Learning Progressions as Empirically Inadequate but Productive Models for Classroom Instruction

• In Event: Tracing Learning Progressions/Trajectories: Novel Theoretical and Methodological Perspectives at the Intersection of Mathematics and Science

Abstract

LPs are commonly evaluated with respect to an important attribute of scientific models: fit to empirical data. However, as both documented empirically and argued theoretically, LPs often come up short on this criterion (e.g., Sikorski & Hammer, 2010). Indeed, Author (2019b) argues that this inadequacy is inherent to LPs because student thinking and learning are more context-dependent than can be represented in coherent levels.
Philosophers of science highlight utility and generativity as additional attributes of scientific models (e.g., Odenbaugh, 2005). From this perspective, LPs may have considerable value as models, despite their empirical inadequacy. Analyzing teachers’ work with LP-based assessment information in an interview format, Author (2019b) provided preliminary evidence of LPs’ utility and generativity for teachers. In this paper, I move from the laboratory setting to real classrooms, asking:
1) How do teachers make sense of LPs’ characterization of student thinking and learning?
2) To what extent does teachers’ use of LPs provide evidence of the criteria of a) utility and b) generativity?
This study focuses on six high school physics teachers who participated in a two-year professional development program focused on LPs for three topics in mechanics (see Figure 5). A key feature of the program is that it did not prescribe particular applications of LPs; rather, the research team worked to support teachers to use these tools as they perceived relevance to their individual instructional contexts.
The primary data source is transcripts of teacher interviews conducted at the end of Years 1 and 2 (interviews 3 and 4 in Figure 5). Using codes from Author (2019b) and Author (2018), I coded idea units in three successive rounds of coding, corresponding to research questions 1, 2a, and 2b, respectively: characterization of student thinking and learning, LP utility for instruction, and LP generativity. Table 2 includes illustrative examples of each type of code.
The results of this study provide support for the utility and generativity of LPs suggested by Author’s (2018) laboratory study. Selected evidence for the following claims is presented in Table 1 and will be more fully explicated in the paper.
Claim 1: Teachers each characterized students’ thinking and learning with respect to the LPs in multiple ways, often moving fluidly between characterizations congruent and incongruent with LP assumptions.
Claim 2a: LPs had utility in supporting instruction that took typical student ideas into account; teachers relied on multiple ways of interpreting LP-based information about student thinking and learning to draw instructional implications.
Claim 2b: Teachers’ sensemaking about student thinking and learning with respect to the LPs included generation of insights and hypotheses about student thinking and learning that extended beyond information directly available in the LPs.
This work has implications for ongoing conversations about the nature of students’ thinking and learning and its consequences for LP research and applications (e.g., Hammer & Sikorski, 2015), particularly since teachers found utility in a range of different interpretations of LP-based information about student thinking and learning. Questions remain as to the quality of instruction associated with different interpretations and insights.

Author

• Alicia C. Alonzo, Michigan State University