As I mentioned earlier, with children so young, many issues impact the reliability and validity of the exam. We cannot entirely solve these difficulties, but we can take some measures to make our assessment as reliable and valid as possible under the circumstances.
An exam is considered to have construct validity when it measures only what it is supposed to. In our case, this means that I will read aloud the exam questions and directions so that students are not penalised for difficulty in reading and understanding them independently, as this is not a skill that I intend to measure. Also, the images used must be clear and easy to interpret.
Content validity means that the exam contains a representative sample of what it is to measure. To this end, we can create a specification of skills that the exam is meant to cover, and in this case the specification would be closely related to the assessment criteria.
To increase reliability of the exam, it is helpful if the students are familiar with the exam format. For this reason, the exam tasks are very similar to activities we have done in class. I will also make sure scoring is objective by not including ambiguous items. Since I create my own exam, it will be helpful to have some of my colleagues look at it first to see if they can spot any ambiguities. In this case, these might include foods that could arguably belong to more than one category (if they are included, it should be specified in the exam key that both answers are acceptable.)
An exam is more reliable when it has a greater number of items, rather than just a few. However, with such young learners, having a lot of items on the test would lead them to get fatigued, bored, etc. and would lower the overall reliability of the exam rather than raising it. It is a difficult balance, but we can compensate to some degree by taking into consideration other evidence of student performance (eg. class activities) to complement the information gained from the exam.
No hay comentarios:
Publicar un comentario