Education
is a very important element in developing the potential of every human being.
In the world of education, learning consists of three stages, namely planning,
implementation, and evaluation. Evaluation aims to identify the efforts that
have been made in the learning process that are carried out well or not. To
achieve learning objectives, the evaluation process needs to be carried out
systematically and continuously in order to find out good learning evaluation
procedures. Questions as an evaluation tool are said to have good quality if
they have validity, reliability. So in the Introduction to the Schooling
Environment (PLP) an analysis of the items is carried out which aims to
determine the validity, reliability, index of item difficulty, and index of
discriminating power.
This
research is classified as ex post facto research with a quantitative
descriptive approach using the Rasch model to analyze validity, reliability,
item difficulty index, and discriminatory index. The research data were
obtained from the results of the Middle Semester Assessment (PTS) class XI
Automotive Management and Maintenance Engineering (TMPO) 1 and 2 which
consisted of 15 multiple choice questions. The results of the answers from the
students if they are correct then they get a value of 1 and if they are wrong
they get a value of 0. The research data was analyzed using Winstep 3.08.1
software.
Research
results from the validator test show that the instrument items are very valid.
Meanwhile, from the results of the winstep analysis, the logit value of the
person or measure is 0.13 and the item measure is 0, which means that the
person measure is greater than the item measure. It can be stated that
students' abilities tend to be higher than the difficulty level of the
questions. Meanwhile item reliability was 0.94 and person reliability was 0.49
and Alpha Cronbach was 0.81. from this value it can be concluded that the level
of consistency of student answers is quite high and the quality of the item
items has a reliability value of 0.94. In the misfit order item table, there
are 8 items that do not meet the correlation requirements so that it is
possible to have items that students might easily guess because there are
similar characteristics of the items. Item number 2 has a high index of
difficulty so that it is possible for students to have language that is not
understood by students, making it difficult to answer correctly. While item
number 10 has the easiest index of difficulty, it is possible that the
characteristics of the questions are easy to guess because there are clues to
the correct answer. So that many students answered correctly. From the graph of
the information function, it shows that the question instrument is suitable or
optimal for use by students who have moderate abilities.
Keywords: item analysis, learing evaluation, rasch model.