Monday, June 03, 2019

Place Value Paper Presented at the 2019 Annual Conference of the Jean Piaget Society


Here's the extended abstract for the paper. The Full Paper is under "Elementary Mathematics".

Extended Abstract

This study of 76 6th, 7th, and 8th grade students investigated students’ understanding of place value numerals. It was the research component of the Mollala Math Project in Oregon as generated by the opportunity to conduct embedded research into math development. The 11 teachers of the Molalla River District Math Leadership Team provided district coordination and surprising contributions to the conceptualization of place value understanding in students and the several years of oversight, editorial work, and field testing in converting the interviews into assessment tasks. Their primary purpose was assist teachers in advancing their understand of their students’ math development.
The purpose of this research project was to investigate why math test scores declined from 3rd to 10th grade. About 70% of 3rd graders passed the Oregon state mathematics test. At 10th grade less than a third passed. The decline accelerated after 5th grade. Our investigation entailed three parts: the use of developmental theory to uncover and describe the development of understanding in mathematics, the use of Rasch modelling to verify and extend the developmental findings, and the integration of the theory and the Rasch modelling to improve methods of differentiated instruction that match teaching to each student’s needs.
The first investigation was immediately sidetracked when we could not find a valid definition of place value numeration or its operational components in the literature or instructional materials, This lack was also a complaint of teachers so we undertook a project with the Math Leadership Team to abstract and define the necessary and sufficient elements and operations that would specify the place value system and distinguish it from the number system, a distinction surprisingly almost never made in instructional materials or curriculum guidelines. Numerals are the symbol systems for generating representations of numbers and all depend more or less on some system for mathematically generating and transforming numerals. The mathematical validity of place value system is exactly the same as the criteria of any valid number systems since it has full group properties including reversibility and conservation. We arrived at four defining elements and five basic operations. This system of place value operations then makes possible two derivative operations. These coordinate the movement of values across places within a numeral.
Next the committee formulated and refined five interviews organized as Piagetian type of tasks. Each task posed a specific problem for students which triggered reasoning we could follow by why and how questions directed to determining the complete features of their understanding: The Place’s Values (What do the digits in 16 mean?); Re-grouping of One’s Into Ten (How are place totals re-grouped in 28 + 3?); Column Values in Long Division (Is place value maintained in columns?); Conservation of Place Value (Is the numeral’s total value maintained after borrowing?); and Naming of Transformed Numerals (What language does the student use for meaning of a transformed numeral after borrowing?) These results were condensed into the criteria that defined the stages: “The intra phase leads to the discovery of a set of properties in objects and events finding only local and particular explanations. The ‘reasons’ to be established can thus be found only in the relations between objects which means that they can be found only in ‘transformations’. These, by their nature, are characteristic of the inter level. Once discovered, these transformations require the establishment of relations between each other which leads to the construction of ‘structures’ characteristic of the trans level.” We then had descriptions of how students understood key aspects of place value numeration at three stages. For this place value study, only two of these stages were used, the intra-numeral stage of understanding (stage 1), and the inter-numeral stage subdivided into two levels, a beginning (stage 2) and a completed form (stage 3). The trans-numeral stage was not used since it extends considerably beyond the elementary level in complexity to the level of explicit, formal, abstract understanding of place value as a system. 
The sample for these five studies was composed of two cohorts, the first was selected as an epistemic sample designed to produce the widest possible range of place value understanding. This teacher selected epistemic sample were students teachers selected as high, medium, or low math ability. They selected 17 remedial 7th and 8th grade students and 11 7th and 8th grade students judged of high ability. The second cohort was a random sampling of 34 students from  within three groups of 6th grade students, a group of high (n = 12), medium (n = 11), and low (n = 11) students identified according to their state test scores.
The results revealed a complexity of the place value system that was mathematically beyond the understanding of most of these middle school students. At the intra-figural stage 1, students made no distinction between number and numeral, understood numerals as unitary figural symbols, and used simple procedural notions in problem solving. For example, Chad, grade 8, showed 6 chips for the 6 and 1 chip for the 1 in 16. When asked about the remaining 9 of his 16 chips, he said, “You don’t need the rest.” (But will we both have sixteen if you don’t use those?) “I don’t know,” he replied.
At the beginning inter-figural stage 2, students used a beginning but weak and unstable form of place value understanding that embedded numerals with limited place value relationships. Ross, a grade 8 student, showed an early form of stage 2 thinking in the 16’s task, “If I was explaining it to someone I would say, I took out 6 and then ask him how many are left and say…” (indicates two piles of five). “If he thought it was a 1, I’d show him it (the two piles of 5) couldn’t be. There is 5 and 5 and the 10, and the 1 means a ten and not just a plain 1.” He showed a pile of six and a pile of five and another pile of five. Places for him had significance but only as empirically necessary to account for all the chips, not because a base was explicitly involved.
At the completed inter-figural stage 3, the conservation of place value appeared and students extended place value operations to new problems with reasoning controlled by place value logic. For example, Konstantin in grade 7 was an unusual case because English was not his native tongue. In #1, he showed 1 chip for the 1 in 16 but then almost immediately seemed to understand. He self-corrected before being questioned and said “10.” From his performance on the other tasks, this misstep was seen as one of miscommunication, not misunderstanding.
The combined results of the three equal sized samples of high, medium, and low 6th grade students showed about half the students had completed understanding on the 16’s task. Since the borrowing problem depends on the conservation of place value of a numeral, it is the best indicator of a completed conceptual understanding of place value numerals, and least susceptible to learning effects. In spite of the relative overweighting of the high student sample, the results still showed only 13% of the combined samples had the completed understanding of stage 3. Although appearing as rather simple, (and often presented as simple in instructional demonstrations) place value numeration is, in fact, a mathematically rich system whose complexity is not mastered by the majority of students up through middle school.
In the second investigation, place value data were combined with other data, and Rasch modeling was used to measure the difficulty of the five place value tasks and locate them within a developmental math continuum of 30 concepts. The interview data pool from 159 kindergarten through 8th grade students were coded into the 3 stages and used in the Rasch modelling. With the conservation of number anchored to 2.0 logits, the place value concept and volume both had a difficulty score of 4.5 logits with 14% of the middle school cohort succeeding on the conservation of place value task. The task fit scores  confirmed the validity and difficulty of the place value tasks (infit M = 1.01, S.D. = 0.2, t = – 0.02) and supported the conclusion that these public school students did not understand the concept of place value until well into the middle school years.
In the third part, individual mind maps showing each concept’s stages, achieved and not yet achieved by the student, graphically displayed a zone of development around each student’s developmental level. The mind map method of reporting also showed teachers how quantitative test scores could be diagnostically related to stage descriptions of actual understandings. Combining developmental systems theory with Rasch modeling supports the possibility for developing a continuous progress model of math instruction by matching teaching and learning to each student’s specific developmental needs and abilities.
-->