| KALT CRITERIAL |
Suppose that the students of the school of airplane pilots had a 30% punctuation mean, with a deviation of 10%. The school decides to score using the "curve", that will approve the students that obtained scores above of the mean plus one standard deviation. The best student approved by the school obtained a 45% of ability level. Would you accept to get into the airplane with this pilot? What can be said about the quality of the school of pilots?
When it should not be scored in population's terms, but in terms of minimum ability requirements, the suitable tool is KALT CRITERIAL®, a system of questionnaire scoring referred to a criterion; unique program in its kind that combines the experience of KALT® and the support of the models offered by international investigators of different knowledge areas and according to the most recent publications.
 |
In KALT CRITERIAL® the characteristics of the test are defined by reasoning in terms of value of diverse ability levels required to guarantee that the person is qualified to carry out a specific work, to be promoted to the next scholar level, to have a minimum acceptable level of quality. In this case the ability levels are compared with the scale proposed by the evaluator in a basis of numerical scales (0-10) or literal scales (insufficient, enough, good). The program allows to make the comparison between a person and the established ability levels in a questionnaire and it emits group reports when it is an implementation to a numerous group of people. With the help of KALT CRITERIAL® you can analyze the questionnaire and a report of quality is obtained, with which the evaluator will be able to feedback his work, to correct the deficiencies and to debug the instrument.
|
|
Scoring referred to a criterion not only takes into consideration the number of items (how many) but the form in that the candidate answers the test (which), with an unique model of KALT CRITERIAL® that has the elements for the determination of the error of each person's scoring, in absolute terms and in function of the difficulty of the instrument, facilitating to the evaluator the task of assigning the level of each person's ability. |
 |
The system is developed for the Windows environment (95, 98 or NT), with the support of the most modern programming languages and the best performance at the present time.
The program allows to generate reports of results with very diverse presentation options and data classification, offering possibilities of organizing the information, creation of diverse types of output files, selective printing, etc. It also includes options to create global and individual reports, database creation and diverse printing capabilities.
Reports that KALT CRITERIAL® generates are:
 |
Individual and global. Very detailed for giving feedback to the candidate in each one of the evaluated aspects. Several options to emit the global verdict.
|
.
|
|
|
Item analysis. Including level of difficulty, point-biserial correlation, reliability, modified number of Nedelsky, report of quality, charts of expectatives, measuring error, etc.
|
|
|
|
Technical verdict. It includes a complete analysis of reliability, an estimate of the optimal instrument size, analysis and validation of the cutting point using a "method of commitment" improved by an exclusive model of KALT CRITERIAL®.
|
|
|
|
Statistical. As complete as those of KALT PLUS®, including diagrams of frequencies and curved of accumulated frequencies, global statistical parameters and of every topic, from the mean and the standard deviation to information of the skew and the kurtosis of the distribution of the answers. |
KALT CRITERIAL® stores all and each one of the steps carried out in the calculations, so that you can verify that the results are correctly calculated, besides allowing to use them directly as interface with any database program (Dbase, Access, FoxPro), spreadsheets (Excel, Quattro) or text editors (Word, Edit, etc.)
SOFTWARE FOR GRADING CRITERION REFERENCED TESTS AND MODELS FOR THE DEFINITION OF STANDARDS
"Criterion references tests are widely used actually, mainly related to certification of professional competencies. New specifications for test design as well as for test scoring and analysis are needed; in order to help test designers a comprehensive model and software have been developed. The model for analysis and definition of standards includes several options coming from technical papers, plus new formulations developed specially for the software. The cornerstones are a logistic two-parameter model for item analysis and a quality report of the test. The program has been used by the National Evaluation Center of Mexico (Ceneval) since 1988, scoring and analyzing more than 1.5 million undergraduate and graduate students."