Abstract
Constructing scientific arguments is an important practice for students because it helps them to make sense of data using scientific knowledge and within the conceptual and experimental boundaries of an investigation. In this study, we used a text mining method called Latent Dirichlet Allocation (LDA) to identify underlying patterns in students written scientific arguments about a complex scientific phenomenon called Albedo Effect. We further examined how identified patterns compare to existing frameworks related to explaining evidence to support claims and attributing sources of uncertainty. LDA was applied to electronically stored arguments written by 2472 students and concerning how decreases in sea ice affect global temperatures. The results indicated that each content topic identified in the explanations by the LDA— “data only,” “reasoning only,” “data and reasoning combined,” “wrong reasoning types,” and “restatement of the claim”—could be interpreted using the claim–evidence–reasoning framework. Similarly, each topic identified in the students’ uncertainty attributions— “self-evaluations,” “personal sources related to knowledge and experience,” and “scientific sources related to reasoning and data”—could be interpreted using the taxonomy of uncertainty attribution. These results indicate that LDA can serve as a tool for content analysis that can discover semantic patterns in students’ scientific argumentation in particular science domains and facilitate teachers’ providing help to students.
Similar content being viewed by others
References
Abdous, M., & He, W. (2011). Using text mining to uncover students’ technology-related problems in live video streaming. British Journal of Educational Technology, 40(5), 40–49.
Abdous, M. H., Wu, H., & Yen, C. J. (2012). Using data mining for predicting relationships between online question theme and final grade. Journal of Educational Technology & Society, 15(3), 77–88.
Allchin, D. (2012). Teaching the nature of science through scientific errors. Science Education, 96(5), 904–926.
Akçapınar, G. (2015). How automated feedback through text mining changes plagiaristic behavior in online assignments. Computers & Education, 87, 123–130.
Aufschnaiter, C. V., Erduran, S., Osborne, J., & Simon, S. (2008). Arguing to learn and learning to argue: Case studies of how students’ argumentation relates to their scientific knowledge. Journal of Research in Science Teaching, 45, 101–131.
Beggrow, E. P., Ha, M., Nehm, R. H., Pearl, D., & Boone, W. J. (2014). Assessing scientific practices using machine-learning methods: How closely do they match clinical interview performance? Journal of Science Education and Technology, 23(1), 160–182.
Berland, L. K., & Reiser, B. J. (2009). Making sense of argumentation and explanation. Science Education, 93, 26–55.
Bricker, L. A., & Bell, P. (2008). Conceptualizations of argumentation from science studies and the learning sciences and their implications for the practices of science education. Science Education, 92, 473–493.
Chang, J., Gerrish, S., Wang, C., Boyd-Graber, J. L., & Blei, D. M. (2009). Reading tea leaves: How humans interpret topic models. In Y. Bengio, D. Schuurmans, J. D. Lafferty, C. K. I. Williams, & A. Culotta (Eds.), Advances in neural information processing systems (pp. 288–296). New York: Curran Associates.
Chen, B. (2014). Visualizing semantic space of online discourse: The Knowledge Forum case. In Proceedings of the 4th international conference on Learning Analytics and Knowledge (LAK ’14), 24–28 March 2014, Indianapolis, IN (pp. 271–272). New York: ACM. https://doi.org/10.1145/2567574.2567595.
Chen, B., Chen, X., & Xing, W. (2015). “Twitter archeology” of learning analytics and knowledge conferences. In Proceedings of the 5th international conference on Learning Analytics and Knowledge (LAK ’15), 16–20 March 2015, Poughkeepsie, NY (pp. 340–349). New York: ACM. https://doi.org/10.1145/2723576.2723584.
Chen, N. S., Wei, C. W., & Chen, H. J. (2008). Mining e-Learning domain concept map from academic articles. Computers & Education, 50(3), 1009–1021.
Clark, D. B., & Sampson, V. (2008). Assessing dialogic argumentation in online environments to relate structure, grounds, and conceptual quality. Journal of Research in Science Teaching, 45, 293–321.
de Vries, E., Lund, K., & Baker, M. (2002). Computer-mediated epistemic dialogue: Explanation and argumentation as vehicles for understanding scientific notions. The Journal of the Learning Sciences, 11(1), 63–103.
Dikli, S. (2006). An overview of automated scoring of essays. The Journal of Technology, Learning and Assessment, 5(1), 1–36. Retrieved from https://ejournals.bc.edu/index.php/jtla/article/view/1640.
Dunlosky, J., & Rawson, K. A. (2012). Overconfidence produces underachievement: Inaccurate self evaluations undermine students’ learning and retention. Learning and Instruction, 22(4), 271–280.
Duschl, R. A., Schweingruber, H. A., & Shouse, A. W. (2007). Taking science to school: Learning and teaching science in grades K-8. Washington, DC: National Academy Press.
Erduran, S., Simon, S., & Osborne, J. (2004). TAPping into argumentation: Developments in the application of Toulmin’s argument pattern for studying science discourse. Science Education, 88, 915–933.
Ezen-Can, A., Boyer, K. E., Kellogg, S., & Booth, S. (2015). Unsupervised modeling for understanding MOOC discussion forums: A learning analytics approach. Proceedings of the fifth international conference on learning analytics and knowledge (pp. 146–150). New York: ACM. https://doi.org/10.1145/2723576.2723589.
Feldman, R. D. (1995). Knowledge discovery in textual databases (KDT). Paper presented at the first international conference on knowledge discovery and data mining (KDD-95), August 20–21, Montreal, Canada.
Harish, B. S., Guru, D. S., & Manjunath, S. (2010). Representation and classification of text documents: A brief review. International Journal of Computer Applications, Special Issue on RTIPPR, 2(1), 110–119.
Hotho, A., Nürnberger, A., & Paaß, G. (2005). A brief survey of text mining. Ldv Forum, 20(1), 19–62.
Huff, J. D., & Nietfeld, J. L. (2009). Using strategy instruction and confidence judgments to improve metacognitive monitoring. Metacognition and Learning, 4(2), 161–176.
Hung, J. (2012). Trends of e-learning research from 2000 to 2008: Use of text mining and bibliometrics. British Journal of Educational Technology, 43(1), 5–16.
Janasik, N., Honkela, T., & Bruun, H. (2009). Text mining in qualitative research. Organizational Research Methods, 12(3), 436–460.
Kahneman, D., Slovic, S. P., Slovic, P., & Tversky, A. (Eds.). (1982). Judgment under uncertainty: Heuristics and biases. Cambridge: Cambridge University Press.
Kuhn, D. (1993). Science as argument: Implications for teaching and learning scientific thinking. Science Education, 77(3), 319–337.
Lee, H. S., Liu, O. L., Pallant, A., Roohr, K. C., Pryputniewicz, S., & Buck, Z. E. (2014). Assessment of uncertainty-infused scientific argumentation. Journal of Research in Science Teaching, 51(5), 581–605.
Lee, H. S., Pallant, A., Pryputniewicz, S., & Lord, T. (2017). Articulating uncertainty attribution as part of critical epistemic practice of scientific argumentation. In Proceedings from CSCL 2017: Making a difference: Prioritizing equity and access in CSCL, 12th international conference on Computer Supported Collaborative Learning (CSCL) 2017 (Vol. 1, pp. 135–142). Philadelphia, PA: International Society of the Learning Sciences.
Lin, F. R., Hsieh, L. S., & Chuang, F. T. (2009). Discovering genres of online discussion threads via text mining. Computers & Education, 52(2), 481–495.
Liu, O. L., Rios, J. A., Heilman, M., Gerard, L., & Linn, M. C. (2016). Validation of automated scoring of science assessments. Journal of Research in Science Teaching, 53(2), 215–233.
McNeill, K. L., Lizotte, D. J., Krajcik, J., & Marx, R. W. (2006). Supporting students’ construction of scientific explanations by fading scaffolds in instructional materials. The Journal of the Learning Sciences, 15(2), 153–191.
National Research Council. (2012). A framework for K-12 science education: Practices, crosscutting concepts, and core ideas. Washington, DC: National Academies Press.
NGSS Lead States. (2013). Next generation science standards: For states, by states. Washington, DC: National Academies Press.
Osborne, J., Erduran, S., & Simon, S. (2004). Enhancing the quality of argumentation in school science. Journal of Research in Science Teaching, 41(10), 994–1020.
Rosé, C., Wang, Y. C., Cui, Y., Arguello, J., Stegmann, K., Weinberger, A., et al. (2008). Analyzing collaborative learning processes automatically: Exploiting the advances of computational linguistics in computer-supported collaborative learning. International Journal of Computer-Supported Collaborative Learning, 3(3), 237–271.
Sadler, T. D., & Fowler, S. R. (2006). A threshold model of content knowledge transfer for socioscientific argumentation. Science Education, 90, 986–1004.
Sampson, V., & Clark, D. B. (2008). Assessment of the ways students generate arguments in science education: Current perspectives and recommendations for future directions. Science Education, 92, 447–472.
Sandoval, W. A. (2003). Conceptual and epistemic aspects of students’ scientific explanations. The Journal of the Learning Sciences, 12(1), 5–51.
Sandoval, W. A., & Millwood, K. A. (2005). The quality of students' use of evidence in written scientific explanations. Cognition and Instruction, 23(1), 23–55.
Shermis, M. D., & Burstein, J. C. (Eds.). (2003). Automated essay scoring: A cross-disciplinary perspective. New Jersey: Routledge.
Sievert, C., & Shirley, K. (2014). LDAvis: A method for visualizing and interpreting topics. Proceedings of the workshop on interactive language learning, visualization, and interfaces (pp. 63–70).
Simosi, M. (2003). Using Toulmin’s framework for the analysis of everyday argumentation: Some methodological considerations. Argumentation, 17, 185–202.
Southavilay, V., Yacef, K., Reimann, P., & Calvo, R. A. (2013). Analysis of collaborative writing processes using revision maps and probabilistic topic models. Proceedings of the 3rd international conference on Learning Analytics and Knowledge (LAK ’13), 8–12 April 2013, Leuven, Belgium (pp. 38–47). New York: ACM. https://doi.org/10.1145/2460296.2460307.
Staley, K. W. (2014). Experimental knowledge in the face of theoretical error. In M. Boumans, G. Hon, & A. C. Petersen (Eds.), Error and uncertainty in scientific practice: History and philosophy of technoscience (pp. 39–56). London: Routledge.
Stone, N. J. (2000). Exploring the relationship between calibration and self-regulated learning. Educational Psychology Review, 12(4), 437–475.
Tane, J., Schmitz, C., & Stumme, G. (2004). Semantic resource management for the web: An e-learning application. Proceedings of the WWW conference New York, USA, 2004 (pp. 1–10)
Tawfik, A. A., Law, V., Ge, X., Xing, W., & Kim, K. (2018). The effect of sustained vs. faded scaffolding on students’ argumentation in ill-structured problem solving. Computers in Human Behavior, 87, 436–449.
Toulmin, S. (1958). The uses of argument. New York: Cambridge University Press.
Tseng, Y. H., Chang, C. Y., Rundgren, S. N. C., & Rundgren, C. J. (2010). Mining concept maps from news stories for measuring civic scientific literacy in media. Computers & Education, 55(1), 165–177.
Yu, C. H., Jannasch-Pennell, A., & Digangi, S. (2011). The qualitative report compatibility between text mining and qualitative research in the perspectives of grounded theory, content analysis, and reliability recommended APA citation compatibility between text mining and qualitative research in the perspect. The Qualitative Report, 16(3), 730–744.
Walton, D., Reed, C., & Macagno, F. (2008). Argumentation schemes. New York: Cambridge University Press.
Xing, W., & Gao, F. (2018). Exploring the relationship between online discourse and commitment in Twitter professional learning communities. Computers & Education, 126, 388–398.
Xing, W., Popov, V., Zhu, G., Horwitz, P., & McIntyre, C. (2019a). The effects of transformative and non-transformative discourse on individual performance in collaborative-inquiry learning. Computers in Human Behavior, 98, 267–276.
Xing, W., Tang, H., & Pei, B. (2019b). Beyond positive and negative emotions: Looking into the role of achievement emotions in discussion forums of MOOCs. The Internet and Higher Education, 43, 100690.
Zhang, F., & Litman, D. (2015). Annotation and classification of argumentative writing revisions. Proceedings of the tenth workshop on innovative use of NLP for building educational applications (pp. 133–143).
Zhu, G., Xing, W., Costa, S., Scardamalia, M., & Pei, B. (2019). Exploring emotional and cognitive dynamics of knowledge building in grades 1 and 2. User Modeling and User-Adapted Interaction, 29(4), 789–820.
Acknowledgements
This work is supported by the National Science Foundation (NSF) of the United States under grant numbers DRL-1220756, and DRL-1418019. Any opinions, findings, and conclusions or recommendations expressed in this paper, however, are those of the authors and do not necessarily reflect the views of the NSF.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interests.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
In the scientific argumentation data about the albedo effect described above, each open-ended response a student generates is a text document. That is, each explanation response can include several topics. So does each uncertainty attribution response. Students’ explanation or uncertainty attribution responses are made up of topics that are made up of words. Therefore, LDA describes a document as a probability distribution of a mixture of topics, each of which is expressed with another probability distribution of words. Topics generated by LDA are a combination of words that contribute to the particular topic based on probabilities. LDA analysis results should be further interpreted by human insights about the context in which documents are generated. In this study, LDA is implemented in the following steps:
Step 1: Pre-processing and data preparation
To remove noise in the text data and format the data for input, pre-processing techniques are applied as follows:
-
1)
All non-letter symbols, numbers, and punctuation are removed.
-
2)
Common stop words such as “a,” “and,” “it,” and “the” are removed.
-
3)
Stemming is performed on the text to convert variations of the same word to a non-changing root word form. For instance, the root word “produc” captures several variations of the word like produced, producing, production, etc.
-
4)
Infrequent words are filtered out.
These steps result in a data corpus in the form of a bag-of-words that takes into account the occurrences of words, but does not consider their ordering. Each text document is represented by a document matrix defined as a vector of the words found in the entire corpus along with the frequency of each word found in the document. This is the input for the LDA algorithm.
Step 2: Determining the number of topics K
For the LDA algorithm to work, the number of topics to be extracted from the data corpus, K, needs to be specified. K can be determined by statistical derivations or informed by researchers’ insights about the documents. An optimum value of K can be determined through the Bayesian model selection and approximated using a harmonic mean estimator. The log-likelihood plots can show the best K value for the text corpus. Note that this statistically derived K value is not the absolute measure for K. Expert judgment based on data, knowledge, and experience can be important in selecting the most meaningful K value. In this study, we combine the log-likelihood method with human judgment to determine the K value for the scientific argument data corpus. Different K values were explored before determining the optimal number. Given the K value, LDA generates a list of relevant words for each topic (topical words) and which topics are contained in each document.
Step 3: Application of the LDA algorithm
Collapsed Gibbs sampling is applied as follows:
-
1)
Each word in the corpus is randomly assigned to the K number of topics. Each topic now constitutes an initial random word distribution based on Dirichlet, which will be iteratively improved in the following steps.
-
2)
For each word in a document,
-
•
Compute the proportion of words assigned to a topic in the document, P(topic|document), and the proportion of words assigned to that topic from all documents, P(word|topic).
-
•
Reassign the word to a new topic with the probability of P(topic|document) * P(word|topic).
-
3)
Repeat Step 2 numerous times until the topic-word assignments are stabilized.
-
4)
Use the topic assignments to calculate the proportion of topics in each document.
Distinctive words that appear in a topic and do not appear in other topics can be very useful to characterize the topic. If all the documents contain similar words, it is harder to cluster the words into topics, requiring expert evaluation.
Rights and permissions
About this article
Cite this article
Xing, W., Lee, HS. & Shibani, A. Identifying patterns in students’ scientific argumentation: content analysis through text mining using Latent Dirichlet Allocation. Education Tech Research Dev 68, 2185–2214 (2020). https://doi.org/10.1007/s11423-020-09761-w
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11423-020-09761-w