Student Learning Identities: Developing a Learning Taxonomy for the Political Science Classroom

Abstract

The present article uses Q-Method to uncover, what we refer to as, learning identities in an undergraduate core political science course. The term “learning identities” is employed to highlight the self-referential quality of the learning perspectives revealed in the Q-Sorting exercise. Drawing on a set of 41 objectivist statements derived from the broad field of learning style inventories, a taxonomy of five distinct learning identities is described and discussed below. These include: Adept Learners, Traditionalists, Obliged Pupils, Apathetic Pupils, and Objective Learners. After describing these ideal-typical learner self-conceptions, the authors discuss how this knowledge can be a valuable tool for college and university political science faculty, informing such efforts as curriculum development and instructional approaches. Although learning style inventories have long been a part of the pedagogical literature, few attempts have been made to unify the varying measures of personality type, information-processing approaches, and environmental preferences into a single self-referential tool, and fewer approaches still have sought to focus such a tool specifically on the task of teaching politics at the undergraduate level. The present article seeks to contribute to the task of filling this void.

Keywords:

• inventory
• learning styles
• pedagogical differentiation
• political science
• Q-Methodology
• student perceptions

Notes

Note:

(∗) Those coefficients with t-values greater than |1.98| are significant at p = 0.05.

(∗∗) Those coefficients with t-values greater than |1.67| but less than |1.98| are significant at p = 0.10.

For a comprehensive review and assessment of this literature, see the Learning and Skills Research Centre's 2004 report, entitled “Learning Styles and Pedagogy in Post-16 Learning: A Systematic and Critical Review” (Coffield et al. 2004).

Many statements may easily fit into more than one dimension. Rather than mutually exclusive categories, the framework is meant to provide a useful heuristic tool for negotiating the statement selection process (Brown 1980 1986; Dryzek and Berejikian 1993).

Approximately 60 students are selected at the beginning of each semester to take Advanced American Politics rather than the core introductory American Politics course. The selection is based on student GPA. Thus, this portion of the sophomore class will be missing from any random sampling of the introductory course. Nevertheless, the freshmen selected to replace this group of 60 sophomores are typically among the top performers in their class, providing an exchange that, in terms of GPA, has little impact on the average student achievement for the students in the course.

Varimax rotation is the most often used method of factor rotation. Despite this, some methodologists prefer the use of theoretical rotation to test a priori assumptions. Varimax is employed here to allow structure to emerge from the data, without the imposition of preconceived categories (Brown 1986; Brown, Durning, and Seldon 1998). Five factors had eight or more significant loadings, and together these five factors explained approximately 50% of the total variance in the sample. Nevertheless, the primary reason for the extraction of these five factors was theory driven. The inclusion of additional factors did not reveal any new, theoretically compelling perspective, but simply offered further specified gradations of the first four factors (Brown 1980).

These model sorts or factors were constructed by merging individuals' significant loadings. Some significant loadings were, of course, more closely aligned with the factors than others. Consequently, factor weights had to be computed using the formula w = f/1 − f ², where f is the factor loading and w is the weight (Brown 1980; McKeown and Thomas 1988). Factor weights could then be used along with the data collected from the respondents' statement orderings to produce factor scores for each of the factors on the 41 statements.

In Table 3, an “X” next to a numeric entry indicates statistical significance at the p < .01 level.

Computerized course management and evaluation systems fit into this identity in several ways. First, they can facilitate a better organized course in which information and schedules are widely available and easily accessible. Second, universalizable conceptions of knowledge as a collection of brute facts lend themselves to the rationalized assessment tools seen in most web-based programs.

The conditional probabilities presented in this section were calculated using an Ordinary Least Squares (OLS) regression model. We generated six separate models where each of the five learning identities (and one for those who do not ascribe to any particular identity) provided a dichotomous dependent variable and were arrayed against the demographic factors enumerated in the previous paragraph (gender, GPA, major, etc.).

In fact, in one instance, there was a relationship between a particular instructor and the likelihood of his assigned students to associate with a particular learning identity. Students in one instructor's sections were 11% more likely to ascribe to the Obliged Pupil learning identity but were 11% less likely to ascribe to the Apathetic Pupil learning identity. This indicates that both the common curriculum for the course together with instructor approaches can both have important impacts on student learning identities for the course. Moreover, it indicates the need for additional research to determine what kinds of instructional approaches might lead to changes in course learning self-conceptions.

Of course, if this exercise were administered in a math or engineering course, one might find different patterns among the data. The environmental impact of the political science classroom focuses attention on these types of courses.

As previously mentioned, the limited number of freshmen in the sample course was 12% less likely than their sophomore colleagues to group in this identity.

A preliminary draft of this article was delivered at the Annual Meeting of the Southern Political Science Association in Atlanta, Georgia, January 7, 2006. The authors would like to thank the Center for Teaching Excellence and the Academic Research Division at the U.S. Military Academy, West Point, New York for providing assistance and suggestions in the preparation of this article. The views expressed here are those of the authors and not those of the United States Military Academy or the U.S. Army.