Shohamy E. and Brown 23 and 24

There is so many components that go into assessment especially testing.  The test must be reliable and valid, in accordance to what the test should be measuring.  This goes through domains such as reading, writing, listening, speaking, vocabulary, and grammar; it aslo goes through purpose such as stakeholders of the test, test users, test takers, is it measuring achievement or is it placing the student into a program.  Something that the article concentrates and stresses is the importance of analyzing discourse when testing.  Shohamy gives a short background of research that has been done on discourse and language testing, "quantitative analyses have examined mainly the effect of different discourse elements on test takers' scores" (203).  There is research on the effects of test features such as subject matter, titles, rhetorical structure, missing cohesive ties, question types, oral tests and production tests and how these features effect test scores.  For example, for a reading comprehension test, the question types must take into consideration, "(1) the texts used for a test, (2) the test questions, (3) the answers produced in response to the test questions, and (4) the interpretation of the answers produced and the assignment of scores" (205).  The question types suggested by Perkins in the article are open-ended essay responses that can be collected for a portfolio.  This is similar to the teaching writing article that we read earlier.  In that article they recommend alternative assessment. 
I thought the section on qualitative examination of oral proficiency tests was particularly interesting, perhaps because I will take an partly oral examination to be a bilingual teacher.  On the state test that I will take there will be an oral section in which I have a limited time to speak into a tape recorder.  For a taped test, there is typically more paraphrasing, and for a direct test like the OPI there is more switching to their L1.  I would have thought the opposite, just because there is an actual person that you are speaking in a direct context. 
The last section on new assessment types mentions that "teachers and students are more involved in developing and designing the assessment procedures" (212).  I like the joint development of the test, it is a very transformative way of thinking.  Also there must be a variety of assessment procedures for assessment of discourse skills, which cannot be totally analyzed by a state test.