Abstract:

The main topic addressed in the article is the issue of measuring the success of reading comprehension in a second foreign language - German and French. Reading comprehension can be characterized as a complex process that depends not only on the foreign language competence of students, but also on their cognitive, social and personality variables. In this context, the research team has developed an intervention reading comprehension program in recent years, which appropriately supports and optimizes the different aspects of reading comprehension. The intervention program is intended for upper secondary school students achieving the reference level A2 in a second foreign language. In the near future, an experimental verification of the intervention program and its efficiency in school practice will be carried out. For this purpose (experimental verification of the program), a measurement tool in the form of a language test has been developed and piloted, which needs to be validated and its quality increased so that it can measure the success of the intervention program. That is why in the paper we will focus on the research problem associated with increasing the reliability of the measurement tool designed for the experiment. The research goal is to determine the tool reliability and validate the tool for experimental verification of the reading comprehension intervention program. Subsequently, corrections leading to an increase in the quality of this tool will be introduced. The subject of the performed analyzes are language tests measuring the level of reading comprehension: a test in German and a test in French language.

The methodology is based on the approach of Classical Test Theory, which is mainly associated with the concept of reliability, we point out two possible ways to increase this psychometric indicator. The first of these is the use of the Spearman-Brown correction formula, where the increase in reliability is achieved by a reasonable increase in the number of items in the test. The second way to increase reliability is to identify problematic items in the test and their adequate replacement, or correction based on the item analysis of the test.

The results showed that the measured value of reliability did not reach such a level that would be sufficient for us to be able to unambiguously confirm or exclude the efficiency of the intervention program within the experiment. Both tests had comparable (almost identical) reliability. However, the overall as well as item analysis of both tests revealed some significant differences in psychometric indicators, which were mainly related to the difficulty of the test and its items, functioning of answer keys (correct answer) and distractors (incorrect MCI answers), minimum, maximum and average success and actual student score. These different test parameters will allow us to apply both methods of increasing the tool reliability. The contribution of the paper to the educational process lies primarily in the explicitly formulated methodological procedures usable in pedagogical research, which is based on the creation of quality measurement tools with high reliability. Validation, or increasing the quality of tools for measuring the efficiency of pedagogical intervention (intervention program) is necessary as it is difficult to obtain valid data that demonstrate the success of the program and thus to the successful promotion of reading comprehension.

Keywords:

Reading comprehension, intervention program, classical test theory, reliability, measurement, test, item analysis.