INFORME
High Performance on a Pragmatic Task may
Not Be the Result of Successful Reasoning:
On the Importance of Eliciting Participants’
Reasoning Strategies
Alexandra Mayn1 and Vera Demberg1,2
un acceso abierto
diario
1Department of Language of Science and Technology, Saarland University
2Departamento de Ciencias de la Computación, Saarland University
Palabras clave: experimental stimuli, reasoning, pragmatic inferences
ABSTRACTO
Formal probabilistic models, such as the Rational Speech Act model, are widely used for
formalizing the reasoning involved in various pragmatic phenomena, and when a model
achieves good fit to experimental data, that is interpreted as evidence that the model
successfully captures some of the underlying processes. Yet how can we be sure that
participants’ performance on the task is the result of successful reasoning and not of some
feature of experimental setup? en este estudio, we carefully manipulate the properties of the
stimuli that have been used in several pragmatics studies and elicit participants’ reasoning
estrategias. We show that certain biases in experimental design inflate participants’
performance on the task. We then repeat the experiment with a new version of stimuli which is
less susceptible to the identified biases, obtaining a somewhat smaller effect size and more
reliable estimates of individual-level performance.
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
/
mi
d
tu
oh
pag
metro
i
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
d
oh
i
/
i
.
/
/
1
0
1
1
6
2
oh
pag
metro
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
oh
pag
metro
_
a
_
0
0
0
7
7
pag
d
/
.
i
Citación: Mayn, A., & Demberg, V.
(2023). High Performance on a
Pragmatic Task may Not Be the Result
of Successful Reasoning: Sobre el
Importance of Eliciting Participants’
Reasoning Strategies. Mente abierta:
Descubrimientos en ciencia cognitiva, 7,
156–178. https://doi.org/10.1162/opmi
_a_00077
DOI:
https://doi.org/10.1162/opmi_a_00077
Recibió: 23 Noviembre 2022
Aceptado: 30 Abril 2023
Conflicto de intereses: Los autores
declare no conflict of interest.
Autor correspondiente:
Alexandra Mayn
amayn@lst.uni-saarland.de
Derechos de autor: © 2023
Instituto de Tecnología de Massachusetts
Publicado bajo Creative Commons
Atribución 4.0 Internacional
(CC POR 4.0) licencia
La prensa del MIT
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
INTRODUCCIÓN
Formal probabilistic models, such as the Rational Speech Act model (RSA, Franco & Buen hombre,
2012), are widely used to formalize the reasoning involved in various pragmatic phenomena,
such as scalar implicatures (Buen hombre & Stuhlmüller, 2013), hyperbole (Kao et al., 2014), y
irony (Kao & Buen hombre, 2015). RSA assumes that the speaker and the listener are cooperative
and reason recursively about each other to arrive at a shared interpretation. These models are
then evaluated by being fitted to experimental data, and if the they achieve a close fit, eso es
interpreted as evidence that the model successfully captures some of the underlying processes.
Yet how can we be sure that participants’ high performance on the task is indeed the result of
successful reasoning, and not some other factor related to the experimental setup? Sikos et al.
(2021) revisited the influential RSA study by Frank and Goodman (2012) and found only modest
evidence of people engaging in the assumed pragmatic reasoning. They showed that a simple
literal listener model, which is governed by the salience prior over objects, provided an equally
good fit to the data as the full pragmatic listener model. The authors therefore argued that the
good fit of the RSA model to empirical data was largely due to a combination of non-pragmatic
factores. Neglecting to carefully investigate whether the experimental design contains biases
hence may lead to making unwarranted conclusions about the phenomena we study.
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
Franke and Degen (2016) ask their participants to identify the referent of an ambiguous
message and show that there is quite a lot of individual variability in performance, y eso
the data is better captured by a model which assumes that each participant has their own rea-
soning type, corresponding to predictions of three probabilistic models of different degrees of
complejidad, than by a population-level model that assumes that all participants have the same
reasoning. Por lo tanto, better performance on the task is interpreted as evidence of participants
successfully employing higher complexity of reasoning.
en este estudio, we conduct a series of experiments based on the task used by Franke and
Degen (2016) (originally introduced in Degen & Franke, 2012) where we carefully manipulate
the properties of the stimuli and also elicit participants’ reasoning strategies. We argue that
certain biases are present in the task which allow participants to arrive at the correct answer
without engaging in the assumed pragmatic reasoning. We then design a version of stimuli
aimed at mitigating the identified biases and repeat the experiment, obtaining a somewhat
smaller effect size, y, importantly, more reliable individual-level results. We argue that
probing an experimental design for biases is crucial for ensuring that we can draw meaningful
conclusions, and strategy elicitation is a simple and efficient way of doing so.
All data and analysis scripts for the results reported in this paper are available at https://
github.com/sashamayn/refgame_stimuli_methods.
BACKGROUND
Reference Game
The task which our experiments build on is the reference game, which is the Experiment 1 en
Degen and Franke (2012). Participants’ task is to identify the referent of a message.
On each trial, participants are presented with three objects, each of which is a creature wear-
ing an accessory. There are three possible creatures (green monster, purple monster, and robot)
and three possible accessories (red hat, blue hat, and scarf ). We can think of the objects on the
pantalla, por lo tanto, as varying across two feature dimensions—the creature feature and the acces-
sory feature. Participants also see a message that they are told was sent by the previous partici-
pant. The message is always either a creature (without an accessory) or an accessory. En tono rimbombante,
participants are told that not all creatures and accessories are available as messages: there are no
messages scarf or robot, hence these are so-called inexpressible features. Participants’ task is then
to pick the creature they believe the previous participant was referring to.
The task consists of 66 experimental trials, of which 24 are critical and 42 are fillers. Cada
trial display consists of the target (correct answer), competitor and distractor, presented in
random order.
On the critical trials, the message is ambiguous. Half of the critical trials are simple impli-
cature trials, where only one reasoning step is required to solve them. An example of a simple
implicature trial is presented in the top panel of Figure 1: the message (red hat) may at first
glance be referring to either the robot (objetivo) or to the green monster (competitor). To draw the
implicature, one can reason that if the speaker had meant to refer to the competitor (verde
monster), they could have used the unambiguous message green monster, whereas there is
no way to refer to the target (robot) unambiguously since robot is not an available message,
hence the red hat must be referring to the robot.
Complex trials require two reasoning steps because the target shares one feature with the
competitor and the other feature with the distractor. An example of a complex implicature trial
MENTE ABIERTA: Descubrimientos en ciencia cognitiva
157
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
/
mi
d
tu
oh
pag
metro
i
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
d
oh
i
/
i
/
/
.
1
0
1
1
6
2
oh
pag
metro
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
oh
pag
metro
_
a
_
0
0
0
7
7
pag
d
/
.
i
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
/
mi
d
tu
oh
pag
metro
i
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
d
oh
i
/
i
.
/
/
1
0
1
1
6
2
oh
pag
metro
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
oh
pag
metro
_
a
_
0
0
0
7
7
pag
d
/
.
i
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
Cifra 1.
Example of a simple and a complex critical trial.
is presented in the bottom panel of Figure 1. Aquí, the message (red hat) may at first glance
also be referring to either the target (purple monster with a red hat) or the competitor (verde
monster with a red hat). One would then need to reason that if the speaker had wanted to refer
to the competitor (green monster), they could have used the unambiguous green monster. En
contrast, there is no unambiguous way of referring to the target since its other feature (purple
monster) is shared with the distractor, hence the red hat must be referring to the purple monster
with the red hat.
Del 42 fillers, 33 are completely unambiguous and 9 completely ambiguous. On
completely unambiguous trials, only the target has the feature expressed by the message.
Por ejemplo, the message is a purple monster and there’s only one purple monster on the
pantalla. Those trials are used as an attention check. On completely ambiguous trials, the target
and the competitor are identical. Por ejemplo, the message is a purple monster and there are
two identical purple monsters wearing a blue hat and a robot with a red hat on the screen.
Since there is no way of knowing which of the two identical creatures is the target, perfor-
mance on the ambiguous trials constitutes a random baseline.
Models of Reasoning Complexity in the Reference Game
Franke and Degen (2016) show that formal probabilistic RSA models make distinct predictions
with regards to simple and complex implicatures in the reference game. These models are
defined in terms of a speaker LN and a listener SN−1 who recursively reason about each other
to arrive at an interpretation. The listener’s task is to identify the referent object o given the
message m uttered by the speaker.
The most simple listener model, the literal listener (L0), assigns equal probability to every
(cid:3)
,
object for which the message is literally true: L0 ojmð
(cid:4)
Þ∝ exp λ⋅ U oj o’jm i s
t r u e o f o’gÞ
(cid:1)
(cid:4)
MENTE ABIERTA: Descubrimientos en ciencia cognitiva
158
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
where U is a uniform distribution and is a hyperparameter which governs how strictly an agent
adheres to their utility function as opposed to picking a referent at random. De este modo, L0 will have
chance-level performance on both types of critical trials. Por ejemplo, in the top panel of
Cifra 1, L0 will assign equal probability to the target and the competitor since they are both
wearing a red hat.
(cid:4)
(cid:4)
Þ∝ exp λ⋅ U mj m’jm’
A slightly more sophisticated L1 listener, who reasons about a literal speaker S0, will be able
to successfully solve the simple but not the complex implicature condition. Let’s show why
that is the case. To describe an object o, the literal speaker S0 will pick any message m which is
(cid:3)
literally true of o with equal probability: S0 mjoð
. De este modo,
for the top panel of Figure 1, the literal speaker will always use the message “red hat” to refer
to the target, and will use the messages “red hat” and “green monster” with equal probability
to refer to the competitor. The L1 listener will then reason that the target is more likely than the
competitor, since for the competitor, the S0’s probability is split between two messages. L1 is
not going to be powerful enough to solve the complex implicatures, sin embargo, because for
both target and competitor, S0’s probability of using the red hat message to refer to them is
1
2, as for each of them there is another available message (purple monster and green monster
respectivamente).
t r u e o f ogÞ
(cid:1)
i s
Finalmente, the pragmatic L2 listener, who reasons about a pragmatic speaker S1, OMS, Sucesivamente,
reasons about the literal listener L0, is powerful enough to also solve the complex implicature
tipo. That is because, while S0 is equally likely to use either feature to refer to the target and
the competitor, in the S1, that symmetry is broken. Let’s look at the example in the bottom
panel of Figure 1. S1 knows that L0 will assign equal probabilityto all referents compatible with
the message: if the speaker sends the message green monster, the listener will umabiguously
identify the competitor, whereas if they send the message red hat, the listener will randomly
choose one of twofitting referents. Por lo tanto, S1 will avoid using the red hat to refer to the
competitor. The pragmatic listener L2 knows this, so they will correctly identify the target upon
receipt of the message red hat.
Franke and Degen (2016) fit a hierarchical Bayesian model to their experimental data from
the reference game, whereby one of the three listener types (L0, L1 or L2) was assigned to each
individual participant, and showed that this individual-level model provided a better fit to the
data than a model that assumed that all participants had the same reasoning type. Mayn and
Demberg (2022) found that participants’ performance on the reference game is modulated by
their abstract reasoning ability, as measured by the Raven’s Progressive Matrices and the Cog-
nitive Reflection Test.
Annotation of Reasoning Strategies
Since we were interested in how the participants solve the reference game, in addition to col-
lecting participants responses to the experimental stimuli, in each of our experiments, nosotros también
elicited their reasoning strategies. After completing the main experiment, participants saw one
simple and one complex item again, randomly selected and presented in randomized order.
When they clicked on one of the creatures, a red box appeared around it, along with the ques-
tion “Why did you make that choice?” and a textbox. This was done as a probe into partici-
pants’ reasoning.
All responses were annotated by two annotators, one of whom was blind to the purpose of
the experiment. All disagreements were resolved jointly: the two annotators met and each
made a case for their choice of tag. If one of the annotators successfully convinced the other,
MENTE ABIERTA: Descubrimientos en ciencia cognitiva
159
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
/
mi
d
tu
oh
pag
metro
i
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
d
oh
i
/
i
/
/
.
1
0
1
1
6
2
oh
pag
metro
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
oh
pag
metro
_
a
_
0
0
0
7
7
pag
d
/
.
i
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
the tag was changed accordingly; de lo contrario, the tag unclear was assigned to the item and it
was removed from future analysis.
Annotation scheme. Participants’ responses were assigned one of five tags.
The category correct_reasoning was assigned to hypothetical reasoning about alternatives
of the kind described by the RSA model. An example for the top panel of Figure 1 would be
“The speaker could have used the message green monster if they meant to refer to the other
creature with the red hat, but they didn’t, which makes me think that they wanted me to
choose the robot”.
Random responses were assigned to the category guess. Some participants indicated screen
location as an explanation (often choosing the middle option) or referred to their responses on
previous trials (p.ej., “I have chosen a purple monster a lot in the past so now I will choose the
robot”). Those were categorized as guessing since participants indicate that they had no way
of differentiating between the target and the competitor and then used some other superficial
criterion to break the tie.
Cases where participants described a reason for their choice which was something other
than hypothetical reasoning about alternatives were labeled other_reason. Within that cate-
gory, we further assigned one of the following five subcategories tothe explanations. The tag
visual_resemblance was assigned when the participant stated that they selected the creature
which they found to be “visually the most similar” to the message. Some of the explanations
included a justification revealing the nature of the similarity the participant picked up on. Para
instancia, a common reason for selecting the target in the top panel of Figure 1 was that the
robot’s hat is facing in the same direction as the message, whereas the competitor’s hat is
facing the other way. odd_one_out was assigned when the participant reported selecting a
creature that stands out because it is the only one that has a certain feature. Por ejemplo,
a participant might select the distractor in the top panel of Figure 1 as it is the only creature
with a scarf and without a hat. The tag salience was assigned when the participant reported
selecting a certain creature because it stood out to them. An example response for the top
panel of Figure 1 which was assigned the salience tag is selecting the competitor “Because it
is the biggest picture with a red cap”. The subcategory salience is quite closely related to
odd_one_out since both of those correspond to selecting a response because it stands out.
The difference is that in the odd_one_out case, a selection is made based on a creature being
different from the other two based on a feature (exact match or mismatch), Opuesto a
being bigger or brighter, which is a relative difference. Often odd_one_out corresponded
to selecting the distractor and adapting an inverse interpretation of the speaker’s message,
p.ej., using the red hat message to refer to the only creature that is not wearing a red hat.
The tag preference was assigned when the explanation constituted a personal preference like
“Robots are cool”.
One could argue that preference should be categorized as guessing since the decision did
not involve reasoning about why the speaker sent a given message but instead broke the tie
using the participant’s own preference. Our motivation was the following: guessing is by its
nature not consistent, so we assigned an explanation to the guess category if, when presented
with the same trial again, possibly with different randomization, the participant could have
made a different selection. In the case of selecting a creature because it’s in the middle, desde
the screen location is randomized, a different creature could have been in the middle. En el
case of a personal preference for robots, sin embargo, we assume that the participant would have
consistently selected the robot in this situation if presented with this trial again.
MENTE ABIERTA: Descubrimientos en ciencia cognitiva
160
yo
D
oh
w
norte
oh
a
d
mi
d
F
r
oh
metro
h
t
t
pag
:
/
/
d
i
r
mi
C
t
.
metro
i
t
.
/
mi
d
tu
oh
pag
metro
i
/
yo
a
r
t
i
C
mi
–
pag
d
F
/
d
oh
i
/
i
/
/
.
1
0
1
1
6
2
oh
pag
metro
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
oh
pag
metro
_
a
_
0
0
0
7
7
pag
d
.
/
i
F
b
y
gramo
tu
mi
s
t
t
oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
Responses that involved a strategy that did not fall into one of the aforementioned catego-
ries were assigned to the subcategory other. An example would be a participant attempting to
reason but their reasoning not being sound. For the bottom panel of Figure 1, incorrect rea-
soning could be selecting the competitor since “if [the speaker] had meant to communicate
one of the other creatures, they would have used the purple monster message”.
Sometimes participants’ answers revealed that they misunderstood the instructions and
took the fact that a feature is inexpressible (es decir., that it cannot be referred to directly) a
mean that a creature that has the inexpressible feature could not be referred to at all. Un
example of an answer in that category for the top panel of Figure 1 would be “The speaker
could not choose the robot so it must be the green monster”. Such responses were labelled
misunderstood_instructions.
Answers where it was unclear what the participant meant, p.ej., very brief answers just stat-
ing their selection (“The robot”) were labeled unclear and excluded from further analysis. Nosotros
also excluded items where in their explanation the participant changed their mind as to their
answer since that indicates that their explanation does not reflect their original reasoning
estrategia.
We investigated how internally consistent participants were when they saw one of the
experimental items again during the strategy elicitation part of the experiment. We’d expect
people who used correct_reasoning to be consistent, while guessers may have picked the
target the first time and the competitor the second time, or the other way around. Eso es
indeed what we find. In the simple condition, collapsing across experiments since the pattern
is very similar for all experiments, 78 (88.6%) participants who used the correct_reasoning
strategy are consistent, selecting the target both times, and the remaining 10 (11.4%) selected
the competitor the first time and the target the second time around. As predicted, guessers
are more mixed: 17 (32%) participants selected the target both times (presumably by
chance), 15 (28.3%) participants selected the competitor both times, and the remaining
21 (39.6%) were inconsistent. Those breakdowns of responses were significantly different,
with correct_reasoners consistently selecting the target both times more often ((cid:1)2(2,
norte = 141) = 52.63, pag < 0.0001).
Among other_reason responses, it is notable that 33 (86.8%) participants who relied on
visual_resemblance picked the target, 3 (7.9%) picked the competitor both times, and 2
(5.3%) were inconsistent. Other other_reason subcategories were pretty mixed.
In the complex condition, 44 (84.6%) correct_reasoners picked the target both times and
8 (15.4%) were inconsistent and picked the target the second time. guessers, on the other
hand, are pretty evenly split, with 34 participants (37.8%) selecting the target both times,
29 (32.2%) of participants selecting the competitor both times, and 27 (30%) inconsistent
participants. The difference between correct_reasoners and guessers is again significant
χ2 2; N ¼ 142
ð
Þ ¼ 32:77; p < 0:0001
Þ:
ð
It is worth asking how reliable strategy explanations obtained through introspection are.
Post-hoc explanations have been criticized for not always accurately reflecting the reasoning
in the moment (Cushman, 2020). Nisbett and Wilson (1977) argued that humans lack access to
introspective processes in the moment and therefore post-hoc explanations are rationalizations
made after the fact. As an example, in the Wason selection task, people appear to be influ-
enced by what has been termed textitmatching bias: when they are asked to indicate which
cards need to be turned over to verify a logical rule (e.g., “If a card has a D on one side, it has a
3 on the other”), they are more like to select a card if it was mentioned before (a D or a 3 in this
OPEN MIND: Discoveries in Cognitive Science
161
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
/
.
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
example). However, subjects appear to not be conscious of this bias, as it never comes up in
post-hoc explanations (Evans, 2019).
We argue that, despite these limitations, post-hoc explanations have the potential to yield
important insights. In our experiments below, we show that strategy elicitation can help reveal
certain biases in the task, which some subjects are conscious of and which may be influencing
other subjects subconsciously. We also see that people whose explanation featured correct_
reasoning tend to be quite consistent in that they select the target both times when they
see the same trial twice, and in fact a lot more consistent than those who reported guessing.
Later in the paper we also show that there is considerable alignment between reported strat-
egies and performance once biases present in the stimuli are accounted for. So whether or not
the correct_reasoning explanation reflects correct hypothetical reasoning about alternatives
about in the moment, it does seem to be a good predictor of solving the task correctly. Also,
in these experiments we combine analysis of strategy explanations with careful manipulation
of the stimuli and comparison of effect sizes. The fact that these two measures together yield
consistent results gives more confidence in the validity of the provided explanations.
EXPERIMENT 1: REPLICATION OF FRANKE AND DEGEN (2016) WITH
REASONING ELICITATION
In this experiment, we replicated the reference game experiment by Franke and Degen (2016),
additionally eliciting participants’ reasoning strategies in order to get an insight into how par-
ticipants solve the task.
Participants
60 native speakers of English with an approval rating of at least 95%, were recruited through
the crowdsourcing platform Prolific.
Methods
Participants completed the reference game described in Reference game. After completing the
main experiment, participants saw one simple and one complex item again, presented in ran-
domized order. When they made their selection, a red box appeared around it, along with the
question “Why did you make that choice?” and a textbox. Participants’ strategies were then
annotated as described in Annotation scheme.
Results
One participant’s data was not saved on the server. Two further participants were excluded
from analysis for not having paid enough attention because their performance on the unam-
biguous filled trials was below 80%. The data from the remaining 57 participants entered the
analysis.
Franke and Degen (2016) fit a logistic mixed-effect regression model to verify that partici-
pants perform significantly better on the simple condition than they do on the complex, and
that their performance on the complex condition is above the chance baseline (the ambiguous
filler condition). The model in Franke and Degen (2016) included the maximal effect structure
that allowed it to converge, which, in addition to per-participant random intercepts, included
random slopes for message type (accessory or species) and trial number. It did not include a
random slope for condition, or a per-item random intercept, presumably for convergence rea-
sons. We believe, however, that it is important to include a random slope for condition in the
OPEN MIND: Discoveries in Cognitive Science
162
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
/
.
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
model since the motivation for individual-level modeling is that individual participants may
perform differently in the two conditions.
We also faced convergence issues when attempting to fit generalized linear mixed-effects
models for our data despite using an optimizer (bobyqa, Powell (2009)): the only random effect
structure with which the models for all 4 of our experiments converged was one that included
only per-participant random intercepts. Therefore, in order to be able to keep the random
effect structure maximal, we fit all our models using Bayesian regression using the brms pack-
age in R (Bürkner, 2017). Like in Franke and Degen (2016), we exclude from analysis trials on
which distractors were selected (for Experiment 1, that corresponds to 2.1% of trials) and
regress the binary correctness variable (whether target or competitor was selected) onto con-
dition (simple, complex, or ambiguous, dummy-coded, with complex as the reference level)1,
trial number, the interaction between trial number and condition, message type (creature or
accessory), and position of the target creature on the screen (left, center, or right, dummy-
coded, with left as the reference level). The random effect structure was maximal and included
per-participant random intercepts and random slopes for condition, message type and trial
number, and per-item random intercepts. The results are reported in Table 1.
As in the original study by Franke and Degen (2016), in the replication we find strong evi-
dence that participants performed better on the simple trials than on the complex ones (β =
2.16 (0.37), 95% CI [1.47, 2.94]); the effect size estimate is larger than in the original study
(2.16 vs. 1.28). We do not find a significant effect of the difference between the population-
level performance on the complex trials and the chance baseline (β = −0.19 (0.27), 95% CI
[−0.74, 0.34], whereas in the original study, that difference was significant (β = −0.44 (0.13),
p < 0.001). We think that this difference may in part be due to a different participant sample,
but also due to the original study using a different strategy to code the ambiguous condition.
Since in the ambiguous condition, two of the three creatures are identical, it needs to be
decided somehow which one is target and which one is competitor. Since Franke and Degen
(2016) report an uneven split in the ambiguous condition (46% of target choices vs. 51% of
competitor choices), it appears that their implementation was to randomly decide on each trial
which one of the two identical creatures was target and which one was competitor (let’s call
this method coin-flipping). We, on the other hand, ensure an even split at the population level
by having exactly half of the ambiguous non-distractor responses be correct. We chose this
randomization method since we believe that it is a better approximation of a chance baseline
and has smaller variance depending on the initialization2. Our simulation results suggest that
the difference between the ambiguous and the complex conditions in the original study was
amplified by the fact that Franke and Degen (2016)’s random initialization happened to result
in more competitor responses.
Like in the original study, we find evidence of participants being more likely to select the
target if it was in the center (β = 0.52 (0.15), 95% CI [0.23, 0.80]]) or on the right (β = 0.54
(0.15), 95% CI [0.25, 0.83]). We find no evidence of aneffect of trial or of the interaction
between trial and condition, indicating absence of learning effects, nor an effect of message
type (creature or accessory).
1 Franke and Degen (2016) use Helmert coding where simple is compared to the rest and complex is com-
pared to ambiguous, but since brms uses dummy coding, we set complex to be the reference level so that we
can obtain both of simple vs. complex and complex vs. ambiguous comparisons.
2 When we ran each randomization method 100 times on our replication data, the standard deviations of
effect size estimates were larger for the coin-flipping method. Related to that, the difference between the com-
plex and the ambiguous conditions came out significant 10 out of the 100 times when the coin-flipping method
was used and none of the times when we ensured a 50–50 split.
OPEN MIND: Discoveries in Cognitive Science
163
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
.
/
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Table 1.
original study by Franke and Degen (2016). Participants who did not reach 80%accuracy threshold on unambiguous trials were excluded from analysis.
Effect size estimates, standard errors, and 95% confidence intervals for the 4 experiments, as well as effect size estimates, standard errors, and p-values for the
Intercept
condition (simple
vs. complex)
condition (ambig
vs. complex)
trial number
simple vs. complex:
trial
ambig vs. complex:
trial
target pos
(middle vs. left)
target pos
(right vs. left)
msg type (accessory
vs. species)
F & D (2016) (51)
−0.15 (0.11),
p = 0.18
1.28 (0.12),
p < 0.0001
−0.44 (0.13),
p < 0.001
0.00 (0.00),
p < 0.3
0.00 (0.01),
p < 0.9
0.01 (0.01),
p < 0.33
1.28 (0.14),
p < 0.0001
0.74 (0.13),
p < 0.0001
−0.02 (0.12),
p < 0.85
replication (57)
−0.29 (0.27),
[−0.83, 0.24]
remapped (55)
0.33 (0.17),
[−0.00, 0.66]
all messages (56)
−0.21 (0.21),
[−0.63, 0.21]
2.16 (0.37),
[1.47, 2.94]
−0.19 (0.27),
[−0.74, 0.34]
0.00 (0.01),
[−0.01, 0.01]
0.00 (0.01),
[−0.01, 0.02]
0.00 (0.01),
[−0.01, 0.02]
0.52 (0.15),
[0.23, 0.80]
0.54 (0.15),
[0.25, 0.83]
0.25 (0.21),
[−0.15, 0.68]
0.46 (0.33),
[−0.17, 1.15]
−0.33 (0.17),
[−0.68, 0.01]
−0.00 (0.00),
[−0.01, 0.01]
0.02 (0.01),
[0.01, 0.04]
−0.01 (0.01),
[−0.02, 0.01]
0.20 (0.14),
[−0.07, 0.47]
0.09 (0.13),
[−0.17, 0.34]
−0.19 (0.14),
[−0.48, 0.09]
1.15 (0.26),
[0.66, 1.68]
−0.06 (0.24),
[−0.54, 0.41]
−0.00 (0.00),
[−0.01, 0.01]
−0.01 (0.01),
[−0.02, 0.01]
0.01 (0.01),
[−0.01, 0.02]
0.30 (0.13),
[0.04, 0.56]
−0.20 (0.13),
[−0.46, 0.06]
0.48 (0.20),
[0.08, 0.89]
shapes (60)
0.20 (0.20),
[−0.20, 0.61]
1.24 (0.30),
[0.67, 1.86]
−0.42 (0.21),
[−0.85, −0.03]
0.01 (0.00),
[0.00, 0.02]
−0.02 (0.01),
[−0.03, −0.00]
−0.02 (0.01),
[−0.03, −0.00]
0.38 (0.13),
[0.11, 0.64]
−0.04 (0.14),
[−0.31, 0.23]
0.22 (0.13),
[−0.03, 0.47]
O
n
t
h
e
I
m
p
o
r
t
a
n
c
e
o
f
E
l
i
c
i
t
i
n
g
P
a
r
t
i
c
i
p
a
n
t
s
’
R
e
a
s
o
n
i
n
g
M
a
y
n
a
n
d
D
e
m
b
e
r
g
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
.
/
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
O
P
E
N
M
N
D
I
:
i
D
s
c
o
v
e
r
i
e
s
i
n
C
o
g
n
i
t
i
v
e
S
c
e
n
c
e
i
1
6
4
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
Each participant’s average performance in Experiment 1, with the color corresponding
Figure 2.
to the strategy label on the simple condition. The red dots correspond to participants who applied a
strategy other than guess or correct_reasoning. Quite a few of the red dots are pretty far on the right,
indicating that these participants got a trial correct for the wrong reason at least some of the time; in
other words, these participants’ performance is likely inflated. Based on the performance alone,
however, they are indistinguishable from correct reasoners.
We now take a look at the annotations. Inter-annotator agreement was substantial, Cohen’s
κ = 0.65, 95% CI [0.56, 0.75]. In the simple condition, 1 response labeled exclude and 8
responses labeled unclear out of 57 annotations were excluded. 48of the 57 annotations
(84%) entered the analysis. In the complex condition, 2 exclude and 5 unclear responses were
excluded, and 50 of the 57 annotations (88%) entered the analysis.
As can be seen in the top panel of Figure 3, 27.5% of correct responses (and 27.1% of total
responses) in the simple condition fell into the other_reason category, suggesting that some
participants arrived at the correct answer not via the assumed reasoning. When we examine
the correct other_reason responses (bottom panel of Figure 3), we see that the majority of
them used the visual_resemblance strategy. For the remainder of this paper, we will be
discussing the simple condition, which turned out to be more susceptible to bias and more
helpful for revealing it, but corresponding graphs for the complex condition can be found in
the Appendix.
In Figure 2, we plot participants’ average performance on the simple and complex trials.
We see that the majority of participants who gave an other_reason explanation in the simple
condition exhibit near-ceiling performance, so based on performance alone they would be
indistinguishable from correct reasoners, although their answers are correct due to factors
unrelated to RSA-style reasoning about alternatives at least some of the time.
EXPERIMENT 2: REMAPPED VERSION
We observed that in the original stimuli, the pairs of expressible features (that is, the features for
which there is a message available to the speaker) constitute a kind of conceptual grouping –
the two monsters are expressible as messages but the robot is not, the two hats are expressible
as messages but the scarf is not.
The inexpressible features, therefore, because they are the odd-ones-out – robot is the only
non-monster creature and the scarf is the only non-hat accessory – may stand out a lot more.
To further probe the effect of stimuli on this task, we, therefore, swapped the expressible and
inexpressible features around, breaking this conceptual grouping: the robot and the scarf were
swapped for the purple monster and the blue hat respectively. The swapping is illustrated in
OPEN MIND: Discoveries in Cognitive Science
165
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
/
.
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
.
/
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Figure 3. Participants’ explanations by tag in the simple condition (top – all explanations, bottom – other_reason explanations).
Figure 5: along the creature dimension, we made the robot expressible and the purple monster
inexpressible instead, and along the accessory dimension, we made the scarf expressible and
the blue hat inexpressible instead.
So in this experiment, the simple trial in Figure 1 looks as follows (illustrated in Figure 4):
the message stays the red hat since no swapping had been performed there; the target is now a
purple monster with a red hat (since we swapped the robot and the purple monster around),
the competitor remains the same since no swapping has been performed for either the green
monster or the red hat, and the distractor is now a robot with a blue hat (since the purple
monster is swapped with the robot and the scarf is swapped with the blue hat). Therefore,
underlyingly, all trials remained the same as in Experiment 1, the only difference is that the
inexpressible features are represented with different images. We used the images from the orig-
inal study so no changes were performed e.g., to the hat orientation.
Participants
60 native speakers of English with an approval rating of at least 95%, were recruited through
the crowdsourcing platform Prolific.
Methods
We changed which features were expressible as messages, swapping the robot and the scarf
for the purple monster and the blue hat respectively. Otherwise the experiment was identical
to the replication. An example of how the swapping was performed is shown in Figure 4.
OPEN MIND: Discoveries in Cognitive Science
166
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
Figure 4. An example of a remapped trial (the simple condition from Figure 1). Because the robot
gets swapped with the purple monster, the target becomes purple monster with the red hat, the
competitor stays the same, and the distractor becomes a robot(swapped with the purple monster)
with a blue hat (swapped with a scarf ).
i
/
.
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Results
One participant’s data was not saved on the server. Four further participants were excluded
from analysis for not having paid enough attention because their performance on the unam-
biguous filler trials was below 80%. The data from the remaining 55 participants entered the
analysis.
As can be seen in Table 1, we observe a stark change in the effect for simple vs. complex
condition: the effect size is a lot smaller than in the replication, and the effect is no longer
significant (β = 0.46 (0.33), 95%CI [−0.17, 1.15]). Like in the replication, performance on
the complex condition is not significantly different from the chance baseline (β = −0.33
(0.17), 95% CI [−0.68, 0.01]).
Strategy responses were again annotated by two annotators, one of whom was blind to the
purpose of the experiment. Inter-annotator agreement was again substantial, Cohen’s κ = 0.77,
95% CI [0.68, 0.86]. In the simple condition, 4 responses labeled exclude and 6 responses
labeled unclear out of 55 annotations were excluded. 45 of the 55 annotations (82%) entered
the analysis. In the complex condition, 3 exclude and 10 unclear responses were excluded,
and 42 of the 55 annotations (76%) entered the analysis.
OPEN MIND: Discoveries in Cognitive Science
167
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
.
/
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Figure 5.
For Experiment 2, we changed which features were expressible as messages.
When we examine the breakdown of annotation tags (Figure 3), we see that a smaller pro-
portion of correct answers and a larger proportion of incorrect answers is comprised of other_
reason responses, and while in the original study, visual_resemblance led participants to
always select the correct answer, in the remapped version, it was now incorrect some of
the time; salience, preference, and odd_one_out reasoning now pointed participants to
the competitor instead of the target. For instance, in the simple trial depicted in Figure 4,
in the original version, the creature that is the odd-one-out and therefore a more salient one
is the robot, which happens to be the target, so if a participant selects it based on salience or
odd-one-out reasoning, they will happen to be correct. In the remapped version, on the other
hand, the target is less visually salient because it is one of the two monsters, therefore, the
same salience or xtitodd_one_out reasoning may lead the participant to select the distractor,
the robot, which would be incorrect. This suggests that in the remapped version of the task,
some of the existing biases in the stimuli were now having the opposite effect from the original
study, that is, “deflating” participant performance.
EXPERIMENT 3: ORIGINAL EXPERIMENT, ALL MESSAGES AVAILABLE
We wanted to know how often participants would make the right choice coincidentally in the
original experiment, in a setting where correct hypothetical reasoning about alternatives is
made impossible, thus isolating stimuli effects.
In order to isolate stimuli effects and make correct_reasoning impossible, we made all six
features expressible. As a result, correct_reasoning is no longer possible in the simple
OPEN MIND: Discoveries in Cognitive Science
168
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
condition, since there are now unambiguous messages available to refer to the target as well as
the competitor. For example, in the top panel of Figure 1, both the target and the competitor
can now be referred to unambiguously via the messages “robot” and “green monster” respec-
tively, so the message “red hat” is now completely ambiguous.
Note that in the complex condition, correct_reasoning is still possible because the target still
cannot be unambiguously identified by a message, while both the competitor and the distrac-
tor can be.
Thus, in this case, none of the formal models should be able to solve the simple implicature
condition and, like in the original setup, only L2 should be able to solve the complex one.
Participants
60 native speakers of English with an approval rating of at least 95%, were recruited through
the crowdsourcing platform Prolific.
Methods
One participant’s data was not saved on the server. Three further participants were excluded from
analysis for not having paid enough attention because their performance on the unambiguous
filled trials was below 80%. The data from the remaining 56 participants entered the analysis.
Apart from the feature expressibility manipulation, the experiment was identical to the
replication.
Results
Table 1 shows that, with all messages available, the correct option is still selected for reasons
that are not correct hypothetical reasoning quite often, as evidenced by the larger significant
effect of condition (complex vs. simple, β = 1.15 (0.26), 95% CI [0.66, 1.68]), notably, con-
siderably more often than in the remapped experiment (β = 0.46 (0.33), 95% CI [−0.17, 1.15]).
Inter-annotator agreement was again substantial, Cohen’s κ = 0.67, 95% CI [0.57, 0.78]. In
the simple condition, 1 response labeled exclude and 7 responses labeled unclear out of 56
annotations were excluded. 48 of the 56 annotations (86%) entered the analysis. In the com-
plex condition, 6 unclear responses were excluded, and 50 of the 56 annotations (89%)
entered the analysis.
When we take a look at the performance by tag (Top panel of Figure 3), we see that when
participants used an other_reason strategy in the simple condition, the vast majority of the time
(89.2% of cases) their strategy coincidentally led them to select the target. When we examine
the other_reason tags in more detail, we see that the strategy that resulted in accidental correct
responses is visual_resemblance, that is, similarity of the target to the message based on the
head of a scarf-wearing creature being uncovered, similar to the message (that creature with-
out an accessory) or the orientation of the hat message matching that of the referent.
This corroborates the claim that factors unrelated to reasoning, most importantly incidental
visual resemblance of the message to the target, inflate performance in the original experiment.
EXPERIMENT 4: MITIGATING BIAS BY USING ABSTRACT STIMULI
Having shown that the visual resemblance bias accounts for the vast majority of coincidental
target choices, we repeated the original experiment with a version of the stimuli that we hoped
would be less susceptible to the visual_resemblance bias.
OPEN MIND: Discoveries in Cognitive Science
169
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
/
/
.
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
Participants
60 native speakers of English with an approval rating of at least 95%, were recruited through
the crowdsourcing platform Prolific.
Methods
There were no exclusions based on accuracy in this experiment, as all 60 participants scored
above 80% on the unambiguous filler trials.
Having identified the biases in the original stimuli, we attempted to mitigate them. We
hypothesized that using more abstract stimuli, of the kind used in the original RSA study (Frank
& Goodman, 2012), would not create opportunity for a visual resemblance bias between the
message and the target.
We therefore ran a version of the experiment with abstract stimuli, using geometric shapes
and colors instead of creatures and accessories: square, triangle and circle corresponding to
the robot, green monster and purple monster respectively, and the colors blue, green and red
corresponding to the scarf, red hat and blue hat respectively. When the message was a color, it
was represented by a color tube, and when it was a shape, it was represented by a shape
contour. An example of trial from Experiment 4, corresponding directly to Figure 1 from Exper-
iment 1, is included in Figure 6.
Thus, underlyingly, the experiment remained the same as the original, the only difference
being what images were used to represent the messages and the referents.
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
.
/
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Figure 6. Example of a simple and a complex trial for Experiment 4. These trials directly corre-
spond to those depicted in Figure 1 for Experiment 1.
OPEN MIND: Discoveries in Cognitive Science
170
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
Results
Inter-annotator agreement was again substantial, Cohen’s κ = 0.80, 95% CI [0.71,0.89]. In the
simple condition, 2 responses labeled exclude and 3 responses labeled unclear out of 60
annotations were excluded. 55 of the 60 annotations (92%) entered the analysis. In the com-
plex condition, 1 exclude and 6 unclear responses were excluded, and 53 of the 60 annota-
tions (88%) entered the analysis.
We see in Figure 3 that with abstract stimuli, barely any other_reason explanations are
given, indicating that people do not seem to rely on clues like visual similarity but instead
employ either guessing or correct hypothetical reasoning about alternatives, as assumed by
formal probabilistic models. Only one visual_resemblance response is given, and it does hap-
pen to accidentally lead to selecting the target: when the message is a circle, the blue circle
isselected with the justification that the blue color is “nearest to clear”, so a clear circle is most
similar to the blue circle.
We conclude the abstract stimuli appear to be less susceptible to biases that inflated per-
formance on the original experiment, and in particular, the visual resemblance bias. Therefore,
we can have more confidence that the task results more accurately reflect participants’ prag-
matic reasoning ability. In terms of the effect sizes, this study’s results closely resemble those of
the original Franke and Degen (2016) study and our replication: β = 1.24 (0.30), 95% CI [0.57,
1.86] for simple vs. complex condition in this study vs. β = 2.16 in the replication; like in the
original study and unlike in our replication, there is evidence for a difference between the
complex condition and chance (β = −0.42 (0.21), 95% CI [−0.85, −0.03]). Thus, it appears
that the original study, despite the biases, fairly accurately captures the effects at the popula-
tion level. However, now we can be more certain that they accurately reflect individual-level
performance as well.
DISCUSSION
In this study, we elicited participants’ strategies and manipulated which messages were
expressible to probe the stimuli first introduced by Degen and Franke (2012) and used in
Degen et al. (2013) and Franke and Degen (2016) for biases. Effect sizes varied greatly
between experiments and appear to be inflated in the original task by factors unrelated to rea-
soning. The most prominent bias was visual resemblance, where participants selected the
creature most similar to the message, which in most cases happened to be the target. We then
replicated the experiment with a version of the stimuli that we showed to be less susceptible to
the visual resemblance bias.
There remains the question of whether the effect sizes in Experiments 1 and 4 are even
comparable: the stimuli in the original task have biases, inflating the effect of condition, but
they are also visually more complex, potentially leading to higher task difficulty and lower
effect of condition than the abstract stimuli in Experiment 4. In other words, there appears to
be no way to know exactly how much smaller the “true” effect size of condition in the
original experiment would have been if there were no stimuli-related biases, because when
we use different stimuli that do not have the biases, those stimuli also have different per-
ceptual properties which affect task complexity, and, correspondingly, effect sizes. We
would expect the “true" effect sizes to lie somewhere between those in Experiment 1
and Experiment 2, since in the original experiment, biases inflate participants’ performance,
and in the remapped version (Experiment 2), biases still exist but they lead participants to
choose incorrectly.
OPEN MIND: Discoveries in Cognitive Science
171
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
/
.
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
The final experiment, which used different stimuli, was shown to be much less susceptible
to biases present in the original stimuli. That means that we can have more confidence that
participants’ performance on the task is a more reliable reflection of their reasoning ability, at
the population level, and especially at the individual level.
To investigate individual-level performance in Experiments 1 and 4 in more detail, we ran
Latent Profile Analysis on participants’ average performance on simple and complex condi-
tions using the tidyLPA package in R to obtain reasoner classes. We decided to use LPA and
not the Bayesian model from Franke and Degen (2016) because the latter is too strict for our
purposes: it presupposes the existence of only three classes (the theoretically motivated classes
L0, L1, and L2), and since we saw that many participants used other strategies, we wanted to
have a data-driven way of identifying classes which may potentially include classes other than
the ones corresponding to the three probabilistic listener models.3 Recall that Experiments 1
and 4 are underlyingly the same and the only difference is what images are used as the rep-
resentation of the messages and referents, hence L0, L1 and L2 RSA models make the same
predictions for each trial of the two experiments.4
For both experiments, the best fit, as measured by AIC and BIC, was obtained by models
with 4 classes (model fit for each number of classes can be found in the Appendix).5 As can be
seen in Figure 7, three of the four identified classes approximately correspond to the predic-
tions of the three formally defined reasoning types from Franke and Degen (2016), L0 (at
chance on both conditions), L1 (can solve the simple but not the complex trials), and L2
(can solve both conditions). Additionally, there’s the smallest fourth class, consisting of 4
and 2 participants respectively for the two experiments, of participants who perform below
chance on the simple condition and at or below chance on the complex. Strategy annotations
also approximately match our expectations, but the pattern is more pronounced for Experiment
4: participants assigned to the class L0 guess in the simple condition while participants assigned
to classes corresponding to L1 and L2 apply correct reasoning in the simple condition. The
corresponding scatterplot for the complex condition can be found in the Appendix.
We see that in Experiment 1, of the participants whose strategy in the simple condition was
labeled other_reason, 3 are assigned to the L0 class, 5 to L1, and 3 to L2. As we saw in the
above experiments, these participants’ performance was likely inflated by factors unrelated
to reasoning; therefore, they were possibly estimated to have a more advanced reasoning type
than they actually have. In Experiment 1, only one participant used an other_reason strategy in
the simple condition; that participant was assigned to L0, although they are on the border with
L2 (bottom panel of Figure 7). As we saw in Experiment 4, this participant happened to choose
the target for the wrong reason (a creative application of the visual_resemblance strategy);
therefore, their performance is also inflated. However, the fact that this is only 1 participant
is a large improvement compared to Experiment 1.
When we look at the breakdown by tag for the identified classes (Figure 8), we see that the
responses in Experiment 4 much more cleanly map onto the corresponding formal classes than
3 Incidentally, applying Franke and Degen (2016)’s class assignment model to our data yielded uninterpretable
classes.
4 That is also true for Experiment 2, but not for Experiment 3, where neither L1 nor L2 are predicted to be able
to solve the simple condition since it is made completely ambiguous.
5 For Experiment 4, the 5-class model has the best fit according to AIC whilethe 4-class model has the best fit
according to BIC. The only difference between those two clusterings is that for the 5-class model, one partic-
ipant whose accuracy is 0 for both simple and complex is in their own separate class. Therefore, for compa-
rability with Experiment 1, we use the 4-class model.
OPEN MIND: Discoveries in Cognitive Science
172
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
/
/
.
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
.
/
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
.
/
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Figure 7. Classes of participants identified by LPA (top – Experiment 1 (replication), bottom – Experiment 4 (debiased stimuli)). Three of the
identified classes approximately correspond to the theoretical predictions of listener models L0, L1, and L2. The fourth LPA class, which is the
smallest for both experiments, corresponds to participants who perform below chance level in the simple condition.
the responses from Experiment 1 (L0 guesses forboth types of trials, L1 applies correct reasoning
about alternatives for the simple condition and guesses on the complex, and L2 applies correct
reasoning in both conditions), again supporting the claim that the debiasing of the stimuli was
successful and constitute a better proxy for participants’ reasoning ability as defined by formal
probabilistic models. In order to quantitatively corroborate this finding, we conducted cluster
analysis. Each participant was assigned a class label based on performance. For that, we used
the four classes identified by LPA: L0, L1, L2, and other (people who ended up in the fourth,
below_chance class). Each participant was also assigned a class label based on the annotated
strategies: L0 if they used the guess strategy in both conditions, L1 if their strategy was labeled
correct_reasoning in the simple condition and guess in the complex, and L2 if their strategy in
both condition was correct_reasoning. Finally, participants who used an other_reason strategy
in either condition were given the label other. Participants whose strategy was labeled unclear
or exclude were excluded from this analysis, resulting in 42 participants for Experiment 1 and
47 participants for Experiment 4 entering the analysis. The motivation for the other label is that
we would ideally want people who use other reasoning to be distinguishable from the three
reasoning types based on performance. In order to see how well the performance-based and
annotation-based classes aligned, we computed cluster homogeneity and completeness for
Experiment 1 and Experiment 4, with performance class labels as reference. For Experiment 1,
OPEN MIND: Discoveries in Cognitive Science
173
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
.
/
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
.
/
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Figure 8. Tags in each condition for the classes identified by LPA (top – Experiment 1 (replication), bottom – Experiment 4 (debiased stimuli)).
homogeneity is 0.13 and completeness is 0.13, and for Experiment 4, homogeneity is 0.56 and
completeness is 0.51. This further supports the observation that more abstract stimuli lead to a
better alignment between performance and the underlying strategy.
While it is unrealistic to expect no stimuli effects, it is important to know what biases are
present and how strong they are because that influences the conclusions that are drawn. This
can be quite important, in particular, for individual difference studies, where spurious corre-
lations of performance with individual measures may emerge or real ones may not be identi-
fied due to such stimulus effects. For instance, the study by Mayn and Demberg (2022), which
explored individual differences that modulate participants’ reasoning complexity in the rea-
soning game and found an effect of non-verbal IQ and reflectivity, was conducted using the
original biased stimuli, suggesting that it would be important to repeat the analysis and confirm
that the findings still hold and that the effect of IQ and reflectivity is not driven or inflated by
the stimuli biases.
We used post-hoc strategy elicitation to probe the experimental design of the reference
game for biases. Post-hoc explanations have been criticized for not to always accurately
reflecting the reasoning in the moment (e.g., Cushman, 2020). If we look at the relationship
between participants’ performance and the elicited strategy explanation in Experiment 1
(Figure 2), we see that while most participants whose strategy explanation on the simple trial
was labeled correct_reasoning performed near ceiling on average, there are a few participants
who show chance-level average performance. One reason why this might be the case is that
the fact that a participant applied a certain strategy, e.g., xtitcorrect_reasoning, on a given trial
does not mean that they applied it successfully on all trials. Another possibility is that these
participants were actually guessing in the moment and came up with the correct reason for
their selection after the fact. This suggests that these explanations on a single trial are not a
OPEN MIND: Discoveries in Cognitive Science
174
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
perfect proxy for participants’ reasoning strategies throughout the experiment. However, the
fact that participants who provided a correct_reasoning reasoning on the simple trial have
much better average performance than guessers (0.87 (0.17) vs. 0.66 (0.32)) suggests that,
while imperfect, post-hoc explanations are a proxy for the strategies participants used during
the experiment. While there are limitations to elicited post-hoc explanations, they can be a
useful tool for revealing biases in an experimental design that could otherwise be missed,
as was the case in this study.
Recently, Ryzhova et al. (2023) applied a method quite similar to the strategy elicitation
method used in this work to a different pragmatic phenomenon, atypicality inferences, in order
to gain more insight into how participants solved the task. When an event which is very pre-
dictable in a given context is overtly mentioned, it may be inferred that the event is actually
atypical for the referent (e.g., “John went to the grocery store. He paid the cashier” may result
in the atypicality inference that John does not usually pay at the grocery store). In addition to
collecting ratings of how typical participants expected the redundantly mentioned activity to
be for the referent, the authors asked participants to justify the ratings they gave. Based on the
provided justifications, they were able to make a distinction between two groups of partici-
pants which would have been unidentifiable based on ratings alone: those who did not make
the atypicality inference and those who initially made it but then rejected it (e.g., “Not paying
would be stealing, therefore it’s unlikely that John does not pay”) and gave a high typicality
rating. Adding explanations to the experimental measure allowed the authors to identify dif-
ferent reasoning which results in the same performance, much like in the current work where
we show that high task performance may be the result of correct reasoning but also of unin-
tended biases in the task where people get away with applying simpler strategies. The fact that
strategy elicitation has proved useful for two different pragmatic phenomena – the reference
game in our work and in the case of atypicality inferences – suggests that it can be useful more
broadly and may be worth applying when investigating other pragmatic phenomena and seek-
ing to better understand the processes underlying pragmatic reasoning.
More broadly, we believe that, when designing experiments, it is very important to ask the
question how to make sure that the measure that is being used is indeed tapping into the con-
struct of interest. Might there be other explanations for why participants are behaving a certain
way? Here, we showed two ways that can be used for probing an experimental design: careful
manipulation of the stimuli and strategy elicitation. Of course, what methods are appropriate
depends on the experimental setup: for instance, strategy elicitation seems to mostly apply when
the studied phenomenon is expected to involve relatively conscious reasoning or reflection;
manipulating stimuli in systematic ways and examining how that affects the results seems to
be applicable more broadly. There may be biases which are not reported because subjects
are not aware of them, just like participants don’t report matching bias in Wason’s selection task.
Therefore, we believe that it is worthwhile to explore the application of on-line measures, such
as eye gaze, to getting a better sense of participant strategies, and learn more about how those
relate to performance and explanations obtained through introspection. How people actually
solve a task is a difficult question as we cannot look directly into people’s heads and have to
rely on proxies like performance or explanations, but it is a very important one, and therefore we
should make it a priority to search for possible ways to get closer to the answer.
ACKNOWLEDGMENTS
We would like to thank the anonymous reviewers for their helpful comments and suggestions
and AriaRay Brown for assistance with annotation and helpful conversations.
OPEN MIND: Discoveries in Cognitive Science
175
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
/
.
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
.
/
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
FUNDING INFORMATION
This project is supported by funding from the European Research Council (ERC) under the European
Union’s Horizon 2020 Research and Innovation Programme (Grant Agreement No. 948878).
REFERENCES
Bürkner, P.-C. (2017). Advanced bayesian multilevel modeling with
the R package brms. arXiv:1705.11123. https://doi.org/10.48550
/arXiv.1705.11123
Cushman, F. (2020). Rationalization is rational. Behavioral and Brain
Sciences, 43, e28. https://doi.org/10.1017/S0140525X19001730,
PubMed: 31133084
Degen, J., & Franke, M. (2012). Optimal reasoning about referential
expressions. In Proceedings of SemDIAL (pp. 2–11). SEMDIAL.
Degen, J., Franke, M., & Jäger, G. (2013). Cost-based pragmatic
inference about referential expressions. Proceedings of the
Annual Meeting of the Cognitive Science Society, 35, 376–
381.
Evans, J. S. B. (2019). Reflections on reflection: the nature and func-
tion of type 2 processes in dual-process theories of reasoning.
Thinking & Reasoning, 25(4), 383–415. https://doi.org/10.1080
/13546783.2019.1623071
Frank, M. C., & Goodman, N. D. (2012). Predicting pragmatic rea-
soning in language games. Science, 336(6084), 998. https://doi
.org/10.1126/science.1218633, PubMed: 22628647
Franke, M., & Degen, J. (2016). Reasoning in reference games:
Individual- vs. population-level probabilistic modeling. PLoS
One, 11(5), e0154854. https://doi.org/10.1371/journal.pone
.0154854, PubMed: 27149675
Goodman, N. D., & Stuhlmüller, A. (2013). Knowledge and impli-
cature: Modeling language understanding as social cognition.
Topics in Cognitive Science, 5(1), 173–184. https://doi.org/10
.1111/tops.12007, PubMed: 23335578
Kao, J. T., & Goodman, N. D. (2015). Let’s talk (ironically) about the
weather: Modeling verbal irony. Cognitive Science.
Kao, J. T., Wu, J. Y., Bergen, L., & Goodman, N. D. (2014). Nonlit-
eral understanding of number words. Proceedings of the National
Academy of Sciences, 111(33), 12002–12007. https://doi.org/10
.1073/pnas.1407479111, PubMed: 25092304
Mayn, A., & Demberg, V. (2022). Individual differences in a prag-
matic reference game. Proceedings of the Annual Meeting of the
Cognitive Science Society, 44, 3016–3022.
Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can
know: Verbal reports on mental processes. Psychological Review,
84(3), 231–259. https://doi.org/10.1037/0033-295X.84.3.231
Powell, M. J. (2009). The BOBYQA algorithm for bound con-
strained optimization without derivatives. Cambridge NA Report
NA2009/06. University of Cambridge.
Ryzhova, M., Mayn, A., & Demberg, V. (2023). What inference do
people actually make upon encountering informationally redun-
dant utterances? An individual difference study. Proceedings of
the Annual Meeting of the Cognitive Science Society.
Sikos, L., Venhuizen, N. J., Drenhaus, H., & Crocker, M. W. (2021).
Reevaluating pragmatic reasoning in language games. PLoS One,
16(3), e0248388. https://doi.org/10.1371/journal.pone.0248388,
PubMed: 33730097
APPENDIX
Annotations for the Complex Condition
In Figure 9 we provide by-tag results in the complex condition for the two experiments. It is
notable that there is a large proportion of odd_one_out responses. Those correspond to the
participant interpreting the message inversely and selecting the distractor. For example, in
the bottom panel of Figure 1, a participant might interpret the message “red hat” as meaning
“the only creature without a red hat” and select the distractor, the purple monster with a scarf.
The fact that we observe a smaller proportion of other_reason responses in Experiment 4 is a
good sign because the employed strategies are predominantly correct_reasoning and guess, as
is assumed by formal probabilistic models.
LPA Model Fit
Table 2 and Table 3 show fits for models with 1 through 6 latent classes for Experiments 1 and
4 respectively, as measured by AIC and BIC. The 4-class model provides the best fit for Exper-
iment 1, and the classes approximately correspond to the predictions of the three probabilistic
listener models: L0, L1, and L2. For Experiment 4, the 4-class model has the best fit according to
BIC and the 5-class model has the best fit according to AIC. The difference between the 4- and
5-class models is that in the 5-class model, one participant who utilized the odd-one-out strat-
egy on all trials, resulting in 0 accuracy on both simple and complex trials, is assigned to their
own class. Since that is the only difference, for comparability in the article we discuss 4-class
models for both experiments.
OPEN MIND: Discoveries in Cognitive Science
176
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
.
/
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
/
.
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
/
.
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Figure 9. Participants’ explanations by tag in the complex condition (top – all explanations, bottom – other_reason explanations).
LPA Classes (Complex Condition)
The 4 latent classes for Experiments 1 and 4 with annotation tags from the complex condition
are included in Figure 10. In Experiment 1 (top panel), we observe a large number of people
who use the odd_one_out strategy, thus performing below chance. They are assigned to the
class corresponding to the L1 model, where the reasoner can solve simple implicatures but
guesses randomly on complex ones. That is not ideal since the odd_one_out strategy where
the message is interpreted inversely is a different, more complex strategy than guessing. In
Experiment 4 (bottom panel), we get cleaner results: there are only a few people who use
odd-one-out reasoning; for the most part, L0 and L1 classes consist of people whose strategy
in the complex condition was guessing, and the L2 class consists of people who applied the
correct_reasoning strategy.
Table 2.
Fit of LPA models for Experiment 1.
Num. classes
1
2
3
4
5
6
AIC
13.89
−4.88
−6.37
−18.83
−16.55
−11.62
OPEN MIND: Discoveries in Cognitive Science
BIC
22.06
9.42
14.06
7.73
16.14
27.2
177
On the Importance of Eliciting Participants’ Reasoning Mayn and Demberg
Table 3.
Fit of LPA models for Experiment 4.
Num. classes
1
2
3
4
5
6
AIC
30.27
17.96
8.55
−6.16
−9.66
−4.87
BIC
38.65
32.62
29.5
21.07
23.85
34.92
l
D
o
w
n
o
a
d
e
d
f
r
o
m
h
t
t
p
:
/
/
d
i
r
e
c
t
.
m
i
t
.
/
e
d
u
o
p
m
i
/
l
a
r
t
i
c
e
-
p
d
f
/
d
o
i
/
i
.
/
/
1
0
1
1
6
2
o
p
m
_
a
_
0
0
0
7
7
2
1
3
3
8
5
5
o
p
m
_
a
_
0
0
0
7
7
p
d
.
/
i
f
b
y
g
u
e
s
t
t
o
n
0
7
S
e
p
e
m
b
e
r
2
0
2
3
Figure 10. Classes of participants identified by LPA (top – Experiment 1 (replication), bottom – Experiment 4 (debiased stimuli)). Shapes
correspond to the tags assigned to the reasoning strategies in the complex condition.
OPEN MIND: Discoveries in Cognitive Science
178
