Abstract :
Objective: By using the scales used in the field of health with this study, it is to examine whether or not the regression tree method can be an alternative approach to the computer adaptive test applications based on psychometric model used to determine the level of examined features. Methods: The predictive performances of Computer Adaptive Test and Regression Tree methods were evaluated on artificial data set derived from 100 people’s dichotom responses (e.g. yes/no) to 100 items. Evaluations were made in terms of estimation values, item numbers and error values. R v.3.6.3 program is used in applications. Results: While the error value was 2.45 in the Computer Adaptive Test method, it was 4.04 in the regression tree method. While the Computer Adaptive test approach predicted with an average of 42 items (minimum:41-maximum43), the regression tree estimated with 23 items. Conclusion: In terms of the compared values, computer adaptive test approach and regression tree methods gave similar results. It has been seen that, as well as the regression trees method can be used as an alternative to the computer adaptive test approach for the data set we use, advanced tree algorithms should be examined further.
