 291–319, 1997. 15. J. G. Dias and J. K. Vermunt, Bootstrap methods for latent profiles such as ‘‘lower-order’’, ‘‘analytical measuring classification uncertainty in latent class analysis, and caring’’, ‘‘higher-order’’, and ‘‘creative and Heidelberg: Springer, 2006. argumentative’’ according to our Latent Profiles 16. S. L. Clark, Mixture modeling with behavioral data, Uni- versity of California, Los Angeles, CA, 2010. Analysis. Of these four types, ‘‘lower-order thin- 17. H. Akaike, A new look at the statistical model identification, kers’’ appeared to have more experiences of instruc- IEEE Trans. Automat., 19(6), pp. 716–723, 1974. tor-centered classes than other three profiles. In 18. G. Schwartz, Estimating dimensions of a model, The Annals of Statististics, 6(2), pp. 461–464, 1978. conclusion, engineering students require more of 19. L. Sclove, Application of model-selection criteria to some the student-centered classes than instructor-cen- problems in multivariate analysis, Psychometrik, 52(3), pp. tered classes for further higher-order thinking 333–343, 1987. 20. Y. Lo, N. R. Mendell and D. B. Rubin, Testing the number of skills including analytical and caring skills. How- components in a normal mixture, Biometrika, 88(3), pp. 767– ever, our statistical analysis did not verify the group 778, 2001. differences at their significance levels, which needs 21. G. J. McLachlan and D. Peel, Finite mixture models, New York: Wiley, 2000. further investigation. 22. J. L. Arbuckle, IBM SPSS AmosTM 22 User’s Guide, Amos 22 User’s Guide, 2013. References 23. S. Kim and K. J. Hong, A study on the welfare attitude of Seoul citizens using the Latent Class Analysis, Korea Journal 1. S. Freeman, S. L. Eddy, M. McDonough, M. K. Smith, N. of Social Welfare, 37(2), pp. 95–121, 2010. Okoroafor, H. Jordt and M. P. Wenderoth, Active learning 24. E. Justo, A. Delgado, M. Vazquez-Boza and L. A. Branda, increases student performance in science, engineering, and Implementation of Problem-Based Learning in structural mathematics, Proc Natl. Acad. Sci. USA, 111(23), https:// engineering: A case study, International Journal of Engineer- doi.org/10.1073/pnas.1319030111, 2014. ing Education, 32(6), pp. 2556–2568, 2016. 2. M. Z. Mokhtar, M. A. A. Tarmizi, R. A. Tarmizi and A. F. 25. D. N. Huntzinger, M. J. Hutchins, J. S. Gierke and J. W. M. Ayub, Problem-based learning in calculus course: percep- Sutherland, Enabling sustainable thinking in undergraduate tion, engagement and performance, Latest Trends Eng. Educ. engineering education, International Journal for Engineering pp. 21–25, 2010. Education, 28(2), p. 218, 2007. Soo Eun Chae is an associate professor in the Teacher Education Division at Gangneung-Wonju National University, South Korea. She received her BA and MA degrees in Education Major and English Literature Minor from SungKyunKwan University, Seoul, Korea, and her PhD degree in Human Development and Quantitative Methodology from the University of Maryland at College Park, USA. Her research interests are cognitive development and technology use in learning. 622 Soo Eun Chae and Mi Suk Lee Mi Suk Lee received her MEd and PhD from Gangneung-Wonju National University, South Korea. She worked as a researcher at CTL (Center for Teaching & Learning) of Gangneung-Wonju National University. Currently, she teaches Educational Technology and Educational Psychology to the students in Gangneung-Wonju National University and Catholic Kwandong University. Her research interests are Higher-Order Thinking, learning strategy, and constructivism.