17 Correcting Student Errors

and Misconceptions
Elizabeth J. Marsh and Emmaline Drew Eliseev

Introduction

Overview and History

Mistakes are common in education and sometimes are even encouraged as part of the
learning process. Young children, for example, are often instructed to “invent”
spellings while writing, as this approach leads them to write longer texts (albeit
with more spelling errors; Clarke, 1988). Many textbook chapters prompt students to
answer “prequestions” about material that has not yet been covered, with the goal of
guiding their reading (Pressley et al., 1990). Learning to debug errors in computer
code is considered integral to learning to program (e.g., Klahr & Carver,
1988). In general, the prevailing zeitgeist is for educational practice to encourage
learners to have a growth mindset and to accept mistakes as part of the learning
process (Yeager & Dweck, 2012).
Such practices differ greatly from what was considered ideal practice prior to the
1970s. Perhaps most notably, the behaviorist B. F. Skinner argued strongly against
a “trial-and-error” view of learning in favor of avoiding errors when possible (Skinner,
1953). The emphasis was on errorless learning (Terrace, 1963), meaning that learning
conditions should allow students to learn without ever making a mistake (under the
logic that errors had the potential to interfere with learning the correct responses). This
perspective is an “early selection” model of errors, in that it aims for errors to never
occur in the first place. Such an approach is insufficient for two reasons: First, in at
least some instances, making a mistake may help learning (e.g., Kornell, Hays, &
Bjork, 2009). Second, and more practically, it is virtually impossible to prevent
a learner from ever making mistakes, necessitating a mechanism for error correction
(a “late correction” model).

The Importance of Error Correction

Understanding how to correct errors is crucial, as uncorrected errors have the potential to
be problematic. For example, students who do not receive feedback on their answers on
a multiple-choice test are at risk for later reproducing some of the multiple-choice lures
they selected (for a review, see Marsh et al., 2007). A student who incorrectly selects

437

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 438 elizabeth j. marsh and emmaline drew eliseev

“mice” as the answer to the multiple-choice question “With the increase in

nutrias in the U.S., which animal’s population decreased?: (a) beavers, (b)
mice, (c) moles, (d) muskrats” will be more likely to answer “mice” to
a later short-answer version of the question (“With the increase in nutrias in
the U.S., which animal’s population decreased?”) than if they had never
answered the multiple-choice question at all. This problem occurs with college
students, high school students (Fazio, Agarwal et al., 2010), and even elemen-
tary school children (Marsh, Fazio, & Goswick, 2012). The problem is not
question-specific, in that incorrectly selecting “gravitation” in response to the
question “What biological term describes fish slowly adjusting to water tem-
perature in a new tank?: (a) acclimation, (b) gravitation, (c) maturation, (d)
migration” affects responses to conceptually similar questions, such as
“Animals that thicken their fur during winter are exhibiting what biological
phenomenon?” (Marsh et al., 2007). Fortunately, there is a relatively simple
solution: Tell students the correct answer (Butler & Roediger, 2008) – preview-
ing one of the main messages of this chapter.
Uncorrected errors have consequences beyond believing a single falsehood is true;
conceptual misunderstandings can interfere with new learning of other related infor-
mation. Students enter the classroom with naïve theories about how things work,
based on their own experiences as well as portrayals in film and television, and such
beliefs are thought to impede acquisition of fundamental concepts in fields such as
chemistry (e.g., Nakhleh, 1992), physics (e.g., Carey, 1986), and computer program-
ming (e.g., Clancy, 2004). It is for this reason that teachers often try to anchor new
concepts on familiar objects and systems (e.g., using the solar system to teach about
the atom), to make sure they are using the correct mental model for the problem.

Defining Feedback
Broadly, we consider feedback to be any information that has the potential to
affirm or update a student’s knowledge (as opposed to feedback regarding
student behavior or motivation, for example). Our definition of feedback is
not tied to a particular type of error and it can take many forms, including but
not limited to marks indicating whether an answer is right or wrong, the correct
answer to a problem or question, an explanation of why an answer is the
correct one, or a new way of thinking about a complex system. It can occur
at different grain sizes, providing only the needed correction or contextualizing
the correction within a larger unit (e.g., an entire textbook chapter or lesson).
Feedback could come from another person (such as a teacher), from places such
as the Internet or a book, or even from oneself (such as when one looks up the
answer to check their work). In short, we endorse a broad definition of feed-
back in this chapter, with the different instantiations tied together by the need
to correct something – meaning that even entire lessons may serve as feedback
if the student enters with a misconception.

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 439

Methods for Studying Error Correction

Overview
There are hundreds of demonstrations of the power of feedback. Feedback is
typically considered to be one of the most powerful tools in the teacher’s toolbox,
yielding a large effect size (average Cohen’s d greater than 0.7; Hattie, 2012, 2015).
The benefits of providing feedback are larger than those of other common recom-
mendations to teachers, such as having students spread out learning over time
(spacing; d = 0.60), create concept maps (d = 0.64), review worked examples (d =
0.37), and engage in peer tutoring (d = 0.55) (Hattie, 2015).
One of our favorite experimental examples comes from Pashler and colleagues
(2005) because they separated the effects of feedback on maintaining correct
answers versus correcting errors. That is, they argued that some of the inconsistent
effects in the literature likely occurred because these two types of responses were
treated as the same, when feedback should have less of an impact on correct answers
(given that, by definition, there is nothing to correct). Their participants received two
chances to study a list of twenty Luganda–English word pairs (e.g., leero–today),
before taking a translation test (leero–?). On this initial test, the groups performed
similarly (as expected), translating about 40 percent of words correctly – meaning
that there were plenty of errors to be corrected. Critically, some subjects received no
feedback about their answers, others were told whether each translation was correct
or not, and a third group was told the correct translation of each Luganda word
(additional controls were run to rule out possible confounds but we will not discuss
those further). Of interest was the learner’s ability to maintain their correct
responses, as well as to correct their errors, both relatively immediately and after
a delay.
While participants did forget some of their correct translations over the course of
one week, this effect was similar across feedback conditions. Error correction, in
contrast, depended on the feedback condition: Errors were only corrected after
receiving answer feedback, not correct/incorrect feedback, and this pattern held
both immediately and on a test one week later (although the expected forgetting
happened over the course of a week). These effects were very large – receiving
correct answer feedback following a mistake improved final performance by 494 per-
cent . Such powerful effects are consistent with past reviews of the literature showing
large benefits of feedback.
Pashler and colleagues’ results highlight our first piece of advice: When students
make an error, it is better to provide the correct answer than to simply mark it as
incorrect. We reached a similar conclusion when we tested students on what they
learned from reading short passages describing history, geography, and science
(Fazio, Huelser et al., 2010). Answer feedback is even the better choice when
correcting errors made on a multiple-choice test (Marsh et al., 2012). This finding
with multiple-choice tests is particularly discouraging about the usefulness of right/
wrong feedback, as such feedback does provide information in this case – it allows
the learner to winnow down the remaining choices. But of course this benefit is

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 440 elizabeth j. marsh and emmaline drew eliseev

contingent on learners’ ability to correctly eliminate lures, leading to the finding

shown in Figure 17.1: Following errors on a multiple-choice test, receiving right/
wrong feedback led to intermediate performance on a final test, between the levels
observed following no feedback versus answer feedback. The benefit of right/wrong
feedback dropped as narrowing down the choices became harder (with three remain-
ing choices, for example, as opposed to one) – after being told one made an error on
a four-alternative multiple-choice question, performance was not much higher than if
one had received no feedback at all.
In another study, we examined whether receiving right/wrong feedback (and the
knowledge of what one does vs. does not know) indirectly benefits learning, by
guiding learners’ future study efforts – but, unfortunately, our data did not support
this idea. After reading texts and receiving feedback, readers received a second
chance to read the passages, at their own pace. However, all readers benefited from
rereading, even if they had not received any feedback at all. One possibility is that
right/wrong feedback often does not provide any additional information to the
learner – there are many cases where learners have a good sense of what they do
versus do not know without feedback, such as translations of foreign vocabulary or
simple facts (e.g., Hart, 1967). However, there are also cases where people do not
realize they are making mistakes, something we turn to in the next section of this
chapter.

Errors Made with Confidence

We know intuitively that not all errors are the same – it is one thing to correct an
incorrect translation and another to correct a misunderstanding of why seasons occur.
These two example errors differ in many ways, including but not limited to the

1.0
Proportion of Errors Corrected

.8

.6

.4

.2

.0
Two Three Four
Number of MC Alternatives
No Feedback Correct/Incorrect
Answer

Figure 17.1 Proportion of errors corrected on the final test based on feedback
condition (no feedback, correct/incorrect, answer feedback) and the number of
multiple-choice alternative answer choices (two, three, four) on an initial
multiple-choice test (after Marsh et al., 2012)

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 441

complexity of the error (a simple factual error vs. a conceptual misunderstanding of

a system), confidence in one’s response (which likely will be lower for the failed
translation than one’s faulty explanation of the seasons), content domain (language
vs. science), and so on.
As a starting point, we can examine other errors that are similar to the incorrect
translations in most ways but differ in one key aspect. For example, many people
believe that George Washington had wooden teeth or that Marie Antoinette said “let
them eat cake,” even though neither is true. Structurally, these errors are like the
incorrect translations, in that they are relatively simple paired associates, with the
“George Washington–wooden teeth” association needing to be updated to “George
Washington–dentures made of bone and other nonwood materials.” The difference is
that these misconceptions are likely believed with much higher confidence than
errors made when translating recently learned foreign words. In other words, we can
examine either materials that elicit a range of confidence or a set designed to elicit
high-confidence errors, to examine whether one’s confidence in an error affects one’s
ability to correct it.
Intuitively, one might expect that beliefs held with high confidence would be harder
to correct, as they likely reflect stronger representations in memory. However, numer-
ous studies demonstrate that people are more likely to correct high-confidence errors
than low-confidence ones (the hypercorrection effect; Butterfield & Metcalfe,
2001). Given feedback, most people, for example, are more likely to correct their
misconception that Sydney is the capital of Australia than an error about the capital of
Botswana. This pattern is observed across ages; for example, adolescents are more
likely to correct high-confidence misconceptions about science, such as “The largest
part of the Sahara consists of sand” and “When in a heavy thunderstorm, it is safest to
lie down flat on the ground,” than misconceptions that were held with lower confidence
(Van Loon et al., 2015). Even young children show a higher correction rate for high-
confidence errors than erroneous guesses (Marsh et al., 2012; Metcalfe & Finn, 2012).
One explanation is that people likely have greater confidence when answering
questions about topics they already know something about (i.e., most Americans
know more about Australia than Botswana) and that knowledge supports encoding of
the feedback (for related ideas about how knowledge may support hypercorrection,
see Metcalfe & Finn, 2011). A second explanation involves people’s subjective
feelings when faced with a large discrepancy between their confidence and accu-
racy – a surprising error increases attention to the feedback, with consequent benefits
for memory. This claim is supported by experiments showing that people take longer
to respond to a secondary tone detection task when feedback mismatches their
expectations (suggesting they were distracted by the feedback; Butterfield &
Metcalfe, 2006). Similarly, people are more likely to remember the color of the
feedback when receiving feedback in response to correct guesses and high-
confidence errors (Fazio & Marsh, 2009), supporting the hypothesis that surprising
feedback directs attention toward the feedback.
The classroom implications of these results are less clear; we do not wish
educators to resort to “gimmicky” feedback in order to “surprise” their students.

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 442 elizabeth j. marsh and emmaline drew eliseev

Rather, such results should provide reassurance that answer feedback will be suffi-
cient even in cases where students are confident in their answers.

Misunderstandings
Misunderstanding how echolocation works is a different problem from not being
able to produce the term echolocation in response to a description of how bats
navigate. The first involves misunderstanding a process whereas the second is
a fairly straightforward memory problem, involving forgetting of a specific term.
This difference has implications for what information the feedback should convey;
simply telling someone the answer works in the case of simple errors (as described in
the previous two sections) but may be insufficient with more complex errors.
Many educators and researchers assume that more feedback is better. For example,
educational software programs (especially Intelligent Tutoring systems) often
respond to wrong answers with an explanation of why the answer is wrong (e.g.,
Graesser et al., 2005). Educators provide in-line comments on student essays,
summary statements on student work, and comments on exams (e.g., Tomas,
2014). However, the experimental evidence is mixed as to whether there are added
benefits from providing information beyond the correct answer (for a review, see
Kulhavy & Stock, 1989; for a meta-analysis, see Bangert-Drowns et al., 1991). For
example, students whose multiple-choice selections revealed misconceptions about
science (such as the belief that an individual insect can become immune to pesti-
cides) benefited as much from correct answer feedback as explanations of why their
choices were wrong (i.e., that natural selection operates at the species level, not the
individual level; Gilman, 1969). Similarly, it was just as effective to tell middle
school children the correct answers to factual questions as to have them find the
answers in their text (with line numbers to ensure they could find them; Peeck, 1979).
One issue that makes it hard to draw conclusions across studies is that the “extra
information” added to the feedback takes many forms or, other times, is not specified
in enough detail to evaluate. Students might receive explanations of why answers are
incorrect (e.g., Kulhavy et al., 1985), reread exact texts to discover why their errors
were wrong (Andre & Thieman, 1988), or be directed to look at a particular place in
a text to find the correct information (Peeck, 1979), among other variations. This
issue is highlighted in a study where introductory psychology students were assigned
to one of four conditions, so that after each multiple-choice exam they either received
no feedback, compared their answers with the correct answers written on the board,
listened to the instructor discuss each question, or were directed to reread textbook
passages relevant to the questions they missed (Sassenrath & Garverick, 1965).
Feedback helped performance on later exam questions that tapped retention (ques-
tions that were repeated from the midterm) but it did not matter whether students
received answer feedback or listened to the instructor’s discussion. It is hard to draw
strong conclusions about feedback content from this study, however, because it is not
clear whether or not students interacted during the instructor’s discussion of the
feedback, nor what content was discussed. (Did the teacher focus on elaborating the
correct answer by explaining why some answers were wrong? etc.) Furthermore,

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 443

even the answer feedback condition was unusual, in that it involved self-grading
(comparing one’s answers with those on the blackboard) and as such involved the
students more than simply viewing answer feedback.
Logically, it is not clear why students would need more information than the
answer to improve on a final test containing exactly the same questions as before – in
such a situation, retention is required, not explanation or elaboration of knowledge.
The advantage of elaborated feedback should be greatest on final tests that require
going beyond retention to demonstrate a deeper understanding of the key concepts
and applications of one’s knowledge to novel situations. Returning to the study of
introductory psychology students just discussed, the data hint at this possibility.
In addition to repeating questions from the midterm, the final test included new
questions that were conceptually related to some of the midterm questions, to test
transfer of learning. On these transfer questions, performance was best in the
discussion condition, with checking one’s answers a close second. Again, it is not
clear exactly what the teacher discussed but it seems reasonable to assume that
a word “discussion” would involve more than just providing the answer.
We tested these ideas more directly in our own work, where students learned about
complex scientific processes such as understanding how bread rises or how torna-
does form (Butler, Godbole, & Marsh, 2013). After reading the scientific texts,
students took an open-ended test probing definitions of the concepts; critically,
after each response, students received no feedback, were told the correct answer to
the question, or received the answer in combination with an explanation that had
been presented in the earlier text (Butler et al., 2013). Two days later, students took
a final test that included some of the same definitional questions as on the first test, as
well as novel inference questions. For example, the final test required students to
name the process that facilitates gas exchange in the alveoli (definitional question)
and to explain why breathing pure oxygen helps people who have trouble breathing
(inference question). When faced with the same definitional questions as on the
initial test, students benefited from having received feedback – but it did not matter if
the feedback contained the right answer or an explanation (see Figure 17.2).
In contrast, when faced with novel inference questions, students who had received
explanation feedback outperformed those who had only received answer feedback.
The extra information in the explanation feedback was unnecessary when the test
tapped retention of answers; the explanation feedback was needed when the test
required transfer of knowledge to a new context.

Conceptual Change
Even if a student understands a concept, he or she may struggle to understand how
that concept interacts with other concepts or how to generalize that knowledge to
a new problem. Students may sometimes have an incorrect mental model of
a situation; in this case, what needs to be changed is the larger mental representation
not just a specific fact or concept. We discussed earlier how feedback can be useful in
correcting simple errors and here we will focus on how feedback can be used to
correct a student’s flawed mental representation of broad conceptual information.

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 444 elizabeth j. marsh and emmaline drew eliseev

1.0
.9 No Feedback

.8 Correct Answer Feedback

Proportion Correct
.7 Explanation Feedback
.6
.5
.4
.3
.2
.43 .62 .66 .28 .30 .45
.1
.0
Repeated Transfer
Final Question Type

Figure 17.2 The effect of feedback type (no feedback, correct answer feedback,
explanation feedback) on repeated questions compared with transfer questions
(from Butler, Godbole, & Marsh, 2013, p. 292. Adapted with permission from the
American Psychological Association)

Of course, correcting a misconception of an entire system is unlikely to occur with

the kind of simple feedback described thus far – one cannot encapsulate the entirety
of evolution into a single sentence, for example. Here is where the line begins to blur
between learning and correction; the first time students learn about motion and
friction, for example, should it be called “learning” (as it is the first school lesson
on the topic) or a correction (given that the student enters the classroom with naïve
incorrect beliefs about the concepts)?
For example, consider students’ misunderstanding of emergent properties in
science (Chi, 2005). This example requires an understanding of the differences
between direct and emergent processes, so we review these first. Direct processes
involve a series of sequential stages – examples include the cycles of the moon, the
circulation of blood in the body, and the stages of mitosis (Chi et al., 2012). One stage
is the direct result of an agent, prior process, or stage. In contrast, emergent processes
are nonsequential and based on unconstrained, continuous action (Chi et al., 2012).
The end result emerges from the set of actions but is not caused by any one agent or
action. Examples include osmosis, heat flow, and natural selection. Consider how
a student’s response to a question about diffusion (an emergent process) suggests that
he or she is incorrectly applying a direct process model. When describing the
exchange of CO2 and O2 in the lungs, the student stated: “the capillaries that are in
your lungs would . . . let the oxygen come in through the space in its walls and then
the carbon dioxide would go out . . . because . . . it wants to get out into a lower
concentration, so all the carbon dioxide would want to go through so it would be in
a lower concentration” (Chi, 2005, p. 185). The student uses sequential language,
stating that oxygen first comes in and then carbon dioxide goes out. This student also

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 445

treats diffusion as an intentional process where oxygen wants to get in and carbon
dioxide wants to get out. In reality, CO2 and O2 (and all molecules) are in constant
motion (Brownian motion), moving from areas of high to low concentration – and it
is these collective movements that yield diffusion, not one molecule causing another
to move.
To teach emergent processing, Chi and colleagues created a module that (1)
defined and differentiated the two types of processes, (2) gave everyday examples
of each, and (3) prompted the student to examine how the processes played out in the
everyday examples. For direct processes, students learned about wolf pack hunting
and skyscraper construction; for emergent processes, students learned about schools
of fish and movement in crowds. The examples were chosen to have familiar
structures that students could easily understand. In the third part of the module,
students identified whether the examples fit the criteria of direct versus emergent
properties, as they had learned about in the beginning of the module. For example,
students identified whether all agents had an equal role (indicative of an emergent
process) or not (suggesting a direct process) – for example, noting that the architect’s
role is different from the welder (skyscraper example) but that no single fish drives
the school of fish. To test the effectiveness of this module, Chi and colleagues
assigned 8th and 9th grade students to complete the emergent processing module
or a control science module; all students later completed a module on diffusion and
took a test that tapped standard misconceptions about diffusion. Students who
completed the emergent processing module (the “feedback” targeting the miscon-
ception) endorsed fewer of the misconceptions than did the students who completed
the control module, although questions about the generality of this work remain.
Conceptual change is required to fix misconceptions of many complex processes,
such as understanding how evolution works or why we experience seasons. Many
examples can also be drawn from physics, where many students (and adults!) possess
naïve beliefs about force and motion that are incorrect. For example, many people
struggle to understand the concept of curvilinear motion. When people are shown the
image in (the incomplete version of) Figure 17.3 of a ball on a string spun in a circle
and asked to predict the path of the ball if the string were to break, they often
incorrectly assume the untethered ball would continue moving in a circular pattern.
Although many people struggle with abstract physics problems, people are more
likely to correctly solve such problems if they are framed within a familiar context.
For example, most people have at one point in their lives engaged in a water fight,
spraying others with a garden hose. If Figure 17.3 is interpreted as a person holding
a garden hose (black line), people do not predict that the water will follow the curve
of the hose but, instead, correctly predict that the water will shoot directly out of the
hose (in a straight line, independent of the curvature of the hose). In this and other
examples, the feedback involves drawing students’ attention to familiar past experi-
ences, to help them understand their mistakes in more abstract situations.
This section emphasizes the many different forms feedback can take, from simply
indicating whether an answer is correct or incorrect to an entire learning module.
The challenge is for educators to know what is needed, and when – points we cover in
the next section.

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 446 elizabeth j. marsh and emmaline drew eliseev

Correct
Incorrect

Figure 17.3 The ball and string problem, which tests a person’s understanding of
curvilinear motion (after Kaiser, Jonides, & Alexander, 1986)
This schematic shows correct and incorrect responses (dashed lines) to the ball
and string problem, which tests a person’s understanding of curvilinear motion.

Advice for Educators

Overview
The laboratory is obviously very different from the classroom – and even with
that expectation, we were shocked at the differences we experienced when we
started working in undergraduate engineering classrooms. Our experiments did
not translate in the ways we expected them to – students worked together on
assignments, teachers unknowingly ruined manipulations, and people cheated
(Butler et al., 2014). We can only imagine the challenges involved in working in
classrooms with younger students. We are not surprised that some laboratory
findings do not translate to the classroom; a once-significant effect may be
swamped in the classroom when the experimenter can no longer control for
other factors that carry more weight.
Fortunately, it is clear that receiving feedback benefits students learning authentic
educational materials, although sometimes learning is measured outside of assign-
ments contributing to course grades. The benefits occur with such varied educational
topics as soil ecology (Farragher & Szabo, 1986), army ants (Surber & Anderson,
1975), glaciers (More, 1969), the human eye (Kulhavy, Yekovich, & Dyer, 1976),
and introductory psychology (e.g., Kulhavy & Anderson, 1972). Increased exposure
to feedback matters; for example, the number of optional homework feedback
sessions attended is correlated with final course grades (Núñez-Peña, Bono, &
Suárez-Pellicioni, 2015). Quality of feedback also matters; 6th grade students’
math achievement improved after teachers received training on what to include in
their written feedback to homework (Elawar & Corno, 1985).

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 447

Given that feedback helps, what do educators need to know? What might surprise
them? In the following section, we try to be more specific than simply advising
teachers to “give feedback.”

Know the Few Situations Where Feedback May Be Unnecessary

The title of this section may appear to directly contradict everything we have written
thus far – so it is worth reiterating that most of the time feedback is important to
provide. However, in schools, time is a precious and limited resource – critically,
time spent doing one activity is at the cost of another. It remains an open question
how educators should decide to spend valuable class time, given that the time used to
distribute feedback could be used for other learning activities known to benefit
learning, such as retrieving information from memory (i.e., flashcards, quizzes;
McDaniel et al., 2011) or writing to learn activities (for a meta-analysis, see Bangert-
Drowns, Hurley, & Wilkinson, 2004; see also Klein and Van Dijk, Chapter 8, this
volume).
One situation was alluded to earlier in this chapter: Feedback has relatively little
effect on the maintenance of correct answers, unless they were guesses (Butler &
Roediger, 2008). In contrast, feedback is absolutely crucial for error correction and
should not be skipped. The logical extension of these findings is that feedback may
not be necessary if student performance is uniformly high. This point was captured in
a laboratory study where undergraduates received a fixed amount of time to learn two
lists of Swahili–English translations (Hays, Kornell, & Bjork, 2010). After an initial
study phase, students completed a series of test-feedback trials (akin to going
through a pack of flashcards repeatedly). Feedback viewing was required for one
list (every flashcard had to be turned over), whereas it could be skipped for the other
list, as determined by either the computer or the learner. In this case, skipping
feedback (whether determined by the computer or the student) meant that students
spent more time practicing retrieval (as opposed to reading feedback), which led to
more correct translations a day later (reflecting the known memorial benefits of
retrieving information from memory; e.g., Roediger & Butler, 2011). It should be
noted, however, that participants’ judgments about whether or not to skip feedback
viewing were excellent, with 85 percent of their feedback choices lining up with their
actual performance. As covered in the next section, students are not always so
calibrated at judging their need for feedback, meaning it is likely safer for the teacher
to make the decision that feedback is unnecessary.

Ensure That Students Look at the Feedback

In the section “Errors Made With Confidence”, we argued that surprising feedback
was better attended to, with consequent benefits for later correction (Fazio & Marsh,
2009). Yet it is difficult to directly generalize these results to educational practice –
not all feedback surprises the learner and, in many cases, the teacher has no idea if
students even looked at the feedback, let alone understood it. In the laboratory, we
can address these problems, requiring feedback to stay on the computer screen for

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 448 elizabeth j. marsh and emmaline drew eliseev

a set period of time, instructing students to respond after reading it, or asking students
to make a judgment that shows they processed it for meaning (e.g., Lhyle & Kulhavy,
1987). However, educators normally do not have the option of forcing students to
spend a set amount of time reading feedback, given that time spent on review means
less time for new learning activities.
Furthermore, attention is not guaranteed, even in a relatively captive classroom
given that laptops distract (Fried, 2008) and minds wander (see review by Szpunar,
Moulton, & Schacter, 2013). The laboratory likely underestimates the challenges of
directing a learner’s attention to processing feedback in authentic educational set-
tings, where feedback processing is often left to the discretion of students. We have
all seen students recycle their commented papers as they walk out the classroom
door, or experienced empty office hours when no students stop by to view their
exams (which were not returned in order to protect a bank of test questions). Existing
data confirm these anecdotal impressions. For example, one set of researchers took
advantage of the practice of allowing third-year medical students to submit a self-
addressed stamped envelope (SASE) so feedback on their essay could be mailed to
them. This essay was a five-page review piece and a passing grade was required to
advance to the next grade. Unfortunately, less than half of the students provided the
SASE, meaning most of the students did not receive any feedback on their written
work (Sinclair & Cleland, 2007). Perhaps most discouraging, students who likely
needed the feedback the most (the ones with lower course grades) were the least
likely to provide an envelope so that the feedback could be mailed.
In our study in an engineering classroom, we used an online homework system
that allowed us to automatically collect records of whether or not individual students
clicked on problem-by-problem feedback (Mullet et al., 2014). This system also
made it possible to require feedback viewing, if desired. In one section of the course,
feedback viewing was required and counted toward the class grade. In the other
section, feedback viewing was optional, as is the norm in most college courses.
The results were striking: When feedback viewing was required, 94 percent of
students clicked on the links. In contrast, students in the feedback-optional section
only clicked on the feedback links for 47 percent of the problems. For a given
problem, students viewed the feedback sooner and more frequently in the feedback-
required section than in the optional section. These different behaviors were asso-
ciated with differences in performance on the course exams – students who had been
required to view the feedback answered 10 percent more exam questions correctly
than did the students for whom feedback viewing was optional.
Why did students sometimes fail to access the feedback provided to them? One
possibility is that students simply punt on tasks that do not contribute directly to their
grades. A second possibility is that students may not always be calibrated enough to
know when to look at the feedback. It is entirely possible that the students in the
engineering classroom study believed they looked at the feedback when they really
needed it – which would suggest a metacognitive problem, not a laziness problem.
Both are likely involved, but here we focus on the metacognitive issue.
Correctly skipping feedback depends critically on people’s awareness of when
they made mistakes. In the Hays and colleagues (2010) study discussed in the

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 449

previous section, students benefited from the ability to skip feedback – but those
students were very good at knowing which Swahili words they could not translate,
with 85 percent of their feedback choices lining up with their actual performance.
We cannot assume similar calibration of learning with more complex materials.
It may be obvious when one does not know the translation of a foreign word but
much harder to judge the quality of one’s essay or whether a math problem was
solved correctly. Second, students in that study likely benefited from skipping feed-
back because they replaced that time with another learning activity – skipping
feedback on some trials allowed them to spend more time retrieving other informa-
tion. We do not have any data on this point but we doubt that the students in our
feedback-optional section used the time they saved on some other additional activity
for the class.

Ensure That Students Process the Feedback Correctly

Our engineering students benefited from clicking on the feedback links – but it
should be noted that we do not know if they actually read the content, whether they
thought about it, and so on. In some instances, additional steps may be necessary to
ensure that students actually process the feedback. For example, in one study,
experimenters attempted to leverage the known benefits of active processing (e.g.,
Slamecka & Katsaiti, 1987) by requiring students to unscramble the words in the
feedback message (that is, the correct multiple-choice alternative was presented in
scrambled format). This strategy only helped when the experimenters added a task to
make sure that students actually processed the feedback for meaning, in this case, by
requiring them to write out the correct unscrambled version of the feedback (Lhyle &
Kulhavy, 1987).
More generally, feedback will fail if students do not understand it or fail to
understand how it contradicts their own answer. This may be particularly important
in situations where people are self-grading or peer-grading – if they do not apply
a rubric correctly, they will not realize an error has been made. In one study, students
were asked to define terms such as “the fundamental attribution error” and then
compare their answers to correct definitions (Rawson & Dunlosky, 2007).
The critical finding is that sometimes students could not understand how their
incorrect answer differed from the correct one. A similar finding has been reported
with example generation, with students asked to generate examples of concepts
failing to understand how their answers fell short when provided with feedback
(Zamary, Rawson, & Dunlosky, 2016).

Don’t Assume That Corrections Are Permanent

There is no “mastery” level that guarantees information will always be known – for
example, we have all experienced tip-of-the-tongue states whereby something we
know is currently “out of reach.” More generally, forgetting increases as the cues in
the environment change and as time passes since learning and using a piece of
information. The most well-known educational example of this problem is the so-

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 450 elizabeth j. marsh and emmaline drew eliseev

called summer learning loss whereby students appear to lose large amounts of
information over summer vacation (Cooper et al., 1996). One issue involves how
to reactivate that previously learned information; another problem involves the
potential resurgence of an earlier error.
First, consider the problem whereby a student has forgotten the meanings of core
concepts from a class. In at least some cases, the simple solution involves simple
reexposure to the material – the same way a tip-of-the-tongue state is often resolved
when someone takes pity on the speaker and fills in the missing word. In our own
work, we have shown that a multiple-choice test can serve the same function; after
knowledge about the Treaty of Versailles is forgotten, solving the multiple-choice
question “What peace treaty ended World War I?” is sufficient to reactivate it (while
also providing formative assessment to the teacher). One problem involves identify-
ing which information should be revisited; when we investigated these issues in
a pharmacology classroom, we relied on the instructor to identify foundational
material that students were supposed to know from prerequisite coursework
(Butler et al., under review). For three of the course’s six units, students answered
multiple-choice questions on related foundational material (which units were
assigned to the intervention versus control were counterbalanced across subjects).
Pretesting indicated that students were unable to produce 75 percent of the founda-
tional material at the start of the course (even though it all had been covered in
prerequisites for the course); multiple-choice testing (without feedback) led to
significant improvement on a later test of that foundational material. To our knowl-
edge, this is the first demonstration of stabilizing access to foundational knowledge
in a classroom setting but there is still much to be learned about how to identify
which knowledge needs to be reactivated and how often – a point we return to at the
end of this chapter.
The second problem involves the resurgence of errors. Spontaneous recovery of
undesirable behaviors is common – the drinker who abstained from alcohol can “fall
off the wagon,” a frequent flyer may become anxious after numerous smooth flights,
and long-debunked beliefs (i.e., that the world is flat) become popular again.
In education, this problem is particularly problematic when dealing with confidently
held misconceptions. Earlier in this chapter, we discussed how people are more
likely to correct high-confidence errors such as “Sydney is the capital of Australia”
than erroneous guesses, given feedback. This result is surprising because confidently
held errors would be expected to be harder to correct – and yet increased attention to
the feedback leads to a higher correction rate. But what happens after time has
passed? Forgetting of the recent event (the feedback) will occur at a faster rate than
forgetting of the misconception (this is Jost’s Law; Jost, 1897) – meaning that, at
some point in time, the misconception will be stronger in memory than the feedback.
A week later, it was errors such as “Sydney is the capital of Australia” that were most
likely to reappear (Butler, Fazio, & Marsh, 2011).
Finally, vestiges of misconceptions may remain even though students show
knowledge of the correct information. For example, young children often believe
that the ability to move is a prerequisite for life, a belief that is at odds with plants
being alive. The remnant of this belief shows up when undergraduates are asked to

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 451

quickly classify a series of items as living versus nonliving; they are slower and less
accurate to classify plants than animals, even after controlling for word frequency
(Goldberg & Thompson-Schill, 2009). Even PhDs in biology show a similar effect
when told to make their decisions quickly, despite having spent an average of twenty
years as biology faculty. Under the right circumstances, an ingrained false belief may
coexist with its correction. Similarly, even though most people know that density
determines whether an object floats or sinks in water, they are slower to make
decisions when mass and density are not positively correlated, reflecting the child-
hood tendency to relate buoyancy and size (Potvin et al., 2015).

Student Preferences Do Not Always Align with the Best Learning

Strategies
Student beliefs about learning are notoriously incorrect. For example, many students
prefer to use inefficient study strategies such as rereading their textbooks and rewriting
their notes instead of engaging in self-testing (Karpicke, Butler, & Roediger, 2009).
Cramming is another popular, yet ineffective, study technique often used by students.
In fact, students believe they learn more when their study sessions are massed together
even though this is not true – they actually learn more when they spread out their study
sessions (Kornell, 2009). These metacognitive illusions likely occur because easy
practice feels good, even if it does not promote long-term learning. That is, rereading
a text is easier than reading it for the first time, and a translation is easily retrieved if
I just tested myself with the same flashcard two trials ago – tricking the learner into
believing they are good study strategies (see Bjork & Schmidt, 1992). It matters what
students like, as their preferences are correlated with their teacher ratings (Beleche,
Fairris, & Marks, 2012) which in turn often play a role in how teachers are evaluated in
their jobs.
Students tend to like feedback that includes written comments, although the need
for them likely varies depending on the nature of the comment and the type of to-be-
learned information. While students value written feedback, not all written feedback
is viewed as helpful (Weaver, 2006). A common complaint of students is that the
written feedback they received on their assignments was unclear, confusing, and
needed to include more details in order for students to understand how to improve
(Ferguson, 2011; Walker, 2009). Earlier we discussed how feedback need not always
contain an explanation of why a particular answer is right or wrong – but students
like to receive feedback that explains their grade. For instance, one student com-
plained that there were “not enough comments to justify the grade given” (Ferguson,
2011, p. 57). Specifically, students prefer feedback that contains comments on the
quality of their main ideas rather than specific details. Unsurprisingly, students like
to see positive comments to build their confidence included with more critical
feedback indicating how to improve their work (Ferguson, 2011).
A second misconception of students is that feedback should be provided as soon as
possible (of course, this preference likely also reflects their desire to know their
grades). This belief is not limited to students; it occurs in contemporary educational
programs and is often considered a positive feature of an assessment if feedback can

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 452 elizabeth j. marsh and emmaline drew eliseev

be immediate – Coursera touts immediate feedback to student responses in massive

open online courses (MOOCs) and teachers can purchase the Immediate Feedback
Assessment Technique (IF-AT) testing system (where students uncover a star if they
scratch off the correct answer to a multiple-choice question; Epstein, Epstein, &
Brosvic, 2001; Epstein & Brosvic, 2002). This belief traces back to B. F. Skinner,
who showed that animals required immediate reinforcement to learn an association
between a lever press and a food reward (see review by Renner, 1964). However, the
results with humans are actually quite mixed, with some studies (mostly classroom
studies) showing a benefit of immediate feedback whereas others (mostly laboratory
studies) show a benefit of delaying feedback (for meta-analysis, see Kulik & Kulik,
1988). Despite the ambiguous success of immediate feedback, students show
a strong preference of using the IF-AT where they receive immediate corrective
feedback over traditional multiple-choice testing (Dibattista, Mitterer, & Gosse,
2004). On written assignments, university students preferred feedback to be returned
in two to three weeks, as long as it was before the start of the next assignment
(Ferguson, 2011).
In our own work on this issue, we used an online homework system that allowed
us to carefully control the timing of feedback (Mullet et al., 2014). Students in upper-
level undergraduate engineering classes received identical feedback on their home-
work assignments; the only difference was whether the feedback was delivered
immediately after the homework deadline or delayed by one week. Grades on course
exams were higher following delayed feedback, even after controlling statistically
for the shorter retention interval. However, students failed to recognize the benefits
of delayed feedback. When asked which feedback schedule they preferred and which
one was more effective, the overwhelming majority of students reported a strong
preference for immediate feedback. This was true regardless of whether they experi-
enced both schedules of feedback within their course (a within-subjects design) or
only experienced one of the two schedules (a between-subjects design). Thus, there
was a metacognitive disconnect between the feedback timing schedule that students
preferred and what actually helped them to learn.
Given the general importance of student ratings in teacher evaluations, we under-
stand why teachers might be loath to implement a strategy that promotes learning but
is almost universally disliked. One possibility is to explain upfront to students the
reasons for one’s choices (which is possible in practice, albeit not in a controlled
experiment) – in teaching, we have found that students interpret delayed feedback as
evidence that the teacher does not care or is procrastinating; it may not be possible to
completely manage this impression but explaining the reasoning will not hurt and
has the potential to help.

Consider Students’ Preexisting Beliefs and Motivations to Change

From the instructor’s perspective, feedback consists of relatively neutral information
(other than the negative affect associated with making a mistake). However, learners
enter the classroom with preexisting beliefs and differ in their motivation to change
them. No one is surprised by a lack of change when one simply tells a climate change

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 453

doubter to believe science on global warming – a different approach is needed. Even

in a study with relatively neutral statements (e.g., telling students that they were
wrong for believing that bulls become enraged by the color red), students indicated
that they did not believe all of the feedback. That is, the more confident students were
in their initial responses, the less likely they were to believe the feedback, as rated on
a 0 (do not believe the feedback at all) to 100 (absolutely believe the feedback) scale
(Rich et al., 2017). Belief in the feedback, in turn, was related to later corrections of
the errors.
This type of situation is also one where explanation feedback can help. It makes
sense that when someone is motivated to believe something, he or she will need more
evidence to reject it. Accordingly, participants who received both the right answer
and the explanation behind it were more successful at correcting their misconcep-
tions than participants who only received feedback with the correct answer (Rich
et al., 2017). That is, receiving the following feedback, “The color red does not
enrage bulls (the correct answer) because bulls do not see the color red, and, instead,
attack because they perceive the matador as a threat (the explanation)” (Rich et al.,
2017, p. 492), is more likely to increase your belief in the feedback and help you
correct your misconception than only receiving “The color red does not enrage bulls
(the correct answer)” (Rich et al., 2017, p. 492). When tackling students’ mistaken
beliefs, feedback is only helpful in correcting misconceptions if it is believed.
Providing explanations in addition to the correct answer is a good way to refute
the misconception and help students accept the feedback as true.

Conclusions

Open Questions
One open question involves the value of personalized feedback. Personalized learn-
ing is, in general, a hot topic in education – it is very appealing to think about
adjusting the curriculum and feedback to a particular learner. While there are many
commercial products touting personalized learning, in most cases we know little
about the nature of the underlying algorithms. From the academic perspective, the
best evidence for personalized learning comes from work looking at the effects of
different practice schedules on the retention of Spanish vocabulary words (Lindsey
et al., 2014). Grade 8 students practiced Spanish vocabulary via an online flashcard
(retrieval followed by feedback) tutoring system called the Colorado Optimized
Language Tutor (COLT). In a standard practice condition, practice was massed –
one-third of the words were assigned to be translated chapter-by-chapter, so that
older chapters were never revisited as the class progressed through the book.
In a spaced condition, the scheduler had one-third of the words revisited later in
the course – practice was spread out over time. In a third condition, analytics were
used to individually determine the practice schedule for one-third of the words; the
analytics drew on both the learner’s data and a large amount of data about past
learners to make predictions about which items needed the most practice. Briefly, the

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 454 elizabeth j. marsh and emmaline drew eliseev

results showed that children learned the words in the personalized spacing condition
best, followed by those in the spaced condition, with the worst retention of words in
the standard practice condition (especially those that had occurred early in the
course).
Feedback was provided in all three COLT learning conditions, so the current data
do not tell us anything about how feedback should be scheduled. However, some
evidence suggests that the temporal spacing of feedback is similar to that of retrieval
(Smith & Kimball, 2010). That is, a fair amount of evidence suggests that the interval
between learning opportunities should be between 5 percent and 20 percent of the
desired retention interval (Cepeda et al., 2008) – and this same formula works for the
timing of feedback.

Final Thoughts
Vague advice to educators can be harmful rather than helpful if principles are
implemented in a way that changes the processing involved. One concern is that
we are simply asking too much of teachers, who are already often following
mandated curricula – especially since we suspect many of the chapters in this volume
are providing other pieces of advice. However, we encourage teachers to keep in
mind (1) the benefits of delaying feedback and (2) the resurgence of errors over time.
To the extent that new topics build on old ones, there may be a natural spacing of
feedback over time. The teacher can also watch for any indicators that students are
regressing, which could trigger a need for review. More generally, we encourage
teachers to think more flexibly about the definition of feedback. Feedback could take
the form of a multiple-choice quiz (Cantor et al., 2015), a student presentation, or an
in-class review game – it does not have to take the form of responses to a test or other
graded assignment.

References
Andre, T. & Thieman, A. (1988). Level of adjunct question, type of feedback, and learning
concepts by reading. Contemporary Educational Psychology, 13(3), 296–307.
Bangert-Drowns, R. L., Hurley, M. M., & Wilkinson, B. (2004). The effects of school-based
writing-to-learn interventions on academic achievement: A meta-analysis. Review
of Educational Research, 74(1), 29–58.
Bangert-Drowns, R. L., Kulik, C. L. C., Kulik, J. A., & Morgan, M. (1991). The instructional
effect of feedback in test-like events. Review of Educational Research, 61(2),
213–238.
Beleche, T., Fairris, D., & Marks, M. (2012). Do course evaluations truly reflect student
learning? Evidence from an objectively graded post-test. Economics of Education
Review, 31(5), 709–719.
Bjork, R. A., Dunlosky, J., & Kornell, N. (2013). Self-regulated learning: Beliefs, techniques,
and illusions. Annual Review of Psychology, 64(1), 417–444.
Butler, A. C., Black-Maier, A. C., Campbell, K., Marsh, E. J., & Persky, A. M. (under review).
Stabilizing access to marginal knowledge in a classroom setting.

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 455

Butler, A. C., Fazio, L. K., & Marsh, E. J. (2011). The hypercorrection effect persists over
a week, but high-confidence errors return. Psychonomic Bulletin and Review, 18(6),
1238–1244.
Butler, A. C., Godbole, N., & Marsh, E. J. (2013). Explanation feedback is better than correct
answer feedback for promoting transfer of learning. Journal of Educational
Psychology, 105(2), 290–298.
Butler, A. C., Karpicke, J. D., & Roediger, H. L. III (2007). The effect of type and timing of
feedback on learning from multiple-choice tests. Journal of Experimental
Psychology: Applied, 13(4), 273–281.
Butler, A. C., Marsh, E. J., Slavinsky, J. P., & Baraniuk, R. G. (2014). Integrating cognitive
science and technology improves learning in a STEM classroom. Educational
Psychology Review, 26(2), 331–340.
Butler, A. C. & Roediger, H. L. (2008). Feedback enhances the positive effects and reduces
the negative effects of multiple-choice testing. Memory and Cognition, 36(3),
604–616.
Butterfield, B. & Metcalfe, J. (2001). Errors committed with high confidence are
hypercorrected. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 27(6), 1491–1494.
(2006). The correction of errors committed with high confidence. Metacognition and
Learning, 1(1), 69–84.
Cantor, A. D., Eslick, A. N., Marsh, E. J., Bjork, R. A., & Bjork, E. L. (2015). Multiple-choice
tests stabilize access to marginal knowledge. Memory and Cognition, 43(2),
193–205. http://dx.doi.org/10.3758/s13421-014–0462-6
Carey, S. (1986). Cognitive science and science education. American Psychologist, 41(10),
1123–1130.
Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in
learning a temporal ridgeline of optimal retention. Psychological Science, 19(11),
1095–1102.
Chi, M. T. (2005). Commonsense conceptions of emergent processes: Why some misconcep-
tions are robust. The Journal of the Learning Sciences, 14(2), 161–199.
Chi, M. T., Roscoe, R. D., Slotta, J. D., Roy, M., & Chase, C. C. (2012). Misconceived causal
explanations for emergent processes. Cognitive science, 36(1), 1–61.
Clancy, M. (2004). Misconceptions and attitudes that interfere with learning to program.
In S. Fincher & M. Petre (eds.), Computer science education research (pp. 85–100).
CRC Press.
Clarke, L. K. (1988). Invented versus traditional spelling in first graders’ writings: Effects on
learning to spell and read. Research in the Teaching of English, 22(3), 281–309.
Cooper, H., Nye, B., Charlton, K., Lindsay, J., & Greathouse, S. (1996). The effects of
summer vacation on achievement test scores: A narrative and meta-analytic
review. Review of Educational Research, 66(3), 227–268.
Dibattista, D., Mitterer, J. O., & Gosse, L. (2004). Acceptance by undergraduates of the
immediate feedback assessment technique for multiple-choice testing. Teaching in
Higher Education, 9(1), 17–28.
Elawar, M. C. & Corno, L. (1985). A factorial experiment in teachers’ written feedback on student
homework: Changing teacher behavior a little rather than a lot. Journal of Educational
Psychology, 77(2), 162–173. https://doi.org/10.1037/0022–0663.77.2.162

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 456 elizabeth j. marsh and emmaline drew eliseev

Epstein, M. L. & Brosvic, G. M. (2002). Students prefer the immediate feedback assessment
technique. Psychological Reports, 90(3, suppl.), 1136–1138.
Epstein, M. L., Epstein, B. B., & Brosvic, G. M. (2001). Immediate feedback during academic
testing. Psychological Reports, 88(3), 889–894.
Farragher, P. & Szabo, M. (1986). Learning environmental science from text aided by
a diagnostic and prescriptive instructional strategy. Journal of Research in
Science Teaching, 23(6), 557–569.
Fazio, L. K., Agarwal, P. K., Marsh, E. J., & Roediger, H. L., III (2010). Memorial con-
sequences of multiple-choice testing on immediate and delayed tests. Memory and
Cognition, 38(4), 407–418. http://dx.doi.org/10.3758/MC.38.4.407
Fazio, L. K., Huelser, B. J., Johnson, A., & Marsh, E. J. (2010). Receiving right/wrong
feedback: Consequences for learning. Memory, 18(3), 335–350.
Fazio, L. K. & Marsh, E. J. (2009). Surprising feedback improves later memory. Psychonomic
Bulletin & Review, 16(1), 88–92.
Ferguson, P. (2011). Student perceptions of quality feedback in teacher education. Assessment
and Evaluation in Higher Education, 36(1), 51–62. https://doi.org/10.1080/
02602930903197883
Fried, C. B. (2008). In-class laptop use and its effects on student learning. Computers and
Education, 50(3), 906–914.
Gilman, D. A. (1969). Comparison of several feedback methods for correcting errors by
computer-assisted instruction. Journal of Educational Psychology, 60(6, Pt. 1),
503–508.
Goldberg, R. F. & Thompson-Schill, S. L. (2009). Developmental “roots” in mature biological
knowledge. Psychological Science, 20(4), 480–487.
Graesser, A. C., Chipman, P., Haynes, B. C., & Olney, A. (2005). AutoTutor: An intelligent
tutoring system with mixed-initiative dialogue. IEEE Transactions on Education,
48(4), 612–618.
Hart, J. T. (1967). Second-try recall, recognition, and the memory-monitoring process.
Journal of Educational Psychology, 58(4), 193–197.
Hattie, J. (2012). Visible learning for teachers: Maximizing impact on learning. New York:
Routledge.
(2015). The applicability of visible learning to higher education. Scholarship of Teaching
and Learning in Psychology, 1(1), 79–91. https://doi.org/10.1037/stl0000021
Hays, M. J., Kornell, N., & Bjork, R. A. (2010). The costs and benefits of providing feedback
during learning. Psychonomic Bulletin and Review, 17(6),797–801.
Jost, A. (1897). Die Assoziationsfestigkeit in ihrer Abha¨ngigkeit von der Verteilung der
Wiederholungen [The strength of associations in their dependence on the distribu-
tion of repetitions]. Zeitschrift fur Psychologie und Physiologie der Sinnesorgane,
16, 436–472.
Kaiser, M. K., Jonides, J., & Alexander, J. (1986). Intuitive reasoning about abstract and
familiar physics problems. Memory and Cognition, 14(4), 308–312.
Karpicke, J. D., Butler, A. C., & Roediger III, H. L. (2009). Metacognitive strategies in
student learning: Do students practise retrieval when they study on their own?.
Memory, 17(4), 471–479.
Klahr, D. & Carver, S. M. (1988). Cognitive objectives in a LOGO debugging curriculum:
Instruction, learning, and transfer. Cognitive Psychology, 20(3), 362–404.

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 457

Kornell, N. (2009). Optimising learning using flashcards: Spacing is more effective than
cramming. Applied Cognitive Psychology, 23(9), 1297–1317. https://doi.org/
10.1002/acp.1537
Kornell, N., Hays, M. J., & Bjork, R. A. (2009). Unsuccessful retrieval attempts enhance
subsequent learning. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 35(4), 989–998. https://doi.org/10.1037/a0015729
Kulhavy, R. W. & Anderson, R. C. (1972). Delay-retention effect with multiple-choice tests.
Journal of Educational Psychology, 63(5), 505–512. https://doi.org/10.1037/
h0033243
Kulhavy, R. W. & Stock, W. A. (1989). Feedback in written instruction: The place of response
certitude. Educational Psychology Review, 1(4), 279–308.
Kulhavy, R. W., White, M. T., Topp, B. W., Chan, A. L., & Adams, J. (1985). Feedback
complexity and corrective efficiency. Contemporary Educational Psychology, 10
(3), 285–291.
Kulhavy, R. W., Yekovich, F. R., & Dyer, J. W. (1976). Feedback and response confidence.
Journal of Educational Psychology, 68(5), 522–528.
Kulik, J. & Kulik, C. (1988). Timing of feedback and verbal learning. Review of Educational
Research, 58(1), 79–97. www.jstor.org/stable/1170349
Lhyle, K. G. & Kulhavy, R. W. (1987). Feedback processing and error correction. Journal of
Educational Psychology, 79(3), 320.
Lindsey, R. V., Shroyer, J. D., Pashler, H., & Mozer, M. C. (2014). Improving students’ long-
term knowledge retention through personalized review. Psychological Science, 25
(3), 639–647.
Maier, U., Wolf, N., & Randler, C. (2016). Effects of a computer-assisted formative assess-
ment intervention based on multiple-tier diagnostic items and different feedback
types. Computers and Education, 95, 85–98.
Marsh, E. J., Fazio, L. K., & Goswick, A. E. (2012). Memorial consequences of testing
school-aged children. Memory, 20(8), 899–906.
Marsh, E. J., Lozito, J. P., Umanath, S., Bjork, E. L., & Bjork, R. A. (2012). Using verification
feedback to correct errors made on a multiple-choice test. Memory, 20(6), 645–653.
Marsh, E. J., Roediger, H. L., Bjork, R. A., & Bjork, E. L. (2007). The memorial consequences
of multiple-choice testing. Psychonomic Bulletin and Review, 14(2), 194–199.
McDaniel, M. A., Agarwal, P. K., Huelser, B. J., McDermott, K. B., & Roediger, H. L., III,
(2011). Test-enhanced learning in a middle school science classroom: The effects of
quiz frequency and placement. Journal of Educational Psychology, 103(2),
399–414.
Metcalfe, J. & Finn, B. (2011). People’s hypercorrection of high-confidence errors: Did they
know it all along? Journal of Experimental Psychology: Learning, Memory, and
Cognition, 37(2), 437–448.
(2012). Hypercorrection of high confidence errors in children. Learning and Instruction, 22
(4), 253–261.
More, A. J. (1969). Delay of feedback and the acquisition and retention of verbal materials in
the classroom. Journal of Educational Psychology, 60(5), 339–342. https://doi.org/
10.1037/h0028318
Mullet, H. G., Butler, A. C., Verdin, B., von Borries, R., & Marsh, E. J. (2014). Delaying
feedback promotes transfer of knowledge despite student preferences to receive
feedback immediately. Journal of Applied Research in Memory and Cognition, 3(3),
222–229.

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 458 elizabeth j. marsh and emmaline drew eliseev

Nakhleh, M. B. (1992). Why some students don’t learn chemistry: Chemical misconceptions.
Journal of Chemical Education, 69(3), 191–196.
Núñez-Peña, M. I., Bono, R., & Suárez-Pellicioni, M. (2015). Feedback on students’ perfor-
mance: A possible way of reducing the negative effect of math anxiety in higher
education. International Journal of Educational Research, 70, 80–87.
Pashler, H., Cepeda, N. J., Wixted, J. T., & Rohrer, D. (2005). When does feedback facilitate
learning of words? Journal of Experimental Psychology: Learning, Memory, and
Cognition, 31(1), 3–8.
Peeck, J. (1979). Effects of differential feedback on the answering of two types of questions by
fifth-and sixth-graders. British Journal of Educational Psychology, 49(1), 87–92.
Potvin, P., Masson, S., Lafortune, S., & Cyr, G. (2015). Persistence of the intuitive conception
that heavier objects sink more: A reaction time study with different levels of
interference. International Journal of Science and Mathematics Education, 13(1),
21–43.
Pressley, M., Tanenbaum, R., McDaniel, M. A., & Wood, E. (1990). What happens when
university students try to answer prequestions that accompany textbook material?
Contemporary Educational Psychology, 15(1), 27–35.
Rawson, K. A. & Dunlosky, J. (2007). Improving students’ self-evaluation of learning for key
concepts in textbook materials. European Journal of Cognitive Psychology, 19
(4–5), 559–579.
Renner, K. E. (1964). Delay of reinforcement: A historical review. Psychological Bulletin, 61
(5), 341–361. https://doi.org/10.1037/h0048335
Rich, P. R., Van Loon, M. H., Dunlosky, J., & Zaragoza, M. S. (2017). Belief in corrective
feedback for common misconceptions: Implications for knowledge revision.
Journal of Experimental Psychology: Learning, Memory, And Cognition, 43(3),
492–501. https://doi.org/10.1037/xlm0000322
Roediger, H. L. & Butler, A. C. (2011). The critical role of retrieval practice in long-term
retention. Trends in Cognitive Sciences, 15(1), 20–27. https://doi.org/10.1016/j
.tics.2010.09.003
Sassenrath, J. M. & Garverick, C. M. (1965). Effects of differential feedback from examina-
tions on retention and transfer. Journal of Educational Psychology, 56(5), 259–263.
Schmidt, R. & Bjork, R. (1992). New conceptualizations of practice: Common principles in
three paradigms suggest new concepts for training. Psychological Science, 3(4),
207–217.
Sinclair, H. K. & Cleland, J. A. (2007). Undergraduate medical students: Who seeks formative
feedback? Medical Education, 41(6), 580–582.
Skinner, B. F. (1953). Science and human behavior. New York: Simon and Schuster.
Slamecka, N. J. & Katsaiti, L. T. (1987). The generation effect as an artifact of selective
displaced rehearsal. Journal of Memory and Language, 26(6), 589–607.
Smith, T. A. & Kimball, D. R. (2010). Learning from feedback: Spacing and the delay–
retention effect. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 36(1), 80–95. https://doi.org/10.1037/a0017407
Surber, J. R. & Anderson, R. C. (1975). Delay-retention effect in natural classroom settings.
Journal of Educational Psychology, 67(2), 170–173. https://doi.org/10.1037/
h0077003
Szpunar, K. K., Moulton, S. T., & Schacter, D. L. (2013). Mind wandering and education:
From the classroom to online learning. Frontiers in Psychology, 4, 495.

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press

 Correcting Student Errors and Misconceptions 459

Terrace, H. S. (1963). Errorless transfer of a discrimination across two continua. Journal of

the Experimental Analysis of Behavior, 6(2), 223–232.
Tomas, C. (2014). Marking and feedback provision on essay-based coursework: A process
perspective. Assessment and Evaluation in Higher Education, 39(5), 611–624.
https://doi.org/10.1080/02602938.2013.860078
Van Loon, M. H., Dunlosky, J., Van Gog, T., Van Merriënboer, J. J., & De Bruin, A. B. (2015).
Refutations in science texts lead to hypercorrection of misconceptions held with
high confidence. Contemporary Educational Psychology, 42, 39–48.
Walker, M. (2009). An investigation into written comments on assignments: Do students find
them usable? Assessment and Evaluation in Higher Education, 34(1), 67–78.
https://doi.org/10.1080/02602930801895752
Weaver, M. (2006). Do students value feedback? Student perceptions of tutors’ written
responses. Assessment and Evaluation in Higher Education, 31(3), 379–394.
https://doi.org/10.1080/02602930500353061
Yeager, D. S. & Dweck, C. S. (2012). Mindsets that promote resilience: When students believe
that personal characteristics can be developed. Educational Psychologist, 47(4),
302–314.
Zamary, A., Rawson, K. A., & Dunlosky, J. (2016). How accurately can students evaluate the
quality of self-generated examples of declarative concepts? Not well, and feedback
does not help. Learning and Instruction, 46, 12–20.

https://doi.org/10.1017/9781108235631.018 Published online by Cambridge University Press