Can observing a Necker cube (really) make you more insightful?
The evidence from objective and subjective indicators of insight
Abstract: Changing a problem’s representation is a crucial process when solving insight problems. Recently, Laukkonen
and Tangen (2017) found that observing ambiguous figures such as a Necker Cube before solving problems can increase
insight frequency. In our research, we extended their procedure by including measures of feelings of insight (e.g.,
confidence and pleasure). This approach allowed us to test the replicability of relationships between perceptual switching
and insight frequency in terms of both accuracy of problem solutions and insight phenomenology. The research took the
form of two studies using two different samples (N
= 68 and N
= 198) using online platforms. Our results consistently
showed no effect of prior Necker cube perception on accuracy. However, we found a significant difference in self-
reported insight (1 - non-aha! experience to 5 – a very strong aha! experience) in our Sample B study. The results suggest
the possibility that viewing ambiguous figures may not have a triggering effect on insight problem-solving performance
but that it may trigger stronger insight experiences when solving insight problems.
Keywords: problem-solving, insight
Have you ever been in a situation where you were
struggling with a problem, and then suddenly and
spontaneously the answer just popped into your mind? If
so, you probably felt an accompanying wave of pleasure and
huge confidence in the correctness of your solution. Such
experiences are termed moments of insight or aha! moments,
and are defined as “occurring when sudden comprehension,
realization or solution of a problem arises from a process
whereby the reorganization of elements of one’s mental
representation of a stimulus, situation, or event yields a non-
obvious or non-dominant interpretation” (Kounios & Bee-
man, 2014, p.71). Classic Gestalt psychologists such as
Duncker (Newell, 1985) and Köhler (1947) believed mental
problem representations to be just one possible way of
interpreting problems, and that restructuring of problem
representations can occur without cognitive control.
Can observing a Necker cube make you more insight-
ful?” The studies of Laukkonen and Tangen (2017)
Laukkonen and Tangen (2017) conducted two inter-
esting studies examining the relationship between insight
problem-solving and perception of the reversal of bistable
figures. The first study examined correlations between
perception of the reversal of bistable figures and the
solving of both insight problems and analytical problems,
and significant correlations were found between all of
these (0.38 < r < 0.48). The second study was an
experimental study in which observing a bistable figure
was shown to have a positive effect on insight problem-
The authors gave three possible reasons as to why
bistable image reversal seems to be linked to insight. First,
they suggested that the “aha!” experiences of perceptual
change occurring in response to a bistable image and
during insight problem-solving have similar phenomen-
ological characteristics. Second, they proposed that
bistable image reversals and some insightful experiences
are preceded by changes in the representation or inter-
pretation of problem elements or assumptions. Thus,
Laukkonen and Tangen suggested that both problem-
solving insights and bistable image reversals result from
the same underlying processes described by Gestalt
psychologists. Third, the authors speculated that the ability
to change perspectives when viewing ambiguous images
seems to be associated with the ability to solve creative
problems (Laukkonen and Tangen, 2017).
The first explanation was examined in Laukkonen
and Tangen’s (2017) first study. Here, the authors used
Original Papers Polish Psychological Bulletin
2021, vol. 52(4) 311–321
Corresponding author: Angelika Olszewska, firstname.lastname@example.org
SWPS University of Social Sciences and Humanities, Wroclaw, Poland; ORCID iD - 0000-0002-7953-3904
SWPS University of Social Sciences and Humanities, Wroclaw, Poland; ORCID iD - 0000-0002-5357-744X
Metcalfe’s (1986) warmth scale to measure suddenness
when making a comparison between phenomenological
experiences of suddenness when perceiving the reversal of
ambiguous figures and experiencing insight while solving
insight problems. Their results showed that both of these
phenomena were experienced more suddenly than the
appearance of solutions during analytical problem-solving.
For both insight problem-solving and ambiguous figures,
representational changes of problem structures / figures
occurred in the absence of feelings that solutions / switches
in perception were becoming increasingly imminent.
Laukkonen and Tangen’s second explanation in-
volves representational change theory (Ohlsson, 1984).
This theory suggests that insight occurs immediately after
a problem’s representation is changed. The theory posits
that a problem solver constructs an initial problem
representation and that their search for a solution is bound
by this representation. The situation does not determine the
problem’s representation: a problem can be represented in
many different ways, depending on how the solver encodes
the initial situation. Ohlsson further distinguished between
external and internal searching for a solution: internal
searches operate in the context of an initial state based on
prior knowledge and mental imagery; external searches are
based on trial and error methods. In both cases, changes in
representation are crucial.
The third explanation seems vague. The authors
speculate that the relationship between insight and bistable
image reversals may involve more than a simple analogy,
and they suggest that the ability to change perspective may
be associated with the ability to solve creative problems.
While Laukkonen and Tangen stressed that this explana-
tion is in want of an explanatory mechanism and further
evidence, they wrote the following: “Our more specific,
but also more speculative hypothesis, is that experiencing
conflict with the Necker cube would elicit conflict
monitoring and cognitive control mechanisms, which
would lead to better performance in the subsequent insight
problem.” (Laukkonen & Tangen, 2017, p. 203). With
respect to conflict monitoring, they proposed that the
reinterpretation processes used in both insight problem
solution and ambiguous figure perception are partly
accounted for by the Conflict Monitoring System de-
scribed by Botvinick et al. (2001).
Laukkonen and Tangen based their third explanation
on a previous study demonstrating that the role of conflict
monitoring and cognitive control during insight problem-
solving is to detect competing options (Kounios &
Beeman, 2014). It is assumed that bistable image reversal
reflects a switching between two competing good gestalts
or two competing interpretations (Kornmeier & Bach,
2005; Long & Toppino, 2004). Moreover, activation of the
anterior cortex prior to problem-solving is associated more
with solving insight problems than analytic problems
(Kounios et al., 2008).
To evaluate their three explanations, Laukkonen and
Tangen examined the causal relationship between experi-
encing conflict during the viewing of ambiguous figures
and performance when solving insight problems. To do
this, they conducted an experimental study (Laukkonen &
Tangen, 2017; Study 2) in which they used the Necker
cube as an example of an ambiguous figure. The Necker
cube is a two-dimensional drawing of a cube, which is
interpreted as being three-dimensional.
The observer can experience two different views of
the cube: it can be perceived as having either the lower-left
square or the upper-right square as its front side. An
ordinary Necker cube was presented in an experimental
(conflict) condition, while in a control (non-conflict)
condition, the authors used two possible interpretations
of the cube and switched them externally. The non-conflict
condition involved two images, one with an unambiguous
cube pointing left and down, and one with an unambiguous
cube pointing right and up. These images were presented
one after the other, so that a switch occurred while the
images were being viewed. In the conflict condition
a single image where both interpretations were possible
was presented. After observing the cube (conflict or non-
conflict, depending on the condition), participants were
asked to solve an insight problem. Results showed that
insight problems were more likely to be solved when
preceded by the original Necker cube.
It is worth noting that it is not entirely clear why
Laukkonen and Tangen (2017) decided to examine the
above causal relationship. The authors did not explain why
engagement of the Conflict Monitoring System during
Necker cube perception should have an impact on insight
problem-solving. However, there are studies showing that
more creative people report more frequent reversals
between possible percepts when viewing ambiguous
figures, although further studies are necessary.
The classic approaches to insightful problem-solving,
on which the assumptions of Laukkonen and Tangen’s
(2017) studies were based, stress the role of spontaneous
restructuring of problem representations. Duncker (1926)
suggested that a problem situation consists of a gap which
needs to be closed by structurally changing a “bad Gestalt”
to a “good Gestalt”. This change occurs rapidly, and he
compared the process by which an immediate realization
of a second perspective is reached to the flipping between
interpretations of a Necker cube.
The Necker cube is not the only reversal figure used
in previous research. For example, Wiseman et al. (2011)
and Olteteanu et al. (2019) have examined relationships
between insight and different ambiguous figures such as
Jastrow’s rabbit-duck and the girl-saxophonist illusion.
Wiseman et al. (2011) found a positive relationship
between self-reported creativity and self-reported ambig-
uous figure reversal. These authors conducted two studies
using different methodologies. In the first study, partici-
pants were asked if they would describe themselves as
artistically creative and as creative problem solvers using
5-point Likert scales, whereas the second study used the
Alternative Uses Task (Guilforfd, 1967) to measure
creativity. A relationship between creativity and self-
Angelika Olszewska, Agata Sobkow
reported ease of ambiguous figure reversal was found in
In their study, Olteteanu et al. (2019) found sig-
nificant correlations between ambiguous figure reversal
and performance on both the Pattern Meanings Test
(Wallach & Kogan, 1965) and the Alternative Uses Test
(Guilford, 1967), but not between ambiguous figure
reversal and insight problems. These results are interesting
because it is possible that participants occasionally
experience switches between categories in the Alternative
Uses Test (Reiter-Palmon, Forthmann, & Barbot, 2019)
and we might expect a similar phenomenon in insight
problem-solving. However, Olteteanu et al. (2019) hy-
pothesized three levels of reinterpretation (the first level
being re-representation of features as different sets of
objects or images, the second level being re-representation
of objects and their properties as different objects, and the
third level being re-representation of objects and scenes
under different problem templates), thus the lack of
a relationship between ambiguous figure reversal and
insight problem performance. They used their ambiguous
figures task as a measure of the first level, the Pattern
Meanings Test and the Alternative Uses Test as measures
of the second level, and insight problems as a measure of
the third level – each of these tasks required the
reinterpretation of an initial thought, idea, or problem
The present study. A replication and extension
of Laukkonen and Tangen’s (2017) second study
Our study revisited the findings of Laukkonen and
Tangen’s (2017) Study 2, and extended their work by
including measurement of insight phenomenology. We
chose Laukkonen and Tangen’s work for several reasons.
First, we appreciated the idea of using two versions of the
Necker cube as an experimental manipulation, thus
providing the opportunity to manipulate bistable images
while retaining a difference between conditions. Second,
the inconsistency of results for relationships between ease
of reversal of bistable images and insight problem-solving
performance suggested the need for further studies. Third,
the small effect size relating to the impact of Necker cube
manipulation on insightfulness required replication. And
finally, as Popper (2005) claimed, a true effect should be
We aimed to clarify the relationship between
perceptual reinterpretations of ambiguous figures and
reinterpretations during problem-solving. We expected to
find differences in accuracy and insight experiences
between experimental conditions. Specifically, we ex-
pected to observe more puzzle solutions after participants’
perceptions switched between the two possible Necker
cube percepts when a cube was presented prior to puzzles.
Additionally, we expected that participants in a conflict
condition would report stronger insight experiences than
those in a non-conflict condition. Our criterion for
a successful replication was the observation of a significant
difference in accuracy between the conflict and non-
Our extension used measures of insight experience
which were more comprehensive than those used in
Laukkonen and Tangen study. These authors used the
warmth scale created by Metcalfe (1986) and a self-report
question about the experience of insight, where partici-
pants answered “yes,” “no,” or “other.” But, we considered
four separate dimensions from previous studies (Danek,
Fraps, von Müller, Grothe, & Öllinger, 2014; Danek &
Wiley, 2017; Webb, Little, & Cropper, 2016). These were
suddenness, pleasure, confidence, and the aha! experience.
Suddenness was included because insightful solutions are
experienced very suddenly: as shown by Metcalfe (1986),
prior to a solution arising, participants do not have a feeling
that they are approaching a solution. Note that it is not only
insight that is experienced suddenly, a recent study by
Danek et al. (2018) also showing the sudden character of
representational changes during problem-solving. A plea-
sure scale was used because positive affect accompanies
feelings of insight (Danek et al., 2014; Gick & Lockhart,
1995). Also, subsequent to insight, a solution seems
obvious, and problem solvers report having strong
confidence in the correctness of solutions, so, similarly
to Danek et al. (2014), we used a confidence scale. Finally,
we also obtained participants’ self-reports of insight.
Although Danek et al. (2014) observed some instances
of false insight in their study, false insight (feelings of
insight when a solution is wrong) is rare, and this rarity
suggests that the phenomenon of insight may be more
connected with changing a problem’s representation and
finding a new path to a solution than with actually
obtaining a correct answer, and we thought that self-
reports would cast further light on this issue.
Summarizing, we conducted a conceptual replication
of Laukkonen and Tangen’s (2017) second study, measur-
ing insight experiences as an extension. We hypothesized
that we would observe more correct solutions of verbal
puzzles in a conflict condition (where participants
switched between two possible Necker cube percepts)
than in a non-conflict condition (where participants
attended to only one of the two possible Necker cube
percepts). Additionally, we expected that participants in
the conflict condition would report more robust insight
experiences than those in the non-conflict condition.
Two studies were conducted using two independent
samples (A and B) on two different online platforms.
However, the methods and procedures were largely the
same in both studies. The only difference was that for
Sample B an attention check was used as described in the
Procedure and Materials section below.
The Sample A study consisted of 68 undergraduate
psychology students participating in exchange for course
= 26.55 years, SD
= 7.34 years;
Can observing a Necker cube (really) make you more insightful? The evidence from objective and subjective... 313
54 females). Data for an additional 18 participants were
excluded from analysis because these participants did not
complete the whole experimental procedure; these data
were excluded before coding of correct answers to the
verbal puzzles was performed. Participants were randomly
assigned to one of two between-subjects conditions, these
conditions being named the conflict condition (n = 36) and
the non-conflict condition (n = 32) in accordance with
Laukkonen and Tangen (2017).
Participants for the Sample B study were drawn from
the Polish population (N = 198; 52 females; M
= 6.63 years) and were recruited for this
online study via the Prolific commercial panel. Participants
took part in the study in exchange for a financial reward
(GBP = 4.00). Of the participants, 34% had graduated,
56% had finished their upper secondary education, 6% had
completed the first stage of education (the equivalent of
secondary school), and 3.5 % had a vocational education.
Participants were randomly assigned to one of two
between-subjects conditions named as for Sample A:
a conflict condition (n = 98) and a non-conflict condition
(n = 100). The above description does not include three
participants who were excluded before the coding of
answers because they had duplicate records and had
probably taken part in the study twice, six participants who
failed an attention check, and 14 participants who did not
finish the whole experimental procedure.
Based on Laukkonen and Tangen’s (2017) paper, we
calculated a Cohen’s d effect size estimate of 0.26. This
small effect size required either a very large sample size or
the application of another solution if we were to reliably
replicate Laukkonen and Tangen’s results. We opted for
the latter: based on Simonsohn’s (2015) suggestions for
designing replication studies, we used the N x 2.5 rule, i.e.,
we sought to obtain a sample size 2.5 times larger than the
original study would have required to attain 80% power in
detecting an effect 33% as large as that actually obtained in
the study. This led to a sample size estimate of N = 200.
However, a priori power analysis was only conducted for
the study involving Sample B. Post-hoc power analysis
was conducted for the Sample A study using G*Power
188.8.131.52 (Faul, Erdfelder, Lang, & Buchner, 2007), which,
for α = .05 – one-tailed, an effect size of d = .26, and
subsample sizes of n
= 36 and n
=33, showed that the
study was underpowered in that power was only 28%.
Procedure and materials
Except where specified, the procedure and materials
were the same for each study. Participants in both con-
ditions received training using Jastrow’s classic duck–
rabbit reversal figure before performing the main task. The
training comprised twenty trials in which participants were
instructed to press the space bar on a keyboard each time
their perception switched from duck to rabbit or from
rabbit to duck. This training was provided so that
participants fully understood the instruction to press the
space bar after ambiguous figure reversal. The rabbit–duck
image was chosen because its reversal is easy to explain.
Participants were randomly assigned to one of the two
between-subjects experimental conditions. The classic
Necker cube was presented in the conflict condition, while
in the non-conflict condition only one of the possible
percepts of the Necker cube was used. In each of 10 trials
(see Figure 2), participants were instructed to observe the
cube (in the presence of conflict or non-conflict) for 90
seconds on a computer monitor and to press a space bar on
a computer keyboard whenever they experienced a reversal
of the cube’s position. Next, they were given 60 seconds in
which to solve a verbal puzzle. The time limits were
identical to the times used by Laukkonen (2017) in study
2. The same set of ten insight puzzles was used in random
order in both conditions. At the end of each trial,
participants were asked about their experience when
solving the puzzle. Finally, participants responded to
demographic questions about their gender and age.
The Necker cube
The nature of the Necker cube viewed was the inde-
pendent variable in our studies. In the conflict condition, we
used the classic Necker cube as used in the Laukkonen and
Tangen (2017) study. This is a well-known ambiguous
figure where there are two perceptual possibilities. A picture
of an unambiguous cube pointing left and down was
attached to the left arrow key of a computer keyboard, and
a picture of an unambiguous cube pointing right and up was
attached to the right arrow key. In the non-conflict
condition, only a picture of the cube pointing left and down
was used (the cubes used are presented in Figure 1).
Participants had one minute to solve each puzzle.
Because of the possible influence of culture and language
on solving the puzzles, insight problems were selected in
two pilot studies. The verbal puzzles from Laukkonen and
Tangen’s (2017) study were translated into Polish and
merged with puzzles previously used by Chuderski and
Jastrzębski (2018a), and Karwowski (2014). The final set
of verbal puzzles contained seven items from Laukkonen
Figure 1. The Necker cube figures used in the non-conflict
and conflict conditions
Angelika Olszewska, Agata Sobkow
and Tangen’s study, two from Karwowski’s study, and one
from Chuderski and Jastrzębski’s study.
Each participant solved ten verbal puzzles. These
puzzles took the form of open-ended questions, participants
having to type out answers. If a response was not given
within 60 seconds the next trial was presented automati-
cally. Subsequent to data collection, participants’ answers
were coded by three independent competent judges (0 –
a wrong answer/no answer, 1 – a correct answer). The
judges were blind to each participant’s experimental
condition: they received a datasheet including only
a participant’s ID, and a participant’s answers. The criterion
for coding was that at least two judges had to agree.
The sum of a participant’s correct answers served as
an index of accuracy and was used as the dependent
variable in our replication. The set of insight puzzles used
had acceptable reliability for both samples (McDonald’s
= .63 and McDonald’s Ω
, and can be found in
the Supplementary materials.
Participants’ insight experiences concerning their
answers were measured by four questions inspired by the
multidimensional approach to qualitative insight phenom-
ena (Danek et al., 2014; Webb, Little, & Cropper, 2017).
After responding to each puzzle, participants answered
these questions on a five-point Likert scale. The four
questions covered confidence (unsure to very sure),
pleasure (very unpleasant to very pleasant), an aha!
experience (no-aha! experience to a very strong aha!
experience), and suddenness (a solution came gradually to
a solution came all of a sudden). Each dimension was
analyzed separately, each variable being considered as
a different dependent variable.
An attention check was only conducted in the Sample
B study, this being done after the main tasks (shown in
Figure 2) but before the measurement of other metrics.
A Necker cube consistent with a participant’s condition
(conflict or non-conflict) was presented for 10 seconds and
participants were subsequently asked to type in the word
“cube” in an open-ended response window. We only
analyzed data for participants who were paying attention.
Analyses for the Sample A and Sample B studies
were conducted separately but were the same for each
study. Similarly to Laukkonen and Tangen (2017), we
aggregated observations across the 10 trials for each
participant by calculating a mean value. Table 1 presents
a comparison of descriptive statistics from our studies and
those of Laukkonen and Tangen. For both Sample A and
Sample B, no difference was observed between the results
of our studies and those of Laukkonen and Tangen in terms
of the median number of switches in the conflict condition.
However, in the non-conflict condition, there was far less
switching in our studies than in Laukkonen and Tangen’s
study. This is probably because there was only one
possible percept of the Necker cube in our control group.
Using (non-parametric) Mann-Whitney tests because of
some negative skewness in switching distributions, we
observed significantly more frequent switching in the
conflict condition than in the non-conflict condition in
both samples (U
= 860, p < .001, and U
p < .001), and can conclude that our manipulation was
effective in both studies.
Descriptive statistics for the insight experience mea-
sures are presented in Tables 2A and 2B. Average values for
all the puzzles were compared between the two conditions
within both samples. No differences were found between the
two conditions with respect to accuracy, t
(66) = 0.89,
p = .38, d = 0.214, and t
(196) = 0.11, p = .91, d = 0.017.
However, we found a significant difference in self-reported
insight (1 – no-aha! experience to 5 – a very strong aha!
experience) in Sample B, t
(196) = 2.57, p < .001,
d = 0.365. Figure 3 presents a comparison of effects. We
observed no significant differences for the other insight
Figure 2. Example trials for the conflict condition
(panel A) and non-conflict condition (panel B)
= 0.61 and Cronbach’s α
Can observing a Necker cube (really) make you more insightful? The evidence from objective and subjective... 315
Figure 3. A comparison of differences between conditions showing means and standard deviations
Table 1 Comparison of the present results with those of Laukkonen and Tangen (2017)
Sample A Sample B Laukkonen & Tangen
Conflict Non-conflict Conflict Non-conflict Conflict Non-conflict
Switching Mdn = 24.90
SD = 22.03
Mdn = 4.50
SD = 14.85
Mdn = 20.25
SD = 21.87
Mdn = 3.75
SD = 23.11
Mdn = 8.20
SD = 26.40
Mdn = 26.10
SD = 4
Accuracy M = 4.33
SD = 2.61
M = 3.82
SD = 2.21
M = 4.89
SD = 2.10
M = 4.86
SD = 2.40
M = 4.24
SD = 1.87
M = 3.76
SD = 1.82
Table 2A Descriptive statistics and t-test results for Sample A measures of insight phenomenology
Conflict Non-conflict t(66) p d
M SD M SD
Confidence 2.79 0.97 2.66 0.89 0.59 .56 0.142
Pleasure 3.04 0.60 3.03 0.79 0.07 .95 0.016
Suddenness 3.02 0.83 3.10 0.74 -0.43 .67 -0.104
Aha! experience 2.59 0.86 2.43 0.94 0.72 .47 0.176
Table 2B Descriptive statistics and t-test results for Sample B measures of insight phenomenology
Conflict Non-conflict t(196) p d
M SD M SD
Confidence 3.10 0.63 3.10 0.67 0.01 .99 0.002
Pleasure 3.48 0.58 3.37 0.68 1.28 .26 0.160
Suddenness 3.33 0.69 3.23 0.59 1.09 .28 0.155
Aha! experience 2.79 0.71 2.54 0.66 2.57 .01 0.365
Angelika Olszewska, Agata Sobkow
We also examined relationships between all the
measures for both samples (see Table 3 and Table 4).
There were no significant Pearson’s correlations between
perceptual switching and either accuracy or insight
experience. However, we observed positive relationships
among dimensions of insight phenomenology. Interest-
ingly we observed significantly weaker correlations among
aha! experience and confidence, pleasure and suddenness
in sample B compared to sample A (see supplementary
materials, Table S4).
The present study aimed to replicate and extend the
findings of Laukkonen and Tangen’s (2017) research. In
their study, Laukkonen and Tangen found that looking at
a Necker cube had an impact on accuracy in solving insight
problems. But both of our studies found no support for this
effect. We found no difference between the conflict and
non-conflict conditions in the number of correct answers
participants provided to verbal puzzles. We also found no
relationship between number of switches and average
number of correct answers in the conflict condition.
However, in an extension of Laukkonen and Tangen’s
study, we found a significant difference in self-reported
insight for Sample B but not for Sample A. For Sample B,
the aha! experience was stronger in the conflict condition
than in the non-conflict condition. Laukkonen and Tangen
(2017) found similar results in their study. After each
puzzle, Laukkonen and Tangen asked participants if they
experienced insight using 3-point scale (1 – no, 2 – other,
3 – yes), and they found that participants in the conflict
condition experienced insight more frequently than those in
the non-conflict condition. This effect is consistent with the
result we observed for our extended self-reported insight
scale (we used a 5-point Likert scale from 1 – no-aha!
experience to 5 – a very strong aha! experience).
It was surprising that we found no differences
between the conflict and non-conflict conditions with
respect to number of correct answers and the suddenness,
pleasure, and confidence dimensions of insight phenom-
enology. There are at least two possible explanations for
these results. First, it is possible that there was a significant
difference for the aha! experience but not accuracy
because the experiencing of insight is related to the type
of problem-solving involved rather than solving problems
per se, and classical insight problems can be solved both in
a manner involving insight and analytically (Danek,
Wiley, & Öllinger, 2016; Webb et al., 2016). Second,
previous studies have observed differences in self-reported
insight between different types of problems in the absence
of differences in other phenomenological dimensions.
Thus, Webb et al. (2016) observed differences across
insight and non-insight tasks using an aha! experience
scale but did not observe differences on other insight
dimensions. Similarly, Chuderski et. al. (2020), compared
insight problems with Raven’s matrices problems and
observed differences with respect to self-reported intuition
but not pleasure, suddenness, and certainty. Finally, Danek
et al. (2020) found that a stronger aha! experience is
related to sudden changes in problem representation.
Moreover, we would like to highlight the limitations of
Table 3 Pearson’s r correlations for Sample A measures
1 2 3 4 5
2. Pleasure .69**
3. Suddenness .60** .56**
4. Aha! experience .76** .65** .66**
5. Accuracy .59** .27* .30* .48**
6. Switching .13 -.05 .09 .09 .23
Note: ** p < .001, * p < .05,
Table 4 Pearson’s r correlations for Sample B measures
1 2 3 4 5
2. Pleasure .53**
3. Suddenness .58** .38**
4. Aha! experience .28** .37** .31**
5. Accuracy .51** .36** .23** .02
6. Switching -.08 -.05 -.12 .08 -.02
Note: ** p < .001, * p < .05
Can observing a Necker cube (really) make you more insightful? The evidence from objective and subjective... 317
aha! experience scale. As we observed in sample B this
scale has slightly different properties than in sample A. We
observed it in correlation comparison between samples
(see supplementary materials, Table S4), the aha! experi-
ence scale is the only scale with different correlation
coefficients in sample B. We can find a few possible
explanations for this effect. First, the different motivations
(credit points in sample A vs monetary reward in sample
B) may affect the awareness of the Likert scale. The
participants in sample A were less reflective about the
phenomenology and more consistent with the other.
Second, they were psychology students who may have
a different understanding what aha! experience is. Finally,
when we conducted the sensitivity calculation for this
effect using pwr-package (Champely, 2020), assuming
80% power and effect size d = 0.36, 122 (per group) where
required for observing this effect. Thus, the sample A may
It is worth noting that, despite a significant difference
in number of correct answers between conditions, the
mean numbers of correct answers for the two conditions
observed for Sample A were more similar to those in
Laukkonen and Tangen’s study than were those for
Sample B. For Sample A, an average of 38% correct
answers occurred in the non-conflict condition compared
to 37.6% in Laukkonen and Tangen’s study, and in the
conflict condition an average of 43% of correct answers
occurred compared to the 42.4% observed by Laukkonen
and Tangen. However, for Sample B, we observed more
correct answers for both conditions (49% for the conflict
condition, and 48.6% for the non-conflict condition). This
inconsistency may be explained by the nature of
participants in Sample B (non-students sample rather than
psychology undergraduate students) or by a difference in
participants’ motivations: Sample B participants obtained
a financial reward. It might be suspected that one reason
for the higher proportion of correct answers in our study is
that three puzzles were added to the seven (translated)
original puzzles used by Laukkonen and Tangen, and that
these three extra puzzles were easier. However, a test of
whether this was the case showed that the average
proportion of correct answers for the new puzzles did
not differ from the average proportion of correct answers
for the translated Laukkonen and Tangen puzzles in each
of our samples.
Our replication has limitations in that we did not
replicate Laukkonen and Tangen’s study (2017) comple-
tely. First, they used a within-subjects design whereas we
used a between-subjects design. Second, we used only
seven of Laukkonen and Tangen’s puzzles, and these were
translated into a different language. Finally, we used
a slightly different Necker cube manipulation in that we
used only one of the two possible percepts in the non-
Although the above differences may have affected our
results, it seems reasonable to argue that the manipulation
used in our two studies was effective. We observed
a significant difference for both samples. The number of
switches in our conflict condition was comparable with
that reported by Laukkonen and Tangen (2017). However,
the number of switches in our non-conflict condition
= 4.5; Mdn
= 3.75) was far lower than in
Laukkonen and Tangen’s study (Mdn = 26.1). Surpris-
ingly, participants effectively experienced the switching of
a non-ambiguous figure roughly once every 30 seconds,
suggesting the possibility that participants may have
experienced some perceptual changes. It would be useful
to explore the phenomenological and qualitative experi-
ence of such perceptual changes. In ending, this brief
discussion of switching, it is worth highlight the fact that
the relationship between the number of switches partici-
pants experienced and accuracy was non-significant in
both the original study and our replication.
To conclude, the present study did not replicate
Laukkonen and Tangen’s (2017) observation that prior
Necker cube perception has an effect on accuracy during
insight problem-solving. Furthermore, we observed no
relationships between different dimensions of insight
experience and switching when observing a Necker cube.
However, we would argue that the manipulation we used
was effective (as shown by significant differences in the
number of switches across our conflict and non-conflict
conditions), and that our set of puzzles showed good
reliability and validity (in the form of significant correla-
tions with insight experience measures). Our results do not
allow us to conclude that observing a Necker cube triggered
problem-representation changes. A comparison of our
findings with the inconsistent results of previous studies
(Olteteanu et al., 2019; Wiseman et al., 2011) leads us to
suggest the possibility that viewing ambiguous figures may
not have a triggering effect on insight problem-solving
performance but that it may trigger stronger insight
experiences when solving insight problems.
This research was carried out with the financial
support of SWPS University. We would like to thank the
lead author of the original study, Ruben Laukkonen, for
providing all the information required to conduct our
replication, e.g., sharing correct answers for the puzzles he
used, and providing effect size and confidence interval
information relating to the main effect observed in his work.
Botvinick, M. M., S, B. T., Barch, D. M., Carter, C. S., & Cohen, J. D.
(2001). Conflict Monitoring and Cognitive Control. Psychological
Review, 103(3), 624–652.
Champely, S., (2020). pwr: Basic Functions for Power Analysis.
R package version 1.3-0.
Chuderski, A., Jastrzębski, J., Kroczek, B., Kucwaj, H., & Ociepka, M.
(2020). Metacognitive experience on Raven’s matrices versus insight
problems. Metacognition and Learning, 31–44. https://doi.org/
Angelika Olszewska, Agata Sobkow
Danek, A. H. (2018). Magic tricks, sudden restructuring, and the Aha!
experience: A new model of nonmonotonic problem solving. In
F. Valee-Tourangeau (Ed.), Insight (pp. 51-78). Routledge.
Danek, A. H., Fraps, T., von Müller, A., Grothe, B., & Öllinger, M.
(2014). It’s a kind of magic-what self-reports can reveal about the
phenomenology of insight problem solving. Frontiers in Psychology,
5, 1–11. https://doi.org/10.3389/fpsyg.2014.01408
Danek, A. H., & Wiley, J. (2017). What about false insights?
Deconstructing the aha! experience along its multiple dimensions
for correct and incorrect solutions separately. Frontiers in Psychol-
ogy, 7, 1–14. https://doi.org/10.3389/fpsyg.2016.02077
Danek, A. H., Wiley, J., & Öllinger, M. (2016). Solving classical insight
problems without aha! experience: 9 Dot, 8 Coin, and matchstick
arithmetic problems. Journal of Problem Solving, 9(1), 47–57.
Danek, A. H., Williams, J., & Wiley, J. (2020). Closing the gap:
connecting sudden representational change to the subjective Aha!
experience in insightful problem solving. Psychological Research,
0(0), 1–9. https://doi.org/10.1007/s00426-018-0977-8
Duncker, K. (1926). A qualitative (experimental and theoretical) study of
productive thinking (solving of comprehensible problems). Pedago-
gical Seminary and Journal of Genetic Psychology, 33, 642–708.
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power 3:
A flexible statistical power analysis program for the social,
behavioral, and biomedical sciences. Behavior Research Methods,
39(2), 175–191. https://doi.org/10.3758/BF03193146
Gick, M. L., & Lockhart, R. S. (1995). Cognitive and affective
components of insight. In R. J. Sternberg & R. Davidson, John
(Eds.), The nature of insight (pp. 197–228). Cambridge: MIT Press.
Guilforfd, J. P. (1967). The nature of human inteligence. McGraw-Hill.
Köhler, W. (1947). Gestalt Psychology. The definitive statement of the
Gestalt Theory. New York: Liveright.
Kornmeier, J., & Bach, M. (2005). The Necker cube - An ambiguous
figure disambiguated in early visual processing. Vision Research,
45(8), 955–960. https://doi.org/10.1016/j.visres.2004.10.006
Kounios, J, & Beeman, M. (2014). The cognitive neuroscience of insight.
Annual Review of Psychology, 65. https://doi.org/10.1146/annurev-
Kounios, John, Fleck, J. I., Green, D. L., Payne, L., Stevenson, J. L.,
Bowden, E. M., & Jung-Beeman, M. (2008). The origins of insight in
resting-state brain activity. Neuropsychologia, 46(1), 281–291.
Laukkonen, R. E., & Tangen, J. M. (2017). Can observing a Necker cube
make you more insightful? Consciousness and Cognition, 48, 198–
Long, G. M., & Toppino, T. C. (2004). Enduring interest in perceptual
ambiguity: Alternating views of reversible figures. Psychological
Bulletin, 130(5), 748–768. https://doi.org/10.1037/0033-2909
Metcalfe, J. (1986). Feeling of knowing in memory and problem solving.
Journal of Experimental Psychology: Learning, Memory, and
Cognition, 12(2), 288–294. https://doi.org/10.1037/0278-7393
Newell, A. (1985). Duncker on thinking: An inquiry into progress in
cognition. A Century of Psychology as Science., 392–419. https://doi.
Ohlsson, S. (1984). Restructuring revisited: II. An information processing
theory of restructuring and insight. Scandinavian Journal of
Psychology, 25(2), 117–129. https://doi.org/10.1111/j.1467-9450
Olteteanu, A. M., Schöttner, M., & Bahety, A. (2019). Towards a multi-
level exploration of human and computational re-representation in
unified cognitive frameworks. Frontiers in Psychology, 10(APR), 0–
Popper, K. (2005). The Logic of Scientific Discovery. In Hume’s
Reiter-Palmon, R., Forthmann, B., & Barbot, B. (2019). Scoring
divergent thinking tests: A review and systematic framework.
Psychology of Aesthetics, Creativity, and the Arts, 13(2), 144–152.
Wallach, M. A., & Kogan, N. (1965). A new look at the creativity‐
intelligence distinction 1. Journal of Personality, 33(3), 348–369.
Webb, M. E., Little, D. R., & Cropper, S. J. (2016). Insight is not in the
problem: Investigating insight in problem solving across task types.
Frontiers in Psychology, 7, 1–13. https://doi.org/10.3389/
Webb, M. E., Little, D. R., & Cropper, S. J. (2017). Once more with
feeling: Normative data for the aha experience in insight and
noninsight problems. Behavior Research Methods, 50(5), 2035–
Wiseman, R., Watt, C., Gilhooly, K., & Georgiou, G. (2011). Creativity
and ease of ambiguous figural reversal. British Journal of Psychol-
ogy, 102(3), 615–622. https://doi.org/10.1111/j.2044-8295.2011
Can observing a Necker cube (really) make you more insightful? The evidence from objective and subjective... 319
VERBAL PUZZLES USED IN STUDY
1. Bezrobotna kobieta, która nie miała przy sobie prawa jazdy zignorowała nakaz stopu na skrzyżowaniu, następnie nie
zważając na znak informujący o ulicy jednokierunkowej ruszyła pod prąd. To wszystko obserwował policjant na
służbie, który nic z tym nie zrobił. Dlaczego?
2. Morderca został skazany na karę śmierci. Musi wybrać jedno z trzech pomieszczeń. Pierwszy pokój jest wypełniony
płomieniami ognia, drugi jest pełen zabójców z naładowaną bronią, z kolei w trzecim są lwy, które nie jadły od 3 lat.
Który pokój jest dla niego najbezpieczniejszy?
3. Więzień planował ucieczkę z więzienia. W swojej celi znalazł linę, która wystarczała do połowy wysokości budynku.
Podzielił linę na pół, połączył dwie części i uciekł. Jak to możliwe?
4. Jan mył okna wielopiętrowego biurowca, kiedy nagle się pośliznął i spadł z 18-metrowej drabiny na betonowy
chodnik. Na szczęście nic mu się nie stało. Jak to możliwe?
5. Dwóch mężczyzn rozegrało pięć pełnych partii w warcaby. Każdy z nich wygrał równą liczbę partii, przy czym żadna
z gier nie zakończyła się remisem ani rezygnacją któregoś z graczy. Jak to możliwe?
6. Jeśli w szufladzie masz wymieszane czarne i brązowe skarpetki w proporcji 4 do 5 to ile musisz wyjąć skarpetek, aby
mieć pewność, że masz parę w tym samym kolorze?
7. Emerytowany Profesor prowadził swój stary samochód, kiedy nagle auto samo zmieniło bieg. Profesor nie zwrócił na
to uwagi i kontynuował jazdę, jakby nic się nie stało. Dlaczego się nie zaniepokoił?
8. Brat, jego siostra oraz mąż ze swoją żoną znaleźli 4 monety. Każde z nich wzięło po jednej i jedna im została jedna
moneta. Jak to możliwe?"
9. Co to jest: antyczny wynalazek wciąż używany w większości krajów świata, umożliwiający patrzenie przez ściany?
10. Kierowca Fiata ma brata, lecz brat kierowcy Fiata nie ma brata. Kim kierowca Fiata jest dla brata?
ENGLISH VERSION OF VERBAL PUZZLES
1. An unemployed woman did not have her driver’s license with her. She failed to stop at a railroad crossing, then
ignored a one-way traffic sign and traveled three blocks in the wrong direction down the one-way street. All this was
observed by a policeman, who was on duty, yet he made no effort to arrest the woman. Why?
2. A murderer is condemned to death. He has to choose among three rooms. The first is full of raging fires, the second is
full of assassins with loaded guns, and the third is full of lions that haven't eaten in 3 years. Which room is safest for
3. A prisoner was attempting to escape from a tower. He found in his cell a rope that was half long enough to permit him
to reach the ground safely. He divided the rope in half, tied the two parts together, and escaped. How could he have
4. Mr. Hardy was washing windows on a high-rise office building when he slipped and fell off a sixty foot ladder onto
the concrete sidewalk below. Incredibly, he did not injure himself in any way. How is this possible?
5. Two men played five full games of checkers and each won an even number of games, with no ties, draws, or forfeits.
How is that possible?
6. If you have black socks and brown socks in a drawer, mixed in a ratio of 4 to 5, how many socks will you have to take
out to make sure that you have a pair of the same colour?
7. Professor Bumble, who is getting on in years, was driving along in his old car when suddenly it shifted gears by itself.
He paid no attention and kept on driving. Why wasn’t he concerned?
8. A brother with his sister and a husband with his wife found 4 coins. Each of them took 1 coin, but there was still 1
coin left. How is that possible?
9. What is it: an antique invention still used in most countries of the world which makes it possible to look through
10. A car driver has a brother, but the brother of the driver does not have a brother. Who is the driver for the brother?
Puzzles with numbers 1-7 were used in oryginal Laukkonen and Tangen’s (2017) study. Puzzles 8-9 were used by Karwowski (2014) and puzzle
10 was used by Chuderski and Jastrzębski (2018a).
Puzzles with numbers 1-7 were used in oryginal Laukkonen and Tangen’s (2017) study. Puzzles 8-9 were used by Karwowski (2014) and puzzle
10 was used by Chuderski and Jastrzębski (2018a).
Angelika Olszewska, Agata Sobkow
Table S1 Insight phenomenology measures used in study
dimension question answers (5-point Likert scale)
nagłość (Suddenness) W jaki sposób odpowiedź pojawiła się w Twoim
1 - stopniowo byłem/-am coraz bliżej
5 - rozwiązanie pojawiło się nagle
przyjemność (Pleasure) W jakim stopniu rozwiązywanie tej zagadki było dla
1 - bardzo nieprzyjemne
5 - bardzo przyjemne
pewność (Confidence) W jakim stopniu jesteś pewny/-a swojej odpowiedzi? 1 - nie jestem pewny/-a,
5 - jestem bardzo pewny/-a
olśnienie (Aha! experience) W jakim stopniu doznałeś/aś olśnienia?
Olśnieniem nazywa się moment, w którym nagle
wpadasz na rozwiązanie jakiegoś problemu
i spontanicznie mówisz „Aha!”.
1 - nie doznałem-am olśnienia;
5 - całkowicie mnie olśniło
Table S2 Insight phenomenology measures in English
dimension question answers (5-point Likert scale)
suddenness How did the answer come to your mind? 1 - solution came gradually
5 - all of a sudden
pleasure How pleasant was solving this puzzle? 1 - very unpleasant
5 - very pleasant
confidence How strong are you sure that your answer is correct? 1 - unsure
5 - very sure
aha! experience How strongly did you feel an insight?
Insight is called the moment when you realized the
solution suddenly and often spontaneously say “aha!”.
1 - non-aha experience
5 - very strong aha experience
Table S3 Factor loadings, model fit indices, and internal consistency measures for the subjective insight measure for Sample A
and Sample B
Sample A Sample B
Factor loading Factor loading
1. Confidence 0.868 0.795
2. Pleasure 0.774 0.655
3. Suddenness 0.725 0.673
4. Aha! experience 0.875 0.447
(EFA, one factor solution)
(2) = 1.31; p = .519 X
(2) = 9.79; p = .007
RMSEA = 0.00 [0.00; 0.21] RMSEA = 0.14 [0.06; 0.23]
TLI = – 1.01 TLI = 0.86
Table S4 Comparison
of significant correlations observed in two samples
Correlation between Sample A Sample B Z p
Confidence - pleasure .69 .52 1.86 .063
Confidence - suddenness .60 .58 0.30 .761
Confidence – aha! experience .76 .28 4.97 <.001
Confidence - accuracy .59 .51 0.86 .391
Pleasure - suddenness .56 .38 1.65 .099
Pleasure – aha! experience .65 .37 2.76 .005
Pleasure - accuracy .27 .36 -0.69 .489
Suddenness – aha! experience .66 .31 3.24 .001
Suddenness - accuracy .29 .23 0.47 .637
Sample A – r Pearson coefficient for the Sample A,
Sample B – r Pearson coefficient for the Sample B
Diedenhofen, B. & Musch, J. (2015). cocor: A Comprehensive Solution for the Statistical Comparison of Correlations. PLoS ONE, 10(4):
e0121945. doi: 10.1371/journal.pone.0121945 Available: http://dx.doi.org/10.1371/journal.pone.01219
Can observing a Necker cube (really) make you more insightful? The evidence from objective and subjective... 321