Volume: 7, Issue: 1

1/03/2015

Statistical Analysis of the Construct “Creative Self-Efficacy”
Маслак А.A. [about]

KEYWORDS: Creative Self-Efficacy, measurement of latent variables, construct analysis.

ABSTRACT: This research demonstrates the applicability of Rasch models for measurement of the construct “creative self-efficacy” and all its aspects on a common interval scale. All aspects of creative thinking self-efficacy (fluency, flexibility, elaboration, and originality), creative performance self-efficacy (domain, field, and personality), Beghetto's Creative Self-Efficacy (BCSE), and openness to experience are positively correlated that confirms the construct validity of latent variable creative self-efficacy. The measurement properties of indicators after consolidation appear adequate for practical applications with policy and implications. The questionnaires have a good enough discriminative ability of students depending on their gender and chosen major at the university level.


Literature review

Creativity has proved to be a very important competitive advantage of the human mind. The rise of computers and Internet mean that many forms of work are gone, and others are being destroyed. People can no longer monopolize logical thinking in such domains as science, technology, engineering, and mathematics. Individuals without creativity will be left behind in today’s world where non-creative thinking can be automated, but creative solutions to problems are at a premium. Modern research of creativity goes back to a famous Guilford's APA Presidential address (1950). 

Although studies of creativity have been commonplace for decades, but research which would specifically address the internal structure of creativity constructs is still rare (Beghetto, 2013; Mathisen & Bronnick, 2009;  Starko, 2013; Yang & Cheng, 2009).

Goals and objectives

The goal of this research is to measure all structured aspects of creative self-efficacy and find out statistical relationships between them in accordance with Abbott (2010). The research objective is to analyze these measures across students’ university majors and genders. There are two dimensions of creative self-efficacy: creative thinking self-efficacy (CTSE) and creative performance self-efficacy (CPSE). CTSE includes four hypothesized aspects:  fluency, flexibility, elaboration, and originality while CPSE includes three other hypothesized aspects: domain, field, and personality. We also look into Beghetto's Creative Self-Efficacy (BCSE) and openness to experience. All these aspects are defined operationally, i.e. through a set of indicators (items of questionnaires). We conduct our analysis of the questionnaires for estimation of these aspects based on the framework of a classical test theory by Abbott (2010).

To reach these goals we addressed the following objectives:

  1. To assess the degree of compatibility of the questionnaire items with each other within each aspect.
  2. To measure the level of all aspects of students’ creative self-efficacy.
  3. To obtain statistical relationships between different aspects of creativity.
  4. To analyze the level of students’ creativity (defined by aspects) across majors and genders.

Theoretical framework

As a theoretical framework we have used a theory of latent variables based on Rasch models (Rasch, 1980), and works by Wright & Masters (1982). The main advantage of Rasch’s theory consists in obtaining measures of latent variables on interval scales. The significance and originality of the research lies in obtaining results of measurement of students’ creative self-efficacy (in all of its aspects) on an interval scale. These measures provide basis for monitoring students’ creativity. It is worth mentioning also that the results of analyses of students’ creative self-efficacy vary depending on their major and gender.

For survey data processing, the interactive software system RUMM2020 (Rasch Unidimensional Measurement Models) was used (Andrich, Sheridan, Luo, 2005).

Method

For each indicator (for all aspects) students rate their degree of confidence on the Likert scale by numbers from 0 to 4. Here ‘0’ means not at all confident, and ‘4’ means highly certain in the ability to fulfill the task. Measures on the Likert scale were transformed to linear measures with a Rasch model for rating scales, which computes a log-odds transformation of indicators and students, then computes differences between indicators and students also guided by the one-parameter logistic function to establish a common dimension.  A rating scale model is implemented for this transformation.

Data

Students from all schools of the Branch of Kuban State University in Slavyansk-on-Kuban participated in this survey. The sample of students used in this research equals N = 148. Among them were 30 male and 104 female, 14 students chose not to indicate their gender.

Results

The results show that indicators of questionnaires for all aspects of creative self-efficacy are well targeted on students without ceiling or floor effects.  Indicators (items), in general, appear to distribute well across the latent trait with a concentration at mid-scale. Aspects on the creative self-efficacy scale vary from .600 logits to .416 logits and are located in the following order: domain, fluency, BCSE, flexibility, elaboration, openness, personality, field, and originality (Table 1).

Table 1. Location of aspects of creative self-efficacy on the common scale

Aspects of Creative Self-Efficacy Location (logit) Standard Error Gender Department
1. Fluency -0,502 0,109 Non significant Significant
2. Flexibility -0,172 0,104 Non significant Non significant
3. Elaboration 0,141 0,099 Significant Significant
4. Originality 0,416 0,094 Non significant Significant
5. Domain -0,600 0,111 Non significant Non significant
6. Field 0,350 0,096 Non significant Significant
7. Personality 0,280 0,096 Non significant Non significant
8. Beghetto's Creative Self-Efficacy -0,254 0,104 Non significant Non significant
9. Openness to Experience 0,277 0,096 Non significant Non significant
10. Creative Self-Efficacy as a whole 0,634 0,621 Non significant Significant

So the aspect domain is best for discriminating students with a low level of creative self-efficacy (-.600 logits) and originality is the best for discriminating students with a high level of creative self-efficacy (.416 logits).

All 9 aspects of creative self-efficacy are positively correlated with each other (Table 2).

There is no significant difference between men and women in all aspects except ‘elaboration’. In this aspect women are more elaborative than men: .714 logits and .168 logits, respectively. There is a significant difference between students from different schools on scales: fluency, elaboration, originality, and field. So the questionnaires have a good enough discrimination ability.

Table 2. Correlation coefficients of creative self-efficacy aspects

Creative self-efficacy aspects 1 2 3 4 5 6 7 8 9 10
1. Fluency 1                  
2. Flexibility 0,44 1                
3. Elaboration 0,34 0,47 1              
4. Originality 0,28 0,34 0,43 1            
5. Domain 0,60 0,38 0,25 0,20 1          
6. Field 0,37 0,41 0,35 0,45 0,40 1        
7. Personality 0,25 0,21 0,10 0,25 0,21 0,55 1      
8. Beghetto's Creative Self-Efficacy 0,46 0,39 0,21 0,35 0,30 0,38 0,55 1    
9. Openness to Experience 0,24 0,30 0,19 0,17 0,13 0,08 0,08 0,38 1  
10. Creative Self-Efficacy as a whole 0,69 0,69 0,60 0,65 0,60 0,70 0,54 0,69 0,46 1

Table 2 shows that creative self-efficacy as a latent variable positively correlates with all its aspects too, but best of all it correlates with fluency, flexibility, field, and BCSE (r varies from 0.60 to 0.70); in a lesser degree it correlates with openness and personality (r varies from 0.46 to 0.54).

Discussion

This research demonstrates the applicability of Rasch models for measuring a latent variable creative self-efficacy by constructing a common scale. The same is true for all aspects of creative self-efficacy. The results suggest that the traditional method of comparing students with separate indicators may not be as advantageous as it does not allow using all available information.  When questionnaire indicators are consolidated into a coherent latent trait, they allow comparisons of creative self-efficacy (and all its aspects) of the students depending on their major and gender. The measurement properties of indicators after consolidation appear adequate for practical applications with policy and implications. Another important advantage is the overall perspective that consolidation provides when indicators and students are brought into a common quantitative framework. As a consequence, we can have a more powerful analysis than isolated, separate indicators without sacrificing information about separate indicators could provide. In addition, the consolidation of indicators may have important educational development implications because the latent trait framework shows which indicators and aspects are more important for development of students’ creative self-efficacy.


References

  • Abbott, D.H. (2010). Constructing a creative self-efficacy inventory: a mixed methods inquiry. Doctoral dissertation, University of Nebraska.
  • Andrich, D., Sheridan, B., Luo, G. (2005). RUMM2020: Rasch Unidimensional Measurement Models software and manual.  Perth, Australia, RUMM Laboratory.
  • Beghetto, R.A. (2013). Killing Ideas Softly? The Promise and Perils of Creativity in the Classroom. Information Age Publishing Inc.
  • Guilford, J.P. (1950). Creativity. American Psychologist, 5(9), 444-454.
  • Mathisen, G.E. & Bronnick, K.S. (2009). Creative self-efficacy: An intervention study. International Journal of Educational Research, 48, 21-29.
  • Rasch, G. (1980). Probabilistic models for some intelligence and attainment tests (Expanded edition, with foreword and afterword by Benjamin D. Wright), Chicago: University of Chicago Press.
  • Starko, A.J. (2014). Creativity in the Classroom: Schools of curious delight. 5th edition. Lawrence Erlbaum Associates, Inc., Mahwah, New Jersey.
  • Wright, B.D., Masters, G.N. (1982). Rating Scale Analysis, Chicago, MESA Press.
  • Yang, Y.-L. & Cheng, H-Y. (2009). Creative self-efficacy and its factors: An empirical study of information systems analysts and programmers. Computers in Human Behavior, 25(2), 429-438.

Home | Copyright © 2024, Russian-American Education Forum