I'm looking for a set of free and (somewhat) easy tools that I can use with my EFL writing students next semester. I want to analyze their initial essays and look for common errors that can be addressed back in the classroom (hence, the DDL - data-driven learning). I had initially thought of using AntConc to search for certain words (i.e. "should", and then transfer it over to a spreadsheet for categorization of grammar or errors (o, but I'm not sure this is the best method for doing it. Are there any other methods you could suggest?
I would definitely recommend using AntConc (download here) for the purpose you have described because AntConc is free, you are indeed able to copy and past concordance lines into a spreadsheet, and there are many tutorials and publications available about specifically using AntConc for language learning.
I would highly recommend this book which was published recently:
Liu, D., & Lei, L. (2017). Using Corpora for Language Learning and Teaching. TESOL International Association. Link to Publisher's Page
I would also like to point out the the differences between "Indirect" and "Direct" Uses of Corpora in Language Learning and Teaching (discussed in Liu and Lei p. 5) because your question seems to suggest you are looking for direct uses . Direct use of language corpora in language teaching would be having student's use the corpus themselves, that is, the students would be preforming queries, analyzing concordance lines, generating frequency lists, etc. I think this can be very helpful for advanced and technically savvy students as corpora can act as native speaker informants available 24/7 (as mentioned by Römer 2011: p.215, cited in Liu and Lei: p. 6). Indirect use of corpora refers to anytime corpora are used in the preparation and planning of pedagogical materials for language learning (for example, the teacher creates vocab lists before class).
However, learning to use a corpora can be cumbersome for language learners, and, as you suggest doing error-analysis of learner corpora might be challenging. I would recommend you consider using indirect methods L2 error analysis, you can find common errors and prepare materials for the students themselves.
There is a lot of thought that goes in the design of a learner corpus, and the tools follow from the design, not the other way around.
Some things to consider:
- The corpus should be sharable among researchers, therefore you need the consent of the learners to publish the corpus (anonymised, of course!)
- You want to collect some metadata on the learners (at what level are the learners, maybe measure learning success, sociological data, other learning experiences)
- You need a scheme for annotating errors and good solutions (don't forget about them!) that you and potential co-annotators/co-correctors can apply consistently.
- Think of research questions to be answered by your corpus and how to answer them. Are your data and metadata sufficient?
Fortunately, you are not the first one to prepare a learner corpus. You can take advantage from other people and groups who have built learner corpora before you, here is a reference Andrea Wurm, Presentation of the KOPTE Corpus and Research Project