Return-Path: <nifl-esl@literacy.nifl.gov> Received: from literacy (localhost [127.0.0.1]) by literacy.nifl.gov (8.9.3/8.9.0.Beta5/980425bjb) with SMTP id RAA23654; Fri, 10 Mar 2000 17:31:14 -0500 (EST) Date: Fri, 10 Mar 2000 17:31:14 -0500 (EST) Message-Id: <s8c9141c.097@epcc.edu> Errors-To: listowner@literacy.nifl.gov Reply-To: nifl-esl@literacy.nifl.gov Originator: nifl-esl@literacy.nifl.gov Sender: nifl-esl@literacy.nifl.gov Precedence: bulk From: "Andres Muro" <AndresM@epcc.edu> To: Multiple recipients of list <nifl-esl@literacy.nifl.gov> Subject: [NIFL-ESL:4228] Re: using TABE to assess advanced ESOL -Reply X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=US-ASCII X-Mailer: Novell GroupWise 5.5.3 Status: OR When I was in high school, I took some chemistry classes, some biology classes, some history classes and so on and so forth. In college I took Plato, Modern Philosophy, Intro to linguistics, etc. etc. At conclusion of all these courses, my teachers gave me a grade that showed if I passed or failed the class. After that, nobody gave me a standardized test to determine if I really had understood Plato according to what the state interpreted my understanding of Plato should have been. The higher the level of the course, the more subjective the interpretation of the teacher. In those times, there was trust placed in the teacher being able to determine wether or not the students had learned the subject at hand. I realize that there are a lot of gaps that we need to address, and a purely subjective method is not necessary the answer. However, I think that teachers need to have both the background, ability and trust to discover together with the students if there have been improvements in language acquisition in contexts, and the utility of that knowledge in other contexts. I place a lot of value in the subjective interpretations of our staff. For example, when I inquire about a student, I may ask one of our educators something like "do you think that s/he is writing, reading, conversing better, using health services, applied for admissions to college, etc, etc". In fact, I ask that they report this at the end of a class. Andres >>> kathleenb@epcc.edu 03/10 2:47 pm >>> Tests are based on samplings of behaviors, nothing more. There are two types of reliability: rater reliability (would all raters rate the item (behavior) the same way? would the same rater rate the item (behavior) the same way when presented with it again?) and subject reliability: if the subject is given the same item (sample behavior) will the subject respond the same way? Any test instrument must meet these requirements for reliability. It is easier to achieve reliability with a paper and pencil test, harder to do with a scoring rubric, and almost impossible with a Likert scale. These are separate issues from validity, which simply means that we are measuring what we think we are measuring. All testing and assessment instruments are judged by these two criteria. Again, most of our problems come when we make assumptions about the validity of certain instruments to give us the information we want rather than the information they are designed to give. The TABE will not tell me about your ability to conduct a job interview in English. The BEST won't tell me about your ability to perform graduate level work. It is also difficult to develop reliable measurements of oral proficiency in a natural environment. Any testing environment is by definition artificial. Natural observation comes closer, but is expensive and impractical for programs. So we are left looking for the best possible approximations of what we want to know to measure something. We went through this process in designing the domestic violence survey. Kathleen Bombach
This archive was generated by hypermail 2b30 : Tue Jan 16 2001 - 14:44:47 EST