Faculty survey on upper-division thermal physics content coverage
FFaculty survey on upper-division thermal physics content coverage
Katherine D. Rainey and Bethany R. Wilcox
Department of Physics, University of Colorado, 390 UCB, Boulder, CO 80309
Thermal physics is a core course requirement for most physics degrees and encompasses both thermodynam-ics and statistical mechanics content. However, the primary content foci of thermal physics courses vary acrossuniversities. This variation can make creation of materials or assessment tools for thermal physics difficult.To determine the scope and content variability of thermal physics courses across institutions, we distributeda survey to over 140 institutions to determine content priorities from faculty and instructors who have taughtupper-division thermodynamics and/or statistical mechanics. We present results from the survey, which high-light key similarities and differences in thermal physics content coverage across institutions. Though we seevariations in content coverage, we found 9 key topical areas covered by all respondents in their upper-divisionthermal physics courses. We discuss implications of these findings for the development of instructional toolsand assessments that are useful to the widest range of institutions and physics instructors. a r X i v : . [ phy s i c s . e d - ph ] N ov . INTRODUCTION Thermal physics, which includes both thermodynamicsand statistical mechanics, is a core course required for attain-ing a physics bachelors’ degree at most institutions. However,anecdotally the material covered in thermal physics coursesoften varies between instructors and across institutions. Thiscontent variability poses a significant challenge in develop-ment of standardized thermal physics assessments and teach-ing tools that can be utilized by a wide range of instruc-tors. Though there is a body of research surrounding studentunderstanding of thermal physics concepts , less is knownabout the breadth of topics covered in upper-division thermalphysics courses.Here, we present findings from a survey distributed withthe purpose of soliciting instructor priorities in upper-divisionthermal physics as a part of a broader research effort to de-velop a standardized upper-division thermal physics assess-ment. Findings may lay an important foundation for otherresearchers interested in developing course materials andassessments for thermal physics, and inform instructors indefining course objectives and content-foci for their thermalphysics courses.In this paper, we begin by describing the process of con-structing and distributing the survey (Sec. II). Then, wepresent results of the survey (Sec. III), including generalcourse information, key concepts covered, and valued sci-entific practices. We also consider response consistency be-tween survey responses and submitted syllabi, followed byan analysis of content variability across institutions. We con-clude with a short consideration of implications of the surveyand future directions (Sec. IV). II. METHODS
The faculty survey was designed to solicit key informationabout thermal physics courses, such as content covered, gen-eral course structure and emphasis (thermodynamics, statis-tical mechanics, or both), and needs or interest in an upper-division thermal physics assessment. This section describesmethods for developing and distributing the survey with anemphasis on creating a format that was accessible and rela-tively short in duration, while still soliciting sufficient infor-mation.
Survey Development:
Prior to constructing the survey, a fo-cus group was conducted with four experts, all with experi-ence teaching thermal physics and researching student diffi-culties in thermal physics. The focus group solicited expertperspectives surrounding upper-division thermal physics, in-cluding textbooks, content coverage, learning goals, and ex-isting thermodynamics assessments. Outcomes from the fo-cus group informed several questions included on the sur-vey. For example, participants discussed notational conven-tions as one major challenge for a thermal physics assessment(e.g. the sign convention of work). To address this concern, one question on the survey solicited specific notational issuesworth considering in development of a thermal physics as-sessment. Additionally, textbooks brought up during the fo-cus group comprised the list of textbook options provided onthe survey.To faciliate ease of responses, the survey was a primarilymultiple-response format with only a select set of questionsbeing free-response. Thus, one of the first steps in survey de-velopment was determining which options to provide for var-ious multiple-response questions. We began by investigatingthe scope of thermal physics in texts; we analyzed six thermalphysics texts brought up during the focus group for keycontent coverage. This process involved reviewing each textand identifying topical areas for each based on chapter titles,section headings, and emphasized key terms. Based on thefrequency of topics appearing across the different texts, weclassified topical areas into core topics and supporting top-ics . To put these into an accessible form for use in the survey,topics were sorted and condensed into 29 core topics, mostwith roughly 4 supporting topics (see Table I). For example,the core topic of “thermodynamic laws” had four supportingtopics: 0th law, 1st law, 2nd law, 3rd law. Some core topicshad no supporting topics (e.g. semiconductors) while somehad as many as seven (e.g. energy and thermodynamic po-tentials); the one exception to this was statistical mechanics,which had 14 supporting topics.In addition to focusing on content, and in response to re-cent calls in science education literature for more consider-ation of scientific practices in course materials, assessment,and instruction , the survey also solicited information on thescientific practices valued by respondents in their thermalphysics courses. The list of scientific practices provided onthe survey was pulled from the Next Generation Science Stan-dards (NGSS) list of science and engineering practices . Intheir list, the NGSS combined similar practices together (e.g.developing and using models); however, in upper-divisioncourses, it is less clear that all paired practices would be tar-geted together. Thus, to collect more specific data about indi-vidual practices, paired NGSS practices were split into sepa-rate categories. For example, “developing and using models”was split into “developing models” and “using models” forthe survey.The survey was administered through the survey platformQualtrics and hosted by the University of Colorado Boulder(CU). The survey was divided into 4 major sections: (1) gen-eral course information, (2) content coverage, (3) scientificpractices, and (4) interest in, and concerns about, an upper-division thermal physics assessment. Respondents also hadthe option to identify their institution and submit their coursesyllabus. Additionally, gender and racial identity informationwere collected at the end of the survey.After initial construction of the survey, we solicited feed-back from CU physics faculty who were familiar with teach-ing upper-division thermal physics. Based on these discus-sions, and informed by the frequency of topical areas ap-pearing across the six different analyzed texts, we groupedhe core topics into two categories: assumed core topics andother core topics. Assumed core topics are topics that onemight expect are covered in every thermal physics course:energy and thermodynamic potentials; engines and refriger-ators; entropy; equilibrium; monatomic gases; heat; temper-ature; thermodynamic laws; and work. The survey presentedthese assumed core topics at the beginning of Section (2) ofthe survey, with their supporting topics shown on the samepage. A free-response textbox followed these assumptionsto allow respondents to indicate disagreement with the as-sumptions made. All other core topics were provided on thefollowing page of the survey without their supporting topicsdisplayed. After selecting from the list of other core topics,associated supporting topics for each of the selected core top-ics were displayed on the following page. This conditionalformatting was motivated by the desire to reduce respondentfatigue due to survey length. Survey Distribution:
To ensure the information collectedwas reflective of a broad range of institutions, we collectedcontact information for a large variety of physics degree-granting institutions, including minority serving institutions(MSIs) and women’s colleges, for use in distributing the sur-vey. Institutions were identified using the American Phys-ical Society’s “Top Educators” lists , each of which iden-tifies 16-20 institutions with the highest average number ofphysics bachelors’ degrees awarded by the institution peryear. We also utilized the overall and underrepresented mi-nority (URM) lists for Ph.D.-granting, MS-granting, and BS-granting institutions. Beyond that, we used the AmericanPhysical Society’s MSIs list , which included a list of His-torically Black Colleges and Universities, Black-serving insi-tutions, and Hispanic-serving institutions, to identify all otherphysics-degree-granting MSIs not on the “Top Educators”lists; the MSI list included institutions with both large andsmall physics departments. We also identified women’s col-leges with the “Women in Physics” report produced by theAmerican Institute of Physics . We note that other smallphysics departments (e.g. those that are not Top Educators, orat MSIs or women’s colleges) were not targeted in the initialdistribution of the survey, but will be targeted in the broaderproject moving forward.After identifying institutions, we obtained contact informa-tion of department chairs from physics department websites.We then emailed the survey solicitation to the departmentchairs, with a specific request for the email to be forwarded toall faculty within their department who were currently teach-ing or had previously taught upper-division thermal physics.In addition to department chairs, the research team solicitedthe help of their professional contacts at different institutionsto take the survey or forward it to faculty in their department. III. RESULTS
The survey was open for response collection for three anda half months. During this time, 59 respondents fully com-
FIG. 1. Highest physics degree offered by Minority-Serving Insti-tution (MSI) or Women’s College classification. Bachelor’s degrees(BS), Master’s degrees (MS), and PhDs are indicated. pleted the survey while 2 completed all of the survey exceptquestions regarding scientific practices and assessment. Onlyresponses that completed the sections with core topics andsupporting topics and beyond were used for analysis. Wedo not report response rate, as it is unclear how many peo-ple recieved the solicitation forwarded from their departmentchairs.Racial demographics of respondents included Asian (16%,N=9), Black/African American (2%, N=1), Caucasian (74%,N=43), and Hispanic (2%, N=1); no other racial identitieswere indicated and 7% (N=4) preferred not to answer. Ad-ditionally, 83% (N=48) of respondents were men and 14%(N=8) were women (no other gender identities were indi-cated); 3% (N=2) preferred not to provide their gender. Threerespondents did not provide any demographic information.We collected institutional information, including selectiv-ity, research activity, student population, and highest physicsdegree offered via the Carnegie Classifications and institu-tions’ physics department websites. From the Carnegie Clas-sifications, we identified 70% (N=34) of identifiable institu-tions as being selective or more selective with regards to ad-missions practices, while 31% (N=15) are considered “inclu-sive” institutions. Additionally, 18 schools are classified ashaving high or very high research activity.Overall, we identified 52 unique institutions from the sur-vey, 28 of which were MSIs and/or women’s colleges; oneinstitution could not be identified and one was not in theCarnegie Classifications database. Figure 1 presents institu-tion type by highest physics degree offered and MSI/women’scollege classification. In a few cases, (N=7) institutions wererepresented by 2-3 responses; it was evident from submittedsyllabi and individual item reponses that these were submit-ted by different people. Course Information:
We asked respondents if their coursefocused on thermodynamics, statistical mechanics, or both(thermal physics); 97% (N=59) selected thermal physics andthe remaining 3% (N=2) of responses were split evenly be-tween thermodynamics and statistical mechanics. Most in-stitutions reported one semester of thermal physics (79%,
ABLE I. Response frequency for thermal physics content. The two left columns show data for assumed topics. All assumed core topicsappeared at a frequency of 100%. The right column shows all other core topics. No supporting topics are presented for other core topics.Topics that appeared on syllabi but not the survey (e.g. ensembles and thermodynamic identities) are also not presented.Assumed Topic % Assumed Topic % Other Core Topic %Energy & Thermodynamic Potentials Equilibrium Statistical Mechanics 92
Chemical Potential Thermal Equilibrium
98 Processes 89
Energy Sources Stable & Unstable Equilibrium
41 Diatomic Gases 84
Enthalpy
89 Heat Fermions 84
Equipartition Heat Capacity
100 Blackbody Radiation 82
Free Energy (Gibbs & Helmholtz) Heat Transfer
72 Bosons 80
Internal Energy
Latent Heat
90 Phases 79
Maxwell’s Relations
77 Temperature Kinetic Theory 75Engines & Refrigerators
Absolute Zero
98 Quantum Phenomenon 75
Heat Engines Negative Temperature
69 Pressure Diagrams 72
Refrigerators Thermodynamic Temperature
89 Scaling 71Entropy
Temperature Measurement
59 Magnetism 64
Boltzmann’s Law
90 Thermodynamic Laws Chemical Reactions 54 dS=dQ/T
89 Conduction, Convection, Radiation 53
Entropy & Information
100 Solids 51
TS Diagrams
100 Pure Substances 49Gases
89 Diffusion 46
Ideal Gas Law
100 Work Cooling Techniques 31
Mixtures of Gases Mechanical
98 Fluids 20 van der Waals Interactions Path dependence
84 Semiconductors 12
N=48); some reported two quarters (10%, N=6) or twosemesters (8%, N=5), while a small minority reported onequarter (3%, N=2). The student population was composedof mostly juniors (N=41) and seniors (N=39), though some(N=12) reported sophomores in the course as well.The majority of respondents (72%, N=44) reported using
An Introduction to Thermal Physics by Daniel V. Schroeder . Thermal Physics by Charles Kittel and Herbert Kroemer wasthe second most frequently cited text (16%, N=10). All othertexts appeared at a frequency of 7% or below. Most of theinstructors (74%, N=45) teach with the assumption that theirstudents have little to no prior exposure to thermal physicscontent. Some (N=19) expected familiarity with topics suchas energy, heat, the first and second laws of thermodynam-ics, and the ideal gas law. A few (N=7) said they expectthermal physics exposure from the introductory physics se-quence, though several noted that thermal physics is onlycovered for a few weeks, and sometimes not at all, in thatsequence.These data show most institutions require one semester ofthermal physics, most instructors use Schroeder’s text , andmany instructors assume their students have no prior expo-sure to thermal physics content. These results suggest two im-plications for PER: (1) development of Schroeder-based ther-mal physics assessments and materials could serve many in-structors and institutions, though would still exclude the siz-able population of instructors and institutions who do not use that text; and (2) pretest administration of an upper-divisionthermal physics assessment may not produce meaningfulmeasurements of student understanding of thermal physicscontent prior to taking the course. Key Topical Areas:
Table I shows frequency of assumedsupporting topics and other core topics. All assumed coretopics (see Section II) appeared at a frequency of 100%; thesefrequencies are not reported in Table I. Frequency of support-ing topics is given relative to the number of times the corre-sponding core topic was selected; the frequency of core top-ics is given relative to the total number of valid responses. Wepresent frequencies of all other core topics, but do not presenttheir 56 associated supporting topics or their frequencies dueto space limitations.Four respondents reported teaching thermal physics but didnot select statistical mechanics as a core topic. This resultmay be due to statistical mechanics being covered in theircourse but not seen as a core focus by the respondent; wenote one of these respondents mentioned statistical distribu-tion functions in a textbox but did not select statistical me-chanics as a core topic.These results are relevant for researchers interested inmaterials and assessment development in upper-divisionthermal physics, and can be used to guide content-focifor those endeavors such that they serve a wide range ofinstructors and institutions. cientific Practices:
Of the 16 practices presented on thesurvey, three appeared at a frequency of over 85%: usingmathematical thinking (98%, N=58), asking questions (95%,N=56), and using models (86%, N=51). Review of syllabiindicates the practice of “asking questions” may have beenmisinterpreted; the NGSS practice refers to asking scientificquestions (namely for scientific investigations), but we sus-pect respondents may have interpreted this practice as refer-ring to asking questions about content during class or officehours. The next most frequently appearing practices wereconstructing explanations (70%, N=41), communicating in-formation (64%, N=38), and computational thinking (61%,N=36). The remaining 10 practices appeared at a frequencyof 56% or less.These results highlight at most three scientific practicesthat stand out as valued by nearly all thermal physics in-structors in our sample and demonstrate many other scien-tific practices are less of a universal focus for thermal physicscourses at the upper-division level. Thus, researchers shouldpay particular attention to including opportunities for studentsto demonstrate and develop the practices of using models andusing mathematical thinking in thermal physics-oriented ma-terials and assessments.
Response Consistency:
As a verification of the survey data,we checked for consistency between survey responses andsubmitted syllabi for the 39 responses that provided a syl-labus. We looked at key topics on syllabi and compared withthe associated survey response to ensure topics appearing onthe syllabus also appeared on the survey response. No core orsupporting topics had more than 3 discrepancies when com-paring between survey responses and the 39 syllabi. Discrep-ancies could be due to the amount of focus placed on thosetopics in the course. For example, Bose-Einstein conden-sates may appear on the syllabus but may not be seen as amajor content focus for the instructor when completing thesurvey, resulting in a discrepancy between their syllabus andresponse. Some topics, such as large systems (N=10), inter-acting systems (N=8), and Boltzmann and/or quantum statis-tics (N=9), appeared in syllabi but did not appear as explic-itly named core or supporting topics on the survey. However,those who included topics such as these on their syllabus se-lected other topics on the survey that encompass or require thesame idea, such as multiplicity, thermal equilibrium, and sta-tistical mechanics. Canonical ensembles (N=11) and thermo-dynamic identities (N=6) were the other most common topicsthat appeared on syllabi but were not provided as options onthe survey.This analysis shows that the survey reliably captured thescope of content coverage for most survey responses withoutlarge discrepancies.
Content Variability:
To investigate the claim of content vari-ation across upper-division thermal physics courses, we ex-amined survey responses to see how many topics were se-lected by all instructors. We looked at the three groups oftopics laid out in Table I: assumed core topics, assumed coretopics’ supporting topics, and other core topics. We found that 9/9 (100%) of assumed core topics, 5/32(16%) of assumed supporting topics, and 0/20 (0%) of othercore topics were selected by all respondents. When repeatedwith institutions with multiple responses (e.g. different in-structors at the same institution), we saw an average of 72%of assumed supporting topics and 20% of other core topicschosen by all respondents at a given institution.These results support the anecdotal claim that upper-division thermal physics content coverage varies both acrossinstitutions and between instructors at the same institution(though to a lesser extent). It also makes the case, however,that there are some topics, namely our assumed core topics,that all or most instructors prioritize in their upper-divisionthermal physics courses.
IV. CONCLUSIONS
Our data suggest important considerations for researchersand instructors interested in curricular materials and assess-ment development for upper-division thermal physics. De-spite the demonstrated content variability within thermalphysics, our results point to content-foci, scientific practices,and reference texts that can act as baselines for materials thatcan serve a broad range of institutions and instructors. The re-sults presented here will lay the groundwork for developmentof an upper-division thermal physics assessment. In order forthis assessment to be useful broadly, we carefully and deliber-ately collected data from institutions that serve a wide rangeof student populations. We recommend other researchers in-terested in making widely-available upper-division materialsutilize similar methods in collecting input from a wide rangeof institutions to inform their work. Results from this surveycan inform upper-division thermal physics investigations inPER and the methodology can be reproduced for investiga-tion of the scope of other upper-division physics courses.
ACKNOWLEDGMENTS
This work was supported by funding from the Center forSTEM Learning and the Department of Physics at CU. Wethank S. Pollock and M. Dubson for their input in refiningthe survey, J. T. Laverty for his encouragement of includingscientific practices, and all focus group participants. We alsothank the department chairs who distributed the survey andthe faculty who completed it. We are grateful for their time.
B. W. Dreyfus, B. D. Geller, D. E. Meltzer, and V. Sawtelle,Resource letter TTSM-1: Teaching thermodynamics and statis-tical mechanics in introductory physics, chemistry, and biology,American Journal of Physics , 5 (2015). R. Baierlein,
Thermal Physics (Cambridge University Press,1999). A. Carter,
Classical and Statistical Thermodynamics (PrenticeHall, 2001). C. Kittel and H. Kroemer,
Thermal Physics (W. H. Freeman &Co., 1980). F. Sears and G. Salinger,
Thermodynamics, Kinetic Theory andStatistical Thermodynamics (Addison-Wesley Publishing Co.,1975). D. V. Schroeder,
An Introduction to Thermal Physics (AddisonWesley, 1999). M. Zemansky and D. Dittman,
Heat and Thermodynamics (McGraw-Hili, New York, NY, 1997). National Research Council et al. , A framework for K-12 scienceeducation: Practices, crosscutting concepts, and core ideas (Na-tional Academies Press, 2012). NGSS Lead States,
Next Generation Science Standards: ForStates, By States (The National Academies Press, WashingtonDC, 2013). R. Ivie and K. Stowe, Women in Physics, 2000. AIP Report.(2000).13