Development of quantum perspectives in modern physics

Introductory undergraduate courses in classical physics stress a perspective that can be characterized as realist; from this perspective, all physical properties of a classical system can be simultaneously specified and thus determined at all future times. Such a perspective can be problematic for introductory quantum physics students, who must develop new perspectives in order to properly interpret what it means to have knowledge of quantum systems. We document this evolution in student thinking in part through pre- and post-instruction evaluations using the Colorado Learning Attitudes about Science Survey. We further characterize variations in student epistemic and ontological commitments by examining responses to two essay questions, coupled with responses to supplemental quantum attitude statements. We find that, after instruction in modern physics, many students are still exhibiting a realist perspective in contexts where a quantum-mechanical perspective is needed. We further find that this effect can be significantly influenced by instruction, where we observe variations for courses with differing learning goals. We also note that students generally do not employ either a realist or a quantum perspective in a consistent manner.


I. INTRODUCTION
In the last decade, studies of student beliefs about physics have become a focus of interest in the physics education research (PER) community. Several assessment instruments have been developed in order to characterize student beliefs about the nature of physics and of learning physics, [1--4] including the Colorado Learning Attitudes about Science Survey (CLASS). [5] Previous studies of introductory physics courses have used the CLASS to show that student beliefs can be correlated with conceptual understanding, [6,7] as well as with self--reported interest in physics. [8] With notably few exceptions, [1,9,10] studies have found that students tend to shift to more unfavorable (novicelike) beliefs about physics and of learning physics. [1,7] However, relatively little attention has been paid to student beliefs about physics beyond introductory courses in classical mechanics and electrodynamics. [11] We seek to examine the development of student beliefs as they make key transitions from learning introductory classical physics to their first advanced or specialized course in modern physics.
Prior research on modern physics [12--14] has been predominately concerned with identifying student misconceptions and difficulties in learning the formalism of quantum mechanics. Surveys have been developed to assess students of quantum physics, but have generally focused on common difficulties for advanced undergraduate and beginning graduate students, such as the calculation of expectation values or the time evolution of a quantum state, [15,16] or they have studied how students interpret physical meaning from graphical representations of various wave functions. [17] Others have developed conceptual surveys appropriate for lower--division modern physics students based on research on common student misconceptions. [18--21] Still, student commitments with respect to ontology (mental models of the physical world) and epistemology (beliefs about the nature of knowing) in the context of quantum physics have been understudied, particularly regarding the potentially difficult transition students make from learning classical physics to learning quantum physics.
Introductory courses in classical physics promote a perspective that we call local realism. A realist perspective is deterministic, in the sense that all physical quantities describing a system can be simultaneously specified and accurately predicted for all future times. Such a perspective is often employed in the context of classical electrodynamics; for example, students are typically instructed to model an electron as a localized particle having both a well--defined position and momentum.
This idea of locality can sometimes be useful in the context of modern physics: when learning about the photoelectric effect, a particle model for both electrons and photons is required. A particle model would be inappropriate, however, when trying to explain the interference pattern seen in a double--slit diffraction experiment. In this case, from a quantum--mechanical perspective, electrons and photons behave as delocalized waves as they propagate through space and as particles when interacting with a detector. A quantum perspective also recognizes the probabilistic nature of measurements performed on quantum--mechanical systems, in contrast to the determinism assumed by Newtonian mechanics. [22] This paper concerns itself with how student perspectives change as they make the transition from learning classical physics to learning quantum physics. An analysis of student responses to pre-- and post--instruction surveys at various stages of an undergraduate introductory sequence allows us to infer the development and reinforcement of a deterministic perspective during classical physics instruction, as well as the emergence of a probabilistic perspective in students as they progress through a course in modern physics. In addition, from comparative studies of two classes, we find that student perspectives can be significantly influenced by an instructor's choice of learning goals. Students are more likely to apply a quantum perspective following a course where such a perspective is explicitly taught. We also demonstrate that a student's degree of commitment to either a realist or a quantum perspective is not necessarily robust across contexts. We find that students may simultaneously hold both realist and quantum perspectives and not always know when to employ each of these epistemological and ontological frames.
We conclude from the available data that specific attention paid to the ontological interpretation of quantum processes during instruction may aid students in the cultivation of a desired quantum perspective.

II. STUDIES
The University of Colorado offers a three--semester sequence of calculus-based introductory physics: PHYS1 and PHYS2 are large--lecture courses (N ~ 300--600) in classical mechanics and electrodynamics, respectively, [23] and PHYS3 is a course in modern physics, offered in two sections, each with a typical class size of 75 students. At the beginning and end of each semester, students from each of the above courses were asked to respond to a series of survey questions designed to probe their epistemic and ontological commitments. The first of these surveys was an online version of the CLASS, wherein students responded using a five--point Likert--type scale (ranging from strong disagreement to strong agreement) to a series of 42 statements, including:  [24] Among students starting off in PHYS1, many more will agree (40%) with this statement than disagree (26%); yet the number in agreement decreases significantly following instruction in classical physics (to 30%, p < 0.001), while an increasing number of students disagree (39%). This trend then reverses itself over a single semester of modern physics; at the end of which a greater percentage of students agree with this statement (46%) than at the beginning of classical physics instruction. In a longitudinal study of 124 students over three semesters, we observe the same trends, shown in Fig. 1 [ Table  I] The distribution of responses for pre--instruction modern physics students was similar to that of classical physics students, and so the data for both have been combined into a single group, shown in Table II. correct answer) was preferred by those who disagreed (69%). We conclude from this that, among students of introductory classical physics, those who disagree with No. 41 primarily concern themselves with the idea that there can be only one correct result for any physical measurement, while those who agree with the statement are more conscious of the possibility for random hidden variables to influence the outcomes of two otherwise identical experiments. We find that few students invoke quantum phenomena when responding before any formal instruction in modern physics (despite the fact that a majority of these modern physics students reported having heard about quantum mechanics in popular venues, such as books by Greene [26] and Hawking, [27] before enrolling in the course); however, a single semester of modern physics instruction results in a significant increase in the percentage of students who believe that quantum phenomena could allow for two valid, but different, experimental results. Students shift from 10% to 32% in providing quantum--specific reasoning, and from 13% to 49% in referencing quantum or relativistic reasons for agreeing with the statement.
[ Table III] Responses from each population were compared with a chi--square test and were found to be statistically different (p < 0.001).

II.B. Influence of instruction on student perspectives
Students' commitments to either a realist or quantum perspective can vary by context; [28] to see if these commitments can be influenced by different types of instruction and learning goals, we examined data from two recent semesters of PHYS3 for physics majors. Course PHYS3A was taught by a PER instructor who employed in--class research--based reforms, [29] including computer simulations [30] designed to provide students with a visualization of quantum processes.
Course PHYS3B was taught the following semester in the form of traditional lectures delivered from a chalkboard. Both classes provided online and written homework, Student 1: "That blob represents the probability density, so it tells you the probability of where the electron could have been before it hit the screen. We don't know where it was in that blob, but it must have actually been a tiny particle that was traveling in the direction it ended up, somewhere within that blob."

Student 2: "No, the electron isn't inside the blob, the blob represents the electron!
It's not just that we don't know where it is, but that it isn't in any one place. It's really spread out over that large area up until it hits the screen." Student 3: "Quantum mechanics says we'll never know for certain, so you can't ever say anything at all about where the electron is before it hits the screen." FIG. 2. A sequence of screenshots from the quantum wave interference PhET simulation.
In this end--of--term survey question, students were asked to agree or disagree with any or all of the fictional students and to provide evidence in support of their response. Responses were coded according to whether students preferred a realist or a quantum perspective in their argumentation. A random sample of 20 student responses were recoded by a PER researcher unaffiliated with this project as a test for inter--rater reliability; following the discussion of the coding scheme, the two codings were in 100% agreement. The following quotations from two students are illustrative of the types of responses seen: Student response (realist): "We just can't know EXACTLY where the electron is and thus the blob actually represents the probability density of that electron. In the end, only a single dot appears on the screen; thus the electron, wherever it was in the probability density cloud, traveled in its own direction to where it ended up."

Student response (quantum): "The blob is the electron and an electron is a wave
packet that will spread out over time. The electron acts as a wave and will go through both slits and interfere with itself. This is why a distinct interference pattern will show up on the screen after shooting out electrons for a period of time." The distribution of all responses for the two courses is summarized in Table  IV; columns do not add to 100% because some students provided a mixed or otherwise unclassifiable response. For this essay question, there is a strong bias toward a quantum perspective among PHYS3A students, while students from PHYS3B highly preferred a realist perspective. Notably, virtually no student agreed with fictional student 3 (which would be consistent with an agnostic perspective); among those who explicitly disagreed with student 3, most felt that knowing the probability density was a sufficient form of knowledge about this quantum system.  Suppose a machine releases a marble from the same starting point 300 times, and the cumulative results for where the marble ends up are shown in the histogram below (Fig.  3). There is a distribution of possible final outcomes for each drop of the marble, even though the initial conditions for each drop seemed to be the same. Q2--(I) What is the origin of the uncertainty in the final outcome for this classical system? (Explain your answer in 2-3 sentences at most.) Q2--(II) The distribution shown in the histogram above looks similar to a distribution of measurements on a quantum system (for example, one part of an interference pattern created during a double--slit experiment). In what ways is the uncertainty in final outcomes for such a quantum system the same as or different from the classical example given above? This question was designed to cue students to consider how an uncertainty in initial conditions will lead to varying final outcomes in a classical system (specifically, through the phrase "the initial conditions for each drop seemed to be the same"). Students were then asked to compare and contrast this origin of uncertainty (for a chaotic but deterministic system) with an example from quantum physics. We expect that students who are committed to a realist perspective would view the two examples as similar, so that the uncertainty in initial conditions for the Plinko marble would be seen as analogous to the perceived uncertainty in the initial conditions for electrons in a diffraction experiment. Most every student provided a satisfactory response to the first part of the question, stating either that the distribution of final results was due to a random 50/50 probability on how the marble would be deflected at each peg or due to an uncertainty in the initial conditions for this classical system. Student responses to part II were rated using the rubric shown in Table  V; most every student commented on the systems being similar in that there is some kind of distribution of final results (or that both can be described with statistics, probability, etc.), and so the rubric does not code for this response and focuses instead on the argued differences between the two systems.
The results for all student responses are summarized in Table VI.

B (Quantum)
Different because there is no interference or the marble is localized in space. "Electrons behave like waves."

C (Realist)
No statement about differences or thinks they are the same. Implies there are differences, but reasoning is unclear or weak The following is the full response of one student who coded as an A category: Q2--(I) "The origin for the uncertainty comes from the variables of the initial conditions. The Plinko ball can't be dropped exactly the same way every time, and so not all the balls follow the same path." Q2--(II) "In a quantum system, the initial conditions can be exactly the same in every case, but the outcomes can be different. The reason the quantum distribution looks the same as the macro distribution is because quantum distributions follow probabilities which are similar to classical distribution patterns." As can be seen in Table  VI, few students from either semester provided the complete targeted response, which was to recognize that the classical Plinko game is a false analogy to a quantum system, where there can be varying outcomes to measurements even though the initial conditions are identical. Still, a majority of PHYS3A students perceived that there is some difference between the two examples (the most common response is that the classical system does not exhibit interference effects or that electrons behave as waves, while the marble does not), while a majority of PHYS3B students seemed to believe that the origins of uncertainty in both systems were analogous or were unable to articulate why they might be different.   Table IV). The post--data in Table VII from the two courses can be combined and compared with student   responses  to  the  prior  essay  question  on  double--slit  interference.    Table  VIII  shows the student post--instructional responses to QA No. 16, categorized by which perspective they held on the QWI essay question of Fig. 2. Here, we see that students who had preferred a quantum perspective tended to answer QA No. 16 favorably, while the majority of students who preferred a realist perspective chose an unfavorable response. Of particular interest, however, is that students were not necessarily consistent in their responses: 18% of those who disagreed with QA No.
16, and 33% of those who agreed, were offering a response that was inconsistent with their response to the QWI essay question. That is, 18% of students held a quantum perspective on electron position (QA No. 16), but a realist perspective on the quantum wave interference question. 33% of students were the reverse: holding a realist perspective on electron position in QA No. 16 (agree), but a quantum perspective on the interference question.  Quantum  56  11  33  100  Realist  18  18  64 100 Nonetheless, the majority of students held a consistent quantum or realist perspective on the two questions relating to electrons (quantum wave interference and electron position in an atom). By including analysis of student perspectives with respect to the second essay question (sources of classical and quantum uncertainties), we may consider the consistency of student perspective across a third context. The data for both courses (PHYS3A and PHYS3B) have been combined in Table IX, and student responses are categorized as quantum (or realist) if students consistently report an answer coded as quantum (or realist) across all three questions; otherwise, students are reported as mixed (giving at least one quantum and one realist answer). Table IX shows that the majority of students end up with a mixed perspective (p < 0.01 by pairwise proportion test to each of quantum and realist groups), sometimes applying a quantum perspective and sometimes a realist perspective. Intriguingly, more students (p = 0.05, two tailed, and pairwise) end with a consistently realist perspective than a quantum perspective. The dominance of the mixed state holds for each of the PHYS3A and PHYS3B courses independently (71% and 50%, respectively); however, in PHYS3A the quantum state is more prevalent than realist (22% vs 7%) and in PHYS3B the realist state is more prevalent than quantum (39% vs 11%).

III. DISCUSSION AND CONCLUSIONS
The data presented in this paper serve as evidence in support of three key findings. First, student perspectives with respect to measurement and determinism in the contexts of classical physics and quantum mechanics evolve over time. The distribution of reasoning provided by students in response to the CLASS survey statement indicates that the majority of those who disagree with this statement believe that experimental results should be repeatable or that there can be only one correct answer to a physics problem. One could easily imagine that students begin their study of classical physics at the university level with a far more deterministic view of science than is evidenced by their initial responses to the survey statement (after all, most students do arrive with some training in classical science). We take the first significant shift in student responses (a decrease in agreement and an increase in disagreement with this statement, as shown in Fig. 1  While we do not make a valuation of either of these instructional goals, we feel it is worth acknowledging that different goals regarding the interpretation of quantum processes do exist. We believe that the physics community would benefit from a discussion of the pedagogical usefulness of either of these interpretations because our research indicates that students, in this regard, can adopt their instructor's philosophical predisposition when given explicit instruction. We believe that this in itself is a significant finding, considering that there is substantial evidence that students do not necessarily adopt an instructor's views and attitudes in other contexts. For example, students will often not develop a sound conceptual understanding of physics, even if instructors believe in the importance of such, unless conceptual understanding is explicitly taught, as is evidenced by myriad studies. Furthermore, students tend not to develop more sophisticated views on the nature of science and of learning physics, even from reformed instruction in introductory courses. [1,7] In fact, students' views on the nature of physics and learning tend to become less "expertlike" over time, although it has been shown that this trend can be positively influenced by making epistemology an explicit aspect of instruction in introductory physics courses. [10] The results of this study provide further indication that instructors should not take for granted that students will adopt their perspectives on physics unless such learning goals are made explicit in their teaching. Third, we find that most students do not exhibit a consistent perspective on uncertainty and measurement across multiple contexts. While the data shown in Table VIII do demonstrate some consistency of responses when answering two questions on electron position, we see that a significant number of students who preferred the quantum description of an electron in a diffraction experiment would still agree that an electron in an atom has a definite, but unknown, position. When looking across more varied contexts that include a question comparing electron diffraction and a Plinko game, students exhibit a tendency to be less consistent, dominantly holding mixed quantum and realist perspectives. Students likely do not have a robust "concept" of quantum measurement. These findings parallel studies of student epistemic commitment in classical physics [32] and the resources view of student conceptual understanding and understanding the nature of knowing physics. [33] In the end, we believe that a reasonable instructional objective is for students to use the appropriate perspective (deterministic or probabilistic, localized or delocalized) at the appropriate time. This goal seems to require a level of metacognitive awareness that students may not have at the introductory level: we find that few students from either course were able to demonstrate the ability to distinguish between classical uncertainty and the uncertainty that is inherent to quantum systems. While a majority of students from the transformed PHYS3A course demonstrated at least partial understanding of this distinction (by focusing on interference and the wave description of electrons), a majority of PHYS3B students did not make any reasonable distinction between the two systems (which is again consistent with a realist perspective).
These findings suggest that students do not automatically develop the robust understanding of measurement, uncertainty, or metacognitive abilities that we may desire. If we are to include these goals for our classes, it is important to understand how these messages are sent to our students and what instructional practices may promote such understandings. Such investigations are the subject of current studies.