A typical task in language schools is to assign a level of skill to the students with respect to the language they are going to study. For this, the students are evaluated through a set of level tests that measure the student's competence. Normally, the assessment consists of an oral competency test, a written competency test, and a listening comprehension test. A characteristic of this type of test is the possibility of evaluating them remotely (usually they are tests with questions where every question has associated a set of possible answers where only one of the answers is correct) without the necessity of the presence of any evaluator. For this reason, it is possible to create computer applications that automate the evaluation and management of test results. This chapter presents a computer application that has been created with the aim of automating the process of management and evaluation of level tests for students of Spanish as a foreign language. The application has been tested at the University of Zaragoza.
TopIntroduction
This article is contextualized in language schools that exist in universities. In this sense, many universities have language training services that offer students and university staff the opportunity to learn a language (note that this service is very useful and used in the case of foreign students who visit a university and who must learn the native language of the country).
To perform in one of these language courses, students must be leveled by performing a set of tests (Kumaravadivelu, 1994) that allow the school to determine the level of linguistic competence that the student has with respect to the language to be studied. In general, three types of tests are performed (Finocchiaro, 1983). These are an oral comprehension test (McKay, 2002) that measures the ability to understand and be understood by a native person, a test of listening comprehension that measures the student's ability to understand conversations between native people (Rivers 1981) and a written comprehension test (Hadley, 1993) that measures the domain of vocabulary and grammatical constructs in language. Sometimes a fourth test of reading comprehension is also performed which is independent of the written comprehension test and consists of reading a text and then answering questions about the text read.
Usually the oral test (Lado, 1961) is performed in the presence of a teacher with whom a topic is discussed. In the auditory test (Pinto-Llorente, 2016), schools use a multimedia resource such as a video, a song or a recorded conversation that the student must listen to and after which questions are asked to see what he has understood about the audition. Finally, the written test (Nuttall, 1996) usually consists of exercises such as writing an essay on a proposed topic, rearranging the components of a sentence so that the sentence is correct or completing parts of a sentence that are empty with the more appropriate words (Pinto-Llorente, 2014).
Level tests can be automated through some computer applications. In this sense, there are several ways of automating them (Yang, 2009) that can be classified in: specific evaluation tools or general evaluation tools (Squires, 1997).
The specific evaluation tools (Voorheis, 2004) evaluate only one specific language. Its functionality is oriented to the particular characteristics of the language to be evaluated. In most cases, these tools are not free, being necessary to pay some kind of license to use them. The main advantage they present (Gottliebson, 2010) is the specificity with respect to the particularities of the language to be evaluated. However, they present a disadvantage, the impossibility of reusing them to evaluate other languages (Pinto-Llorente, 2015). Likewise, in many cases, they have little possibility of adaptation and configuration (usually it is software that cannot be modified by users, adaptations of their interfaces cannot be made, and they offer a very poor set of user management services).