(Post last updated July 21, 2021)
Review panel summary
The CSCI was developed to measure a student’s “self-concept as a learner of chemistry” [1]. It consists of 40 items where students respond to each item using a seven-point Likert scale, ranging from “very inaccurate of me” to ”very accurate of me”. Many of the items derive from a previously published instrument, Self Description Questionnaire III (SDQIII), with items modified and added to address self-concept in chemistry specifically [1]. Student response data has been collected from multiple courses including university general chemistry [1-4], one-semester nursing chemistry [7], and four variations of high school chemistry [6].
During the development of the instrument, exploratory factor analysis identified five subscales [1]. Using data collected from high school students, an exploratory factor analysis suggested a six-factor solution was more appropriate [6]. These results suggest that there is conflicting evidence of internal structure validity. Comparisons of the scale scores of general chemistry students (non-majors) to chemistry majors [1] and high school students in different chemistry courses [6] have shown expected differences in self-concept among the different groups. Student scores on two of the subscales have been used as an input variable in cluster analysis to establish student groups and membership in these groups has been found to be related to course performance [2, 3, 6]. This provides some validity support in regards to relations to other variables. Coefficient alpha has been used to estimate the single administration reliability of the subscales [1, 2, 4, 5, 7]. Across studies, the Math and Chemistry Self-Concept subscales consistently have higher alpha values (0.89 - 0.93) than the others (0.25 - 0.80). Evidence of suitable test-retest reliability has been reported [1, 5].
Recommendations for use
The CSCI has seen limited use in university general chemistry and high school chemistry courses. The conflicting evidence for internal structure found in the literature leads to the recommendation that users analyze data collected using the instrument for further evidence in support of internal structure. While some evidence of internal structure has been presented based on exploratory factor analysis, future users are encouraged to consider moving to a confirmatory factor analysis framework to explore the fit of an existing structure, thereby providing additional data about the suggested factor structures. In cases where a user cannot collect sufficient data to investigate the internal structure, the Math and Chemistry subscales have demonstrated the most stability, and they could potentially be used separately.
Details from panel review
Principal components analysis was used to identify five factors in a development study, with two items not associating with any of the factors. The five factors of self-concept were labeled: Mathematics, Chemistry, Academic, Academic Enjoyment, and Creativity [1]. Two of the factors were in good agreement with factors found in the SDQIII, while three were different. When data from high school students were analyzed using principal axis factoring, a different factor structure was found with some resemblance to the factor structure of the original SDQIII [5]. In a study with high school students [6], three items were removed, based on an analysis of response distributions, after the initial exploratory factor analysis, but the authors did not address if the fit improved. The evidence for the internal structure of the entire, five-subscale instrument, is therefore limited and conflicting. There is evidence that the Math and Chemistry Self-Concept subscales are robust.
Two studies have collected data that compared the scale scores of students expected to have different levels of self-concept [1, 5]. One study with college students found that the self-concept scores of chemistry majors was higher than that of non-chemistry majors in a general chemistry class [1]. Expected patterns of self-concept scores were also found amongst high school students enrolled in different levels of chemistry courses. Self-concept scores increased from non-college prep chemistry, to college prep chemistry, to honors chemistry, to Advanced Placement chemistry students [5]. In this same study [5], the scores on the subscales were found to be correlated with scores on a chemistry misconceptions assessment. Another study used the increase in subscale scores (pre to post) as evidence of the success of a course redesign [7]. There has been use of the Math and Chemistry subscale scores as input variables in cluster analysis [3, 6]. These reports provide limited evidence for validity with respect to relations with other variables.
Reliability evidence has been based on single administration reliability by providing coefficient alpha values, for each subscale. The coefficient alpha values for the Math and Chemistry Self-Concept have been consistently higher than the other three scales. Test-retest reliability has been evaluated in two ways. In the development process, the instrument was given twice, one week apart, at the beginning of the semester with correlation values ranging from 0.90 - 0.64 [3]. Another study compared scores two months apart by examining the scatter plot and reported an R2 value of 0.6267 [5].
References
[1] Bauer, C. F. (2005). Beyond “Student Attitudes”: Chemistry Self-Concept Inventory for Assessment of the Affective Component of Student Learning. Journal of Chemical Education, 82(12), 1864. https://doi.org/10.1021/ed082p1864
[2] Lewis, S. E., Shaw, J. L., Heitz, J. O., & Webster, G. H. (2009). Attitude Counts: Self-Concept and Success in General Chemistry. Journal of Chemical Education, 86(6), 744. https://doi.org/10.1021/ed086p744
[3] Chan, J. Y. K., & Bauer, C. F. (2014). Identifying At-Risk Students in General Chemistry via Cluster Analysis of Affective Characteristics. Journal of Chemical Education, 91(9), 1417–1425. https://doi.org/10.1021/ed500170x
[4] Chan, J.Y.K. and Bauer, C.F. (2015), Effect of peer-led team learning (PLTL) on student achievement, attitude, and self-concept in college general chemistry in randomized and quasi experimental designs. Journal of Research in Science Teaching, 52: 319-346. https://doi.org/10.1002/tea.21197
[5] Nielsen, S. E., & Yezierski, E. (2015). Exploring the Structure and Function of the Chemistry Self-Concept Inventory with High School Chemistry Students. Journal of Chemical Education, 92(11), 1782–1789. https://doi.org/10.1021/acs.jchemed.5b00302
[6] Nielsen, S. E., & Yezierski, E. J. (2016). Beyond academic tracking: Using cluster analysis and self-organizing maps to investigate secondary students’ chemistry self-concept. Chemistry Education Research and Practice, 17(4), 711–722. https://doi.org/10.1039/C6RP00058D
[7] Smith, A. L., Paddock, J. R., Vaughan, J. M., & Parkin, D. W. (2018). Promoting Nursing Students’ Chemistry Success in a Collegiate Active Learning Environment: “If I Have Hope, I Will Try Harder.” Journal of Chemical Education, 95(11), 1929–1938. https://doi.org/10.1021/acs.jchemed.8b00201