Portal de Periódicos da CAPES

A Rasch Measure of Fostering Creativity

2010; Taylor & Francis; Volume: 22; Issue: 2 Linguagem: Inglês

10.1080/10400419.2010.481534

1532-6934

Lemmy K. C. Teo, Russell Waugh,

Career Development and Diversity

Abstract A general creativity questionnaire of 21 stem-items was designed and based on 7 teaching aspects of fostering creativity in higher education students: (a) teaching the skills and attitudes of creativity; (b) teaching the creative methods of the disciplines; (c) developing a problem-friendly classroom; (d) using prior knowledge; (e) using different types of problems; (f) using multi-disciplinary hands-on projects; and (g) using appropriate lesson plans. Stem-items (21) based on these were conceptualized from easy to hard—3 stem-items for each of the 7 aspects—and answered in 2 perspectives: (a) an attitude self-view (ideally this is what I think I should do) and (b) a behavior self-view (this is what I actually do)—using 3 ordered response categories: none or some of the time (score 1), most of the time (score 2), and all of the time (score 3). This meant that the effective item sample was 42. The general questionnaire was applied to a lecturer sample from a higher institution, N = 124. Data were analyzed with a Rasch measurement model computer program. The item-trait interaction chi-square was not statistically significant (χ2 = 87.6, df = 84, p = 0.37), meaning that a unidimensional trait (called fostering creativity) was measured in which the 42 items each fitted the measurement model with p > 0.02. The Person Separation Index was 0.78 and the Cronbach Alpha was 0.79 showing that the measures were well-separated along the scale in comparison to the errors. Notes Notes. The item means are constrained to zero by the measurement model. The means of both item and person fit statistics are 0.12 and −0.15. The standard deviations of both item and person fit statistics are 0.89 and 1.10. As the item and person fit statistics have means near zero, and standard deviations near one, they indicate that the data fit the Rasch Measurement Model very well. The standard errors are to two decimal places, so numbers are given to two decimal places. Note. The difficulty values of each method are the mean of three supporting items in each aspect. Notes. Location = item difficulties, SE = standard error; Residual = responses predicted from the Rasch measurement parameters and responses provided by the lecturers, df = degrees of freedom, Probability = calculated chance of such occurrence, based on the chi-square value. *Residual exceeds specified limits of greater than 2.0. Notes. Items 1, 4, 5, 6, 12, 23, 29, 30, 31, 32, and 42 have two response categories (1 threshold). The rest have are 3 response categories (2 thresholds). Thresholds are points between adjacent categories where the odds are 1:1 of answering either category.