Abstract:

Creativity, more than ever, is a major competitive advantage in today’s world. The aim of this investigation is to analyze the quality of a questionnaire as a measuring tool for evaluating students’ creative self-efficacy. Creativity is a latent variable, so it is evaluated based on the questionnaire. Significant disadvantages of many methods of estimating latent variables (weighing method, expert estimates, etc.) are the subjectivity of expert weights and the non linear scale. This makes it difficult to apply statistical methods of analysis which use a linear scale of measurement.
Therefore, the evaluation of latent variables is carried out within the framework of the theory of latent variables, which allows to measure students’ creative self-efficacy on a linear scale. The latent variable "creative self-efficacy" is defined operationally by a set of indicators (questionnaire items), each of which characterizes one of the aspects of creativity. It is essential to analyze the quality of the questionnaire as a measuring tool, as this affects the objectivity of the results.

The following aspects of the questionnaire as a means of measurement are considered:
- Adequacy of the set of indicators for the measurement model (Rasch model).
- The presence of extreme indicators in the questionnaire.
- Targeting the set of indicators to the subjects (targeting the average value of indicators to the average value of the measured objects).
- The range of variation of measured objects and indicators on the linear scale of the latent variable.
- Analysis of distinctive indicators (the “easiest”, the "most difficult", the most "adequate", the least "adequate"). The quality analysis of the questionnaire was carried out based on the results of the survey of students of Kuban State University. The statistical analysis showed the following outcomes. The results of the survey are well described by the Rasch model, which indicates the compatibility of the set of indicators. No extreme indicators were found. The person separation index was quite high (0.88), which indicates a high degree of differentiation of students on the measured latent variable creative self-efficacy. The range of variation of students' creative self-sufficiency assessments is quite large and equals 5.0 logit. This suggests that students vary greatly in their creative self-sufficiency. In addition, as expected, the distribution of students’ estimates is close to normal. Indicators vary in a slightly larger range (5.5 logit) and completely cover the range of variation of students' estimates. This ensures high measurement accuracy over the entire range of variation of the latent variable. Between the students’ estimates and indicators’ estimates, there is a small shift – the difference between the respective averages is equal to 0.681 logit. This means that the used set of indicators is informative for measuring creative self-sufficiency. Distinctive indicators (the “easiest”, the most “difficult”, the most "adequate", the least "adequate") are also taken into consideration.
In general, the questionnaire can be used as a measuring tool for evaluating students’ creative self-efficacy.

Keywords:

creativity, latent variable, measurement, Rasch model, questionnaire quality