I thin k that the self-assessment of the students can be ONE measure, but students often think that their skills are better than they actually are.  I think that you need some kind of independent assessment, such as a pre- and post-test methodology.  

I have attached a rubric that I have used to evaluate oral proficiency.  I think that a similar rubric could be developed for other categories of language skills.   Then have more than one person do the evaluation and determine inter-coder reliability.

Data Oral Proficiency Rubric

When working with L2 learners at universitary level, in my experiencie, most troubles will not be grammar mistakes, but rather problems of in-adequacy in the use of idioms. Also, I would expect most troubes to emerge in non-technical language. It is not the discipline-specific use of the language that entails a special difficulty for L2 learners (after all, this is a new language for all new learners of the discipline) but rather the uses of language that reflect a refined knowledge of its uses and mecanisms. (I would venture here to say that a non-native will probably have more difficulties with the use of metadiscourse). Self-evaluation, in this case, is not very effective; using grammar manuals and checking for common mistakes simply will not do. O the other hand, working with peer-reviewers, even non-language specialists, can give an L2 learner many clues about "what sounds odd" or "you just don't use this word in that context".     

You might want to try the Multiple Intelligences Adult Literacy and Education assessment.. Also, North Carolina State Univ. has a modality assessment: http://www.engr.ncsu.edu/learningstyles/ilsweb.html -- hope this helps.

I do agree with Michael, some students have the tendency to over-evaluate their ability, therefore, using a pre and post test approach could produce a more meaningful data than self-evaluation and reflective diaries - you might find it difficult to quantify it as well if you're trying to capture student's language development. Alternatively, you could use Learner's Log, for example you could design it in a form of a table with all the essential skills with specific time frame relevant to their language course (based on existing curriculum), in which students then tick against the column to indicate whether they've acquired the skills or not, then followed by in-depth interviews. You might find it more useful too, if your research is longitudinal, since you're interested in capturing student's development? Hope this helps (:

I think you need to administer IELTS to your students to determine their academic language skills. Most of the language in content areas is discipline specific and this will require a large scientific vocabulary from your students. Orientate them into science language and scientific texts are structured.

Dear Benyoucef,

You may develop your own test of English as Academic Purposes based on the special environment in your area. Please, consider the four English skills as the main part of the test. Peer assessment can also be carried out as part of the test.