Home Search Collections Journals About Contact us My IOPscience

Towards a conceptual diagnostic survey in nuclear physics

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2011 Eur. J. Phys. 32 55

(http://iopscience.iop.org/0143-0807/32/1/006)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 129.79.13.20
This content was downloaded on 21/06/2014 at 14:35

Please note that terms and conditions apply.

IOP PUBLISHING EUROPEAN JOURNAL OF PHYSICS
Eur. J. Phys. 32 (2011) 55–62 doi:10.1088/0143-0807/32/1/006

Towards a conceptual diagnostic

survey in nuclear physics
Antje Kohnle1 , Stewart Mclean1 and Marialuisa Aliotta2
1 School of Physics and Astronomy, University of St Andrews, North Haugh, St Andrews
KY16 9SS, UK
2 School of Physics and Astronomy, University of Edinburgh, JCMB—The King’s Buildings,
Mayfield Road, Edinburgh EH9 3JZ, UK

E-mail: ak81@st-andrews.ac.uk

Received 13 September 2010

Published 8 November 2010
Online at stacks.iop.org/EJP/32/55

Abstract
Understanding students’ prior beliefs in nuclear physics is a first step towards
improving nuclear physics instruction. This paper describes the development
of a diagnostic survey in nuclear physics covering the areas of radioactive
decay, binding energy, properties of the nuclear force and nuclear reactions,
that was administered to students at two institutions in a pre- and post-test
design. The pre-test was given in a free-text entry format, with responses being
used to develop a multiple-choice version that was given as a post-test. We
performed statistical tests to evaluate the reliability and discriminatory power.
Students’ reasoning comments and rated certainties in their responses were used
to determine students’ misconceptions. We give details of misconceptions in
the areas of radioactive decay, binding energy and nuclear density, and discuss
possible underlying reasons for these misconceptions.

1. Introduction

It is well known that pre-existing knowledge and beliefs can strongly influence how new
concepts are understood [1]. If these pre-existing beliefs are inconsistent with current scientific
understanding, they are often called misconceptions, and this term will be used in what follows.
Knowing students’ misconceptions and particular areas of difficulty and possible underlying
reasons for these is an important first step towards tailoring instruction to promote substantial
conceptual change.
Conceptual diagnostic surveys can be a useful tool to assess students’ understanding and to
compare the effectiveness of different types of instruction. Standardized, validated multiple-
choice surveys have been developed, where incorrect choices (so-called distractors) in the
surveys have been designed to reflect common student misconceptions, allowing instructors
0143-0807/11/010055+08$33.00 
c 2011 IOP Publishing Ltd Printed in the UK & the USA 55
56 A Kohnle et al

to determine these latter from survey results. While standardized diagnostic instruments have
been developed for a number of areas of physics including mechanics [2–4], electricity and
magnetism [5, 6] and quantum mechanics [7, 8], only a very limited number of studies have
focussed on nuclear physics [9–12]. Given the ubiquitous applications of nuclear physics
in medicine, analytical and industrial applications and energy production, the ethical debate
surrounding nuclear power and nuclear weapons and its relevance to other areas of physics such
as nuclear astrophysics, it is of prime importance that students come to a correct understanding
of nuclear physics concepts.
As a first step towards improving the nuclear physics curriculum, we have developed a
nuclear physics diagnostic survey, and administered it to students in nuclear physics courses at
two institutions, in a pre- and post-test design. The pre-tests were administered as a free-text
entry survey, with responses being used to develop a multiple-choice version of the survey that
was then administered as a post-test upon completion of the taught component of the course.
This paper is organized as follows: after describing the development of the survey and
analysis methods (section 2), we describe statistical tests to evaluate the survey’s reliability
and discriminatory power, as well as common areas of difficulty and misconceptions found
(section 3). We conclude with a summary and an outlook of future work (section 4).

2. Survey development and analysis methods

The nuclear physics diagnostic survey was created from a pool of questions that we had
developed based on common difficulties we had encountered in our lecturing experience, as
well as from the education research literature and analysis of nuclear physics forum discussions.
The initial survey consisted of 15 questions, covering the topics of radioactive decay
(seven questions), binding energy (two questions), properties of the nuclear force (three
questions) and nuclear reactions (three questions). These questions were chosen as they were
suited to both the St Andrews and Edinburgh courses. Excepting four questions that were
in multiple-choice format, all other questions were free-text responses. For all questions,
students were asked to state their answer, rate their confidence on a scale of 1 (‘certain’) to 4
(‘very uncertain’) and explain their reasoning in a free-text statement. We trialled the survey
with a number of final-year undergraduate students, and this led to minor revisions in the
wording of a number of questions.
The survey was administered to students on nuclear physics courses both in St Andrews
and in Edinburgh. The PH4022 course in St Andrews is an introductory nuclear and particle
physics course taken by physics majors in their third or fourth year of study. The course
assumes that students have taken two intermediate-level quantum mechanics courses as well
as an atomic physics course. There were 49 students enrolled on the module, and all students
had elementary knowledge of nuclear physics from a school course or a first-year university
introductory physics module. At Edinburgh, the survey was administered in the NP4 nuclear
physics course that is offered as an optional module and can be taken either in the fourth (BSc)
or fifth (MPhys) year of study. There were 39 students enrolled on the course, 37 of whom
had taken the introductory subatomic physics course the previous year. All students enrolled
in NP4 were physics majors.
The pre-test was carried out in the first lecture of the PH4022 course at St Andrews,
and during the first course workshop at Edinburgh. Students were given 40–50 min to
complete the survey. We transcribed the responses, certainties and reasoning comments, and
then grouped responses according to common answers for each question. We checked for
ambiguous wording of questions by scrutinizing the reasoning comments. As a result, one of
Towards a conceptual diagnostic survey in nuclear physics 57

the questions on properties of the nuclear force was removed from the post-test questionnaire
as it was deemed to be ambiguous.
We then used common incorrect choices to develop multiple-choice versions of the pre-
test questions for the post-test. We included as many of the responses as possible in the list of
choices, and added additional choices for symmetry wherever sensible. Questions were then
ordered in terms of increasing difficulty, according to the difficulty index determined as the
proportion of correct responses.
The post-test was administered in multiple-choice format, but students were again asked
to rate their certainty and explain their reasoning in a free-text comment for each question.
The post-test was administered in the last workshop slot at Edinburgh. For organizational
reasons, the PH4022 post-test was administered in week 7 of the semester, slightly before the
end of the nuclear physics part of the course. We scrutinized the reasoning comments to check
for possible changes in student reasoning based on the multiple-choice format compared with
the free-text pre-test version of the survey. After the post-test, students at both institutions
were given feedback on the correct answers.
We determined the mean number of correct responses for pre- and post-tests for both
institutions separately, and performed t-tests to test for a significant change in the mean
number of correct responses from pre-test to post-test. For each question, we determined
the gain g = %post − %pre and the normalized gain g = (%post − %pre) / (100 − %pre),
where %pre and %post are the percentages of students with the correct response in the
pre-test and post-test respectively [13]. We tested discriminatory power by determining the
item difficulty index and item discrimination index for each question and averaged over
the survey, as well as Ferguson’s delta for the entire survey [6]. We determined the reliability
by determining the point biserial coefficient for each question and averaged over the survey
[6]. We listed common incorrect choices with high levels of certainty, to determine areas of
student difficulty and misconceptions. We then analysed reasoning comments with the aim of
determining underlying reasons for these difficulties and misconceptions.

3. Student survey outcomes

At St Andrews, 47 students (96%) completed the pre-test and 38 students (78%) completed
the post-test. At Edinburgh, the numbers were 35 (90%) and 17 (44%) respectively. The small
number of students taking the post-test at Edinburgh may have been due to the timing of the
post-test, close to the final-year project report submission date. Thus, comparisons of pre- and
post-test results for Edinburgh may contain systematic biases.
For comparison of pre- and post-test data, we did not take into account the question
removed in the post-test for the pre-test analysis. For St Andrews, the mean number
of questions answered correctly was 5.5 in the pre-test (standard deviation σ = 2.1) and
7.8 (σ = 2.3) in the post-test. For Edinburgh, the numbers were 6.2 (σ = 2.1) for the pre-test and
9.2 (σ = 1.9) in the post-test. Independent t-tests confirmed that the increase in scores from
pre-test to post-test were significant for both St Andrews (t = 3.715, df = 83, p < 0.0005, one-
tailed) and Edinburgh (t = 4.938, df = 50, p < 0.0005, one-tailed). The average normalized
gain from pre-test to post-test was 0.14 for St Andrews and 0.45 for Edinburgh. The low value
for St Andrews may be due to the fact that the post-test was administered somewhat before the
end of the nuclear physics part of the course. Analysing the reasoning comments, we found
no evidence for changes in student reasoning based on the multiple-choice format compared
with the free-text pre-test version of the survey.
Using the post-test results, we calculated the item discrimination index per question and
averaged over the test. The average item discrimination index D was 0.28 for Edinburgh and
58 A Kohnle et al

0.23 for St Andrews. A test is considered to provide good discrimination if D  0.3, so these
values are close to the desired range. The slightly low values arise from a small number
of questions with low discrimination indices, owing to their being either too difficult or too
straightforward. We also determined Ferguson’s delta, a whole-test statistic that measures
discriminatory power by evaluating the breadth of total scores compared with the range of
possible total scores. A test is considered to offer good discrimination if Ferguson’s delta is
0.9. The post-test values for St Andrews and Edinburgh are 0.91 and 0.84 respectively, so
above or very close to the desired value. We determined reliability using the point biserial
coefficient. A value 0.2 is commonly quoted for a reliable test item. Items with coefficents
<0.2 can remain in the test, but there should be few such items. The average point biserial
coefficient for the St Andrews post-test was 0.32, with individual question coefficients ranging
from 0.13 to 0.50. Five questions had coefficients <0.2, all of these being questions with
particularly high or low difficulty indices. We plan to modify these questions in a future
version of the survey.
In the following, we discuss the results of selected individual questions that may represent
widespread misconceptions, focusing on common incorrect choices in pre- and post-tests, and
analysing reasoning comments to elucidate underlying reasons for these responses.

3.1. Radioactive decay law

Question 1 (see the appendix) tests the conceptual understanding of the radioactive decay law.
The correct answer is two half-lives. This result can be obtained by calculation using the
decay law N (t) = N0 exp(−λt) = N0 2−t/T1/2 , where λ is the decay constant and T1/2 is the
half-life. It is however simpler to successively use the definition of a half-life: since only one
quarter of the original nuclei have remained, two half-lives must have elapsed.
Not common, but persistent, incorrect answers were 0.41 (two students in the pre-tests
and five in the post-tests) and 0.5 (seven students in the pre-tests, four in the post-tests).
Common reasoning for the answer 0.5 is ‘half of a half-life’. This reasoning seems to suggest
that N(t) is considered to fall off linearly rather than exponentially, e.g. if 500 atoms have
decayed after 1 half-life, then 250 have decayed after 0.5 half-lives, as well as a confusion
between the number of nuclei that have decayed and the number of nuclei remaining, i.e.
using N(t) = 0.75N0 instead of N(t) = 0.25N0. Students giving the response 0.41 used the
correct decay formula with an exponential decay, but again confused the number of nuclei
that have decayed and the number of nuclei remaining, using N(t) = 0.75N0 instead of N(t) =
0.25N0. In the pre-test, three students only stated ‘less than 1’ as the answer, with reasoning
‘1 half-life: 500’ or similar. Thus, these students seem to know the definition of a half-life but
are unclear how to use the decay law successively. It is noticeable that, of those students with
the correct answer, the great majority argued with the definition of the half-life and did not
use the formula, whereas for students with the incorrect answer, a much greater percentage of
students used the decay law formula.

3.2. Binding energy

Question 3 (see the appendix) was concerned with the mass of a bound state compared with
the mass of two individual constituents. The correct response to this question is that the mass
of the bound state is less than the mass of the constituent particles. The results for this question
are shown in figure 1. Very few students stated that these masses are the same (and none did
so on the post-tests), but a considerable number of students both in pre- and post-tests stated
that the bound state mass was greater than the mass of the individual particles. We analysed
Towards a conceptual diagnostic survey in nuclear physics 59

90
80
70

Percent of students
60 VU
50 SU
40 SC
30 C
20
10
0

E post

E post

E post
SA pre

SA pre

SA pre
SA post

SA post

SA post
E pre

E pre

E pre
same mass smaller mass greater mass

Figure 1. Distribution of pre- and post-test responses to question 3 for both Edinburgh (E) and St
Andrews (SA) students, together with level of certainty (VU = very uncertain; SU = somewhat
uncertain; SC = somewhat certain; C = certain). The correct answer is ‘smaller mass’. Note that
the degree of certainty changes significantly between pre- and post-tests.

the reasoning comments for those students that chose answer C (mass of the bound state is
greater). Most of the comments could be divided into two classes.
• Bound states have extra binding energy which adds to overall mass (58% of comments,
34 comments for choice C in total). Examples of reasoning are ‘the bound state includes
energy between the particles which contributes to the mass’, ‘have to put in binding
energy to bind particles’, ‘due to mass energy conversion. To make a bound state requires
energy which converts to mass’. This may be a false analogy with a mechanical system,
such as bringing two objects close together by hand requiring energy, or considering a
repulsive force between two objects, such as bringing two like charges together.
• In nuclear decay and in fission, energy is released when the nucleus is split into its
constituents. Together with the mass–energy relation, this implies that the mass of the
bound state must be greater. Examples of reasoning include ‘if decay occurs, mass is
converted to energy in E = mc2’ and ‘energy can be liberated via fission of the particles
and mass is energy E = mc2’ (24% of comments, 34 comments for choice C in total).
While energy is indeed liberated in fission, this reasoning neglects to consider (in the
case of induced fission) the kinetic energy of the particle that induces fission, and (for
induced and spontaneous fission) that fission only occurs for heavy nuclei where the
fission fragments have greater binding energy per nucleon than the initial nucleus.
The degree of certainty, averaged over both pre- and post-tests, of students choosing the
incorrect answer C was quite high (1.76, where 1 is ‘certain’ and 2 is ‘somewhat certain’)
and similar to the average certainty of 1.81 of those students choosing the correct answer B.
The pre-test certainty was lower than the post-test certainty, with averages of 1.88 and 1.50
respectively for students choosing C, and 2.26 and 1.37 respectively for students choosing
the correct answer B. Thus, the change in certainty is more pronounced for those students
choosing the correct answer B (see figure 1).

3.3. Electron capture

Question 5 (see the appendix) asked about the fundamental forces of nature responsible for
different decays. In the pre-test results, it was striking that students from both institutions
60 A Kohnle et al

often considered electron capture to be an electromagnetic decay, even though most students
(60% of Edinburgh students, 62% of St Andrews students) correctly identified β + and β −
decays to be governed by the weak force. Of the 11 comments explaining the reasoning of
these students, eight stated that electron capture was due to the attractive electric force between
the positively charged nucleus and the negatively charged electron. Examples of this type of
reasoning are ‘electron capture is due to the capture of an electron in the Coulomb potential
of a nucleus’ and ‘electrons are attracted to the positive nucleus so will be captured if they get
close enough’. Students will be much more familiar with the electromagnetic force than with
the weak force, which may contribute to this error. In our lecturing experience, we have come
across a similar issue in particle physics, where a common error is that quark and anti-quark
in a meson are bound by their electromagnetic attraction and not by the strong force. Students
may also not be considering the neutrino emitted in the decay, which implies that the decay
must be due to the weak force.
In the Edinburgh post-test, 14 of 17 students (82%) chose the correct answer B. However,
two of these students chose both B and C, one of them stating ‘looking at the process, it contains
effects due to both the weak and the electromagnetic force’. One student at St Andrews chose
both B and C in the post-test. These are small student numbers, but may indicate that
instruction does not always lead to a revision of incorrect ideas, but instead to a coexistence
of mutually inconsistent ideas. One of us has found evidence for this elsewhere [14]. In the
St Andrews post-test, 47% of students still chose C, with reasoning again commenting on the
attractive Coulomb force between electron and nucleus. It should be noted, however, that at
the time of the post-test, the four fundamental interactions had not been discussed in detail in
the St Andrews course, but it had been mentioned that neutrinos only interact weakly.

3.4. β − decay

Question 7 (see the appendix) asked for the number of electrons and protons in the daughter
immediately after the β − decay of a 14 −
5 B nucleus. In β , the fundamental decay process is
−
n → p + e + ν e . As the decay leaves the electron shell unchanged, six protons and five
electrons is the correct answer. Common incorrect responses included:
• five protons and four electrons (15 students in the pre-tests, one in the post-tests): common
reasoning is that β − decay is the emission of the electron. Thus, it seems that this response
would be consistent with the electron being emitted from the electron shell and not from
the nucleus. Also, the change in proton number is not considered.
• four protons and four electrons (five students in the pre-tests, one in the post-tests):
common reasoning is that both a proton and an electron are lost in β − decay. Again, this
response seems to imply that the electron is emitted from the electron shell.
• six protons and six electrons (eight students in the pre-tests, 14 in the post-tests): common
reasoning is that charge must be conserved, and noting down the fundamental reaction
n → p +e− +ν e for β − decay. This response does not consider that the daughter is initially
a positive ion. It seems to imply that the electron emitted by the nucleus is captured by
the shell. However, only one reasoning comment explicitly states: ‘assuming the electron
emitted is captured by the atom’.

3.5. Nuclear density

Question 8 (see the appendix) asked about the dependence of nuclear mass density on mass
number, with the correct answer being C. In the pre-test, a substantial fraction of students
(63% of Edinburgh students and 49% of St Andrews students) stated that nuclear mass density
Towards a conceptual diagnostic survey in nuclear physics 61

increases with mass number. In the post-test, these percentages decreased to 41% and 29%
respectively (grouping responses A and B, with B being much more common than A), but are
still quite high. The average certainty of students stating that nuclear density increases with
mass number was 2.5 for the pre-test and 2.7 for the post-test, where 2 is ‘somewhat certain’
and 3 is ‘somewhat uncertain’. Sample reasoning comments from the post-tests are ‘More
nucleons, more tightly bound’, ‘increasing mass number means the nucleus is more and more
dense’. Thus, it seems that these students are drawing a false analogy to atomic systems and
the long-range Coulomb force, and not considering the very different short-range interaction
of the nuclear force.

4. Conclusions and outlook

We have developed a nuclear physics diagnostic survey that covers the areas of radioactive
decay, binding energy, properties of the nuclear force and nuclear reactions. The first step
of this study consisted in devising a free-text survey on specific questions. Responses from
students at two institutions were then used to develop a multiple-choice version of the survey.
This revised version was then used at both institutions as a post-test, after or near completion
of the taught nuclear physics courses. Since incorrect answers found in the free-text responses
were similar at both institutions, we were confident that the multiple-choice distractors
introduced in the post-test would give insight into students’ difficulties and misconceptions.
Students’ reasoning comments entered in the post-test did not appear to have been influenced
by the multiple-choice format. Statistical analysis of the discriminatory power and reliability
of the survey showed that both are satisfactory albeit slightly low, owing to a small number
of questions with very low or very high difficulty index. These questions will be revised in
a future version of the survey to make it more discriminatory and reliable. Certainty ratings
and reasoning comments allowed us to identify a number of misconceptions, with reasoning
comments giving insight into underlying reasons for these misconceptions. Examples of
widespread misconceptions together with possible explanations for their persistence were
given.
In future, we plan to administer the revised survey to a greater number of students,
and aim to use the survey at additional institutions to determine responses for more diverse
student populations. We aim to conduct student interviews to gain more insight into student
thinking. It also remains to be investigated whether well-studied misconceptions in quantum
mechanics, e.g. that a particle tunnelling through a potential barrier loses energy [15, 16],
persist when applied to nuclear physics, such as when considering alpha decay or fission.
However, identifying student misconceptions is only the first step towards an improved nuclear
physics curriculum that explicitly challenges misconceptions, so that students come to a correct
understanding of nuclear physics.

Appendix

Below are selected questions from the survey that are discussed in the text. Instructors can
e-mail the corresponding author to obtain a full version of the survey.
Question 1. Consider a radioactive sample with initially 1000 atoms. How many half-lives
have elapsed when 750 atoms have decayed?
(A) 0.25 (B) 0.41 (C) 0.50 (D) 0.75
(E) 1.3 (F) 2.0 (G) 2.4 (H) 4.0
62 A Kohnle et al

Question 3. How does the mass of a bound state of two particles, A and B, compare with the
sum of the masses mA + mB of the two particles?
(A) the mass of the bound state is identical to mA + mB
(B) the mass of the bound state is less than mA + mB
(C) the mass of the bound state is greater than mA + mB
Question 5. Which of the four known forces of nature is responsible for electron capture?
(A) Strong force (B) weak force (C) electromagnetic force (D) gravity
−
Question 7. A 14
5 B nucleus in a neutral atom undergoes β decay. Immediately after the
decay, how many protons and electrons does the daughter have?
(A) 4p 4e (B) 4p 5e (C) 5p 4e (D) 5p 5e
(E) 5p 8e (F) 6p 4e (G) 6p 5e (H) 6p 6e
Question 8. How does nuclear mass density change with increasing mass number?
(A) it increases exponentially
(B) it increases linearly
(C) it remains constant to a good approximation
(D) it decreases linearly
(E) it decreases exponentially

References
[1] Halloun A and Hestenes D 1985 The initial knowledge state of college physics students Am. J. Phys. 53 1043–55
[2] Hestenes D, Wells M and Swackhammer G 1992 Force concept inventory Phys. Teach. 30 141–58
[3] Hestenes D and Wells M 1992 A mechanics baseline test Phys. Teach. 30 159–66
[4] Beichner R J 1994 Testing student interpretation of kinematics graphs Am. J. Phys. 62 750–62
[5] Maloney D P, O’Kuma T L, Hieggelke C J and Van Heuvelen A 2001 Surveying students’ conceptual knowledge
of electricity and magnetism Am. J. Phys. 69 S12–23
[6] Ding L, Chabay R, Sherwood B and Beichner R 2006 Evaluating an electricity and magnetism assessment tool:
brief electricity and magnetism assessment Phys. Rev. ST Phys. Educ. Res 2 010105-1–7
[7] Cataloglu E and Robinett R W 2002 Testing the development of student conceptual and visualization
understanding in quantum mechanics through the undergraduate career Am. J. Phys. 70 238–51
[8] Wuttiprom S, Sharma M D, Johnston I D, Chitaree R and Soankwan C 2009 Development and use of a conceptual
survey in introductory quantum physics Int. J. Sci. Educ. 31 631–54
[9] Burge E J 1967 Misconceptions in nuclear physics Phys. Educ. 2 184–7
[10] Cooper S, Yeo S and Zadnik M 2003 Australian student’s views on nuclear issues: does teaching alter prior
beliefs? Phys. Educ. 38 123–9
[11] Nakiboglu C and Tekin B B 2006 Identifying students’ misconceptions about nuclear chemistry Chem.
Educ. 83 1712–8
[12] Hernández-Bermejo B 1997 The range of elementary interactions: a criticism of the standard classroom
approach Eur. J. Phys 18 127–8
[13] Hake R R 1998 Interactive-engagement versus traditional methods: a six-thousand-student survey of mechanics
test data for introductory physics courses Am. J. Phys. 66 64–74
[14] Balfour J and Kohnle A 2010 Testing conceptual understanding in introductory astronomy New Directions in
the Teaching of Physical Sciences no 6 (UK: Higher Education Academy UK Physical Sciences Centre)
pp 26–9
[15] McKagan S B, Perkins K K and Wieman C E 2008 Deeper look at student learning of quantum mechanics: the
case of tunneling Phys. Rev. ST Phys. Educ. Res. 4 020103-1–18
[16] Wittmann M C, Morgan J T and Bao L 2005 Addressing student models of energy loss in quantum tunnelling
Eur. J. Phys. 26 939–50