University of Southern California Dissertations and Theses

When photos backfire: truthiness and falsiness effects in comparative judgements

(USC Thesis Other)

When photos backfire: truthiness and falsiness effects in comparative judgements

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Running head: TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS
When Photos Backfire: Truthiness and Falsiness Effects in Comparative Judgements
Lynn Zhang
University of Southern California
A Thesis Presented to the Faculty of the Dornsife School of Letters, Arts and Sciences
University of Southern California
In Partial Fulfillment of the Requirements for the Degree of
Master of Arts (Psychology)
December 2019
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS ii

Table of Contents

Abstract 3

Introduction 4
Background 4
Present Research 8

Experiment 1 9
Methods 9
Results and Discussion 12

Experiment 2 13
Methods 15
Results 18
Discussion 20

Experiment 3 22
Methods 22
Results and Discussion 23

Meta-analysis of Experiment 2 and 3 25

General Discussion 27

References 33

Supplementary Materials 38

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 1
Abstract
Claims are more likely to be judged true when presented with a related nonprobative photo
(Newman, Garry, Bernstein, Kantner, & Lindsay, 2012). According to a fluency account, related
photos facilitate processing and easy processing fosters acceptance of the claim. According to an
illusion-of-evidence account, related photos may increase acceptance of the claim because they
are treated as tentative supportive evidence. To disentangle these potential mechanisms
participants read statements about the future price of easy- or difficult-to-visualize commodities.
Consistent with fluency predictions, photos only increased the acceptance of claims when the
target was difficult to visualize without a photo (Experiment 1). In Experiments 2 and 3 we used
comparative claims. In forming comparative judgments, people first assess attributes of the
linguistic subject of comparison and subsequently compare them to attributes of the referent
(Tversky, 1977). Photos of the linguistic subject in a sentence should facilitate, whereas photos
of the linguistic referent impair, the fluency of this sequence. Consistent with fluency
predictions, photos of the subject in a claim increased acceptance of comparative claims relative
to a no-photo condition (a truthiness effect), and more so when the subject was otherwise
difficult to visualize. Photos of the referent decreased acceptance of comparative claims relative
to a no-photo condition (a falsiness effect), and more so when the subject of comparison was
otherwise easy to visualize. All results are consistent with a context sensitive fluency account:
increases in fluency foster, and decreases in fluency impair, acceptance of a claim as true.

Keywords: truthiness, falsiness, fluency, nonprobative photos, comparative judgements

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 2
When Photos Backfire: Truthiness and Falsiness in Comparative Judgements
The liquid metal inside a thermometer is magnesium. Without consulting external
sources, most people will not know whether this statement is true or false. Previous research
found that statements of this type are more likely to be accepted as true when they are
accompanied by a nonprobative photo of the subject of the claim, here a thermometer (Newman,
Garry, Bernstein, Kantner, & Lindsay, 2012). Although nonprobative photos provide no
meaningful evidence that the claims are true, they produce a sense of “truthiness”, a feeling of
truth that is not based on facts but can nevertheless persist over time (Fenn, Newman, Pezdek, &
Garry, 2013).
One explanation of how a related but nonprobative photo can inflate the truthiness of a
claim is that related photos, regardless of their informational value, facilitate the conceptual
processing of the claim by making it more vivid and imaginable in the recipient’s mind. Failing
to realize that the metacognitive ease arises from an irrelevant source, the experienced fluency
may convey that the claim “feels right”. Indeed, numerous variables that facilitate ease of
processing -- from repetition (e.g., Hasher, Goldstein, & Toppino, 1977) to color contrast (e.g.,
Reber & Schwarz, 1999) and rhyme (e.g., McGlone & Tofighbakhsh, 2000) -- have been found
to increase perceived truth (for reviews, see Schwarz, 2015, 2018).
Another possibility is that nonprobative photos create an illusion of evidence. This may
be the case for several reasons. First, most photos capture real events and photos have long been
offered as evidence that an event has truly occurred (Mnookin, 1998). Second, photos are rich in
perceptual and semantic details, just like real events. Given that people are more likely to assume
that a mental event (e.g., an episodic memory) reflects reality when it features such details
(Johnson, Foley, Suengas, & Raye, 1988), photos may be treated as evidence, consistent with the
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 3
familiar claim that “seeing is believing”. Third, from a conversational perspective, speakers are
expected to communicate only information that is relevant to the ongoing conversation (Grice,
1975; Sperber & Wilson, 1986). People find violations of this relevance principle unnatural
(Davies & Katsos, 2009) and even children as young as three years old are sensitive to such
violations (Eskritt, Whalen, & Lee, 2010). From this perspective, recipients may perceive photos
that are related to a claim as the speaker’s attempt to offer supportive evidence -- or why else
would the photo be presented? This is particularly likely when the speaker is perceived as a
cooperative and trustworthy communicator, which is usually the case when messages are
presented by a researcher (Schwarz, 1994, 1996). Hence, people may treat nonprobative photos
that are presented with a claim as tentative evidence, unless the photo is clearly unrelated to the
claim. When the photo is clearly unrelated, truthiness effects are not observed (Newman, Garry,
Unkelbach, Bernstein, & Lindsay, 2015).
It is difficult to disentangle the fluency account and the illusion-of-evidence account on
the basis of available research. While many studies identified factors that moderate the truthiness
effect of nonprobative photos -- such as the temporal orientation and valence of the claim
(Newman, Azad, & Lindsay, 2016), one’s background knowledge (Abed, Fenn, & Pezdek,
2017), and judgement contexts (Newman et al., 2015) -- few have directly examined the
mechanism(s) underlying truthiness effects. Moreover, both accounts lead to the same
predictions under most conditions -- any related photo is likely to facilitate processing and may
be seen as tentative supporting evidence. However, these difficulties can be overcome by testing
the influence of nonprobative photos on a type of claim that has so far not received attention:
comparative claims about an unobservable difference between two targets.

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 4
Comparative and non-comparative claims
To date, studies have relied on noncomparative claims, such as Turtles are deaf. Such
claims are more likely to be accepted when accompanied by a picture of a turtle, even though the
picture provides no probative information about the turtle’s hearing. This observation is
compatible with a fluency account as well as an illusion-of-evidence account. But these accounts
lead to diverging predictions for comparative claims, such as Turtles have better hearing than
sealions or Sealions have better hearing than turtles. From an illusion-of-evidence perspective, a
photo of a turtle as well as a photo of a sealion might be interpreted as evidence for either claim.
Hence, this account predicts that the influence of a nonprobative photo is independent of the
direction of comparison (here, whether turtles are compared to sealions vs. sealions to turtles)
and the specific target shown in the photo (here, a turtle or a sealion).
In contrast, previous research into the processing of comparative claims (Tversky, 1977;
Tversky & Gati, 1978) suggests that the direction of comparison should influence whether the
target shown in the photo facilitates or impairs fluent processing, resulting in differential
truthiness effects. Evaluating the truth of any comparative claim requires that the subject of the
claim is compared to its referent; in the claim Turtles have better hearing than sealions turtles
constitute the subject and sealions the referent. As Tversky (1977; Tversky & Gati, 1978)
demonstrated, the comparison process begins with an assessment of features of the subject,
which are then checked against features of the referent. Hence, switching subject and referent
results in qualitatively different judgments, such as the memorable observation that North Korea
is judged as being more similar to China than China is to North Korea (Tversky & Gati, 1978).
Similarly, female teachers are judged as more empathetic than male teachers when they serve as
the subject of the comparison, but as less empathetic when they serve as the referent of the
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 5
comparison, reflecting that the direction of comparison influences which features come to
represent the attribute “empathetic” (Wänke, Schwarz, & Noelle-Neumann, 1995). Such findings
highlight that comparative statements are about the subject, which is then compared to a referent.
From a fluency perspective, photos of the subject should facilitate a processing sequence that
begins with assessing attributes of the subject, whereas photos of the referent should impair it.
Hence, the fluency account predicts that, relative to a no-photo condition, a comparative claim is
more likely to be accepted as true when it is accompanied by a photo of the subject of
comparison (i.e., a truthiness effect), but less likely to be accepted as true when it is accompanied
by a photo of the referent of comparison (i.e., a falsiness effect).
To date, falsiness effects of nonprobative photos have only been observed when the photo
was clearly unrelated to the claim; for example, when a claim about macadamia nuts appeared
with a photo of a trash can (Newman et al., 2015). This may reflect that the unrelated photo was
not perceived as tentative supportive evidence or that it impaired processing of the claim. This
ambiguity does not apply in the case of comparative claims, which provide an opportunity to test
whether related photos increase rejection of a claim when they render processing disfluent.
Finally, numerous studies indicate that fluency effects are relative and driven by changes
in fluency rather than the absolute level of fluency (Wänke & Hansen, 2015), which is consistent
with the general observation that people are more sensitive to changes in experience than to
steady states (Berelson & Steiner, 1964). Hence, fluency effects are more reliably observed in
within-participant designs, where the processing experience differs from trial to trial, than in
between-participant designs, where the processing experience for a given participant remains
relatively constant across trials (e.g., Hansen, Dechêne, & Wänke, 2008; Westerman, 2008; for a
review, see Wänke & Hansen, 2015). This robust observation suggests that the size of the
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 6
expected effects of nonprobative related photos should depend on whether the subject or referent
of a claim are familiar and easy to visualize without a photo. This prediction cannot be derived
from an illusion of evidence account. For ease of exposition, we elaborate on all predictions in
more detail in the context of the respective experiments.
Present Research
We tested diverging predictions derived from the fluency and illusion-of-evidence
accounts in three experiments by presenting comparative and non-comparative statements using
targets that are either easy or difficult to process without photos. To identify suitable targets, we
drew on imageability and familiarity ratings in the Medical Research Council Psycholinguistic
Database - Version 2.00 (Wilson, 1988), following Newman et al. (2018). We selected
commodities that are either high or low in both imageability and familiarity (henceforth referred
to as high vs. low imageability conditions).
1
The selected commodities were then combined with
statements that predicted their change in market price in three months. In experiment 1, we
presented non-comparative statements (e.g., Milk [Bhang] will have increased in price in three
months) to ensure that the items would produce the standard truthiness effect observed in
previous studies and to test whether the observed influence of photos is moderated by the ease
with which the claim can be processed in the absence of a photo. After establishing that standard
truthiness effects can be obtained with these materials, we used them to generate comparative
claims (e.g., Milk [Bhang] will have increased more in price than Strawberry [Aster] in three
months). In all experiments, we are interested in whether adding a nonprobative photo to a claim
can increase or decrease its acceptance compared to a condition where the claim is presented
without a photo. Note that the criterion for a truthiness or falsiness effect is whether a

1
The correlation between the ratings of familiarity and imageability for our stimulus set was r
= 0.98, p < 0.001.
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 7
nonprobative photo increases or decreases acceptance of the claim as true relative to a no-photo
condition, not whether the acceptance is above or below chance. For all studies, we report how
we determined the sample size, all data exclusions, all manipulations, and all measures.
Experiment 1: Noncomparative claims
Experiment 1 tests predictions of the illusion-of-evidence account and fluency account
using non-comparative claims. Both accounts predict (i) that a given statement is more likely to
be judged true when it is accompanied by a nonprobative photo of the subject than when it is not.
The fluency account assumes that a given statement is more likely to be judged true when it is
easy to process relative to other statements and attributes the impact of a nonprobative subject
photo to an increase in processing fluency. Accordingly, the truthiness effect should (ii) be larger
when the subject is unfamiliar and difficult to imagine than when it is familiar and easy to
imagine. In contrast, the illusion-of-evidence account assumes that all related nonprobative
photos are perceived as supportive evidence. It predicts, (iii) that the truthiness effect is
independent of how easily the subject can be imagined without a photo.
Method
Participants. Based on the average size (Cohen’s d = .24) of truthiness effects produced
by related but nonprobative photos in Newman et al.’s (2015) within-subject experiments,
G*Power (Faul, Erdfelder, Lang, & Buchner, 2007) indicated that 368 participants were needed
with an alpha = .05 and power = .90. We posted 400 time slots on Amazon Mechanical Turk,
limiting participation to participants with United States IP addresses and approval ratings of 95%
or higher for previous HITS. Participants were compensated with $0.60; N = 396 participants
completed the study and none were excluded from data analysis.
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 8
Design. Experiment 1 followed a 2 (photo: yes, no) x 2 (imageability: high, low) mixed
design, with the presence of a photo as a within-subject variable and imageability as a between-
subject variable. Participants were randomly assigned to the imageability conditions.
Material and procedure. Twenty-four commodities with low imageability and
familiarity ratings (e.g., saltpeter, yucca, tapis, etc.) and twenty-four commodities with high
imageability and familiarity ratings (e.g., cotton, strawberry, onion, etc.) were selected from the
MRC Psycholinguistic Database (Wilson, 1988). A non-comparative statement was created for
each commodity using the following structure: [Commodity] will have increased in price (three
months from today), adapted from Newman et al. (2018). For simplicity, participants were told in
the instructions that the statements would make predictions about commodity prices three
months from today, but the phrase “three months from today” was not repeated for each
statement in the experimental phase. Each statement was presented either without a photo (no-
photo condition) or with a photo of the commodity (photo condition). Figure 1 shows example
statements (Table A1 in Appendix A provides all commodities used and their imageability and
familiarity ratings).
All participants saw twenty-four statements of either high or low imageability, of which
half were paired with a photo. Whether a given statement was presented with a photo or no photo
was counterbalanced and statements were presented in a randomized order.

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 9
Figure 1. Conceptual examples of statements appearing in the low (left) and high (right)
imageability conditions. Photo of Leghorn: Creative Commons License attribution: Naetoru.
Photo of Orange: Creative Commons License attribution: Larissalara400.

The experiment was conducted on Qualtrics. All participants were informed prior to
participation that they must use a computer rather than a smartphone. Those who connected with
a smartphone were automatically identified by Qualtrics and redirected to a page informing them
that they could not participate on a mobile device. This was to ensure that all participants could
see the complete statements and the accompanying photo at the same time.
Participants were informed that they would see statements predicting the prices of
commodities in three months and that each statement would appear either with or without a
photo. They were to decide whether each statement is true or false based on their intuition. Next,
participants made a true/false judgement for each of the twenty-four statements by clicking a
“true” or “false” button below the statement. Each statement appeared on a separate page and
participants had to click “next” to move to the next page. At the end, participants were asked to
report their native language and whether they had looked up the commodities online during the
study.
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 10
Results and Discussion
Figure 2 shows the proportion of times participants responded that the claims were true.
A mixed 2 (photo: yes, no) x 2 (imageability: high, low) repeated measures ANOVA showed
that statements were more likely to be judged true when presented with a photo, F(1, 394) =
41.980, p < .001, 𝜂
"
#
= .096. This main effect replicates the standard truthiness effect first
reported by Newman et al. (2012) and is consistent with the illusion-of-evidence as well as the
fluency account.

Figure 2. Proportion of “true” responses to items that appeared with each photo type. Error bars
represent the 95% confidence intervals for the photo effects.

In addition, a main effect of imageability, F(1, 394) = 8.382, p = .004, 𝜂
"
#
= .021, as well
as an interaction between photo and imageability, F(1,394) = 24.969, p < .001, 𝜂
"
#
= .06,
emerged. Consistent with the fluency account, follow-up tests confirmed that claims presented
without a photo were more likely to be judged true when the subject was easy to imagine (M
= .502, SD = .184) than when the subject was difficult to imagine (M = .399, SD = .220), t (394)
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 11
= 5.033, p < .001, mean difference = .103, 95% CI [.062, .143]. Furthermore, adding a subject
photo to the claim increased acceptance in the low imageability condition, t(198) =7.579, p
< .001, mean difference = .128, 95% CI [.097, .159], but not in the high imageability condition,
t(196) = 1.137, p = .296, mean difference = .016, 95% CI [-.015, .048].
In sum, a given non-comparative claim is more likely to be judged true when its subject
is relatively easy to imagine. In the absence of photos, this advantages claims about high
imageability subjects over claims about low imageability subjects. Adding a subject photo the
claim increases acceptance when the subject is otherwise difficult to imagine, but does not
further increase acceptance when the subject is already easy to imagine. These observations are
consistent with a fluency account of truth judgment and cannot be derived from an illusion-of-
evidence account. Recall, however, that easily imagined commodities are also familiar
commodities, reflecting the high natural correlation between familiarity and imageability. It is
therefore conceivable that familiarity with the subject of the claim limited the impact of
nonprobative subject photos independent of imageability. Experiments 2 and 3 render this
alternative unlikely.
Experiment 2: Comparative claims
Experiment 2 extends experimentation on the influence of nonprobative photos to
comparative claims, which have not been addressed in previous research. The illusion-of-
evidence account predicts (i) that photos of the subject as well as photos of the referent will
increase acceptance of a claim as true, (ii) independent of the initial ease of processing (due to
imageability and familiarity) of the subject or referent. In contrast, the fluency account predicts
that photos of the subject and the referent will produce opposite effects on truth judgement and
that their impact will be moderated by the ease of processing the subject. Because the processing
of comparative claims starts with the subject (Tversky, 1977), a photo of the subject should
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 12
facilitate processing of the claim, especially if the subject is otherwise difficult to imagine. This
predicts (iii) that a given comparative claim is more likely to be judged true when it is presented
with a photo of the subject than without a photo. This truthiness effect should (iv) be more
pronounced under conditions of low rather than high imageability. In contrast, a photo of the
referent is likely to impair the processing of a comparative claim because the photo interferes
with attending to the subject at the initial stage of processing. Hence, a given comparative claim
should (v) be less likely to be judged true when it is presented with a photo of the referent than
without a photo, resulting in a falsiness effect. This adverse impact of a referent photo should be
more apparent when the subject is easy to imagine than when the subject is difficult to imagine
to begin with. Hence, the predicted falsiness effect of referent photos should (vi) be more
pronounced under conditions of high rather than low imageability.
Note that these predictions also bear on a number of other theoretical issues. First, with
regard to comparative judgment, observing the predicted differential effects of subject photos
and referent photos under conditions of high and low imageability would further support the
notion that comparisons begin with an assessment of the subject, as proposed by Tversky (1977)
for judgments of similarity. Second, with regard to the informational value of fluency
experiences, the same pattern of results would provide novel data that highlight that fluency
effects are relative -- experiencing low fluency due to a referent photo is less informative when
low fluency is already expected due to the low imageability and familiarity of the subject than
when it contrasts with the otherwise easy processing of a highly imageable and familiar subject.
Finally, the fluency account predicts that exposure to related, but nonprobative photos can
produce a falsiness effect when they pertain to the referent. To date, falsiness effects have only
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 13
been observed for photos that were completely unrelated to any aspect of the claim with which
they were paired (Newman et al., 2015).
Method
Participants. Since there was no previous study that reported falsiness effects of
nonprobative photos related to the claim, we used the average size (Cohen’s d = .19) of falsiness
effects produced by unrelated photos in Newman et al.’s (2015) within-subject experiments to
estimate the sample size required to capture a falsiness effect, if any, of the referent photos.
G*Power indicated that 586 participants were needed with alpha = .05 and power = .90. We
posted 600 timeslots on Amazon Mechanical Turk, using the same eligibility criteria and
compensation as in experiment 1. A total of 603 participants completed the experiment and no
participant was excluded from analysis.
Design. Experiment 2 used a 3 (photo: subject, referent, no) x 2 (imageability: high, low)
mixed design, manipulating the type of photo within participants and imageability between
participants. Participants were randomly assigned to either the high or low imageability
condition.
Material and procedure. Comparative statements were created from the same forty-
eight commodities used in experiment 1. By pairing commodities that had the closest
imageability ratings, twelve comparative statements were created for each imageability
condition. The statements had the following structure: [Commodity A] will have increased more
in price than [Commodity B] three months from today. Each statement appeared with a photo of
commodity A (subject photo condition), a photo of commodity B (referent photo condition), or
no photo (no-photo condition). Each photo had a label that identified the commodity (see Figure
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 14
3). The phrase “three months from today” was not repeated in the experimental phase after the
initial instructions.
Imageability was manipulated as a between-subject variable and participants saw either
twelve statements with familiar and easy-to-visualize commodities (high imageability condition)
or twelve statements with unfamiliar and difficult-to-visualize commodities (low imageability
condition). To manipulate photo type as a within-subject variable and to counterbalance the
photo type paired with each statement and the direction of comparison, we created six versions
of each statement for both imageability conditions. For example, the statement Shrimp will have
increased more in price than Roses was paired with a photo of the subject in counterbalance 1,
with a photo of the referent in counterbalance 2, and with no photo in counterbalance 3.
Counterbalance 4-6 repeated this arrangement except that the direction of comparison was
reversed (i.e., the statement became Roses will have increased more in price than Shrimp). The
counterbalance was arranged such that 1) each participant in either imageability conditions only
saw one version of each statement, and 2) among the twelve statements they saw, four appeared
with a photo of the subject, four appeared with a photo of the referent, and four appeared with no
photo. The order of the twelve statements was randomized. Figure 3 gives an example of
different photo types paired with the same statement in both the high and low imageability
conditions.

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 15
Figure 3. Conceptual examples of different photo types paired with the same statement for both
the high and low imageability conditions. Photo of Betel: Creative Commons License
Attribution: Ananda Cilianuri. Photo of Leghorn: Creative Commons License Attribution:
Naetoru.

Participants were informed that they must use a computer to participate in the study.
Those connecting with a mobile device were redirected to a page informing them that they could
not participate. Eligible participants received the same instructions used in experiment 1, except
that the example had a comparative structure (i.e., Gold will have increased more in price than
Silver). Participants were randomly assigned to either the high or low imageability condition and
made a true/false judgement about each of the twelve statements on separate pages. At the end of
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 16
the study, they reported their native language and whether they had looked up the commodities
on the internet.
Results
We first calculated the proportion of times participants responded “true” to each claim
across experimental conditions. We then tested the predictions of the illusion-of-evidence
account and fluency account with main effect contrasts between different photo types (subject vs.
referent photo; subject vs. no photo; referent vs. no photo) and simple contrasts of photo type
within each imageability condition. To correct for multiple comparisons, a Bonferroni
adjustment was made to the p-values and 95% confidence intervals. The results of the omnibus
analysis of variance and all interactions are reported in Appendix B.
The left-hand panel of Figure 4 shows the results. Overall, a given comparative claim was
more likely to be judged true when accompanied by a photo of the subject than when
accompanied by a photo of the referent, F(1, 601) = 38.371, p < .001, 𝜂
"
#
= .060, for the main
effect contrast. Photos of the subject increased acceptance of the claim (M = .577, SD = .266)
relative to the no-photo control condition (M = .531, SD = .266) -- a truthiness effect; F(1, 601) =
8.669, p = .003, 𝜂
"
#
= .014, for the main effect contrast. On the other hand, photos of the referent
decreased acceptance of the claim (M = .477, SD = .274) relative to the no-photo condition -- a
falsiness effect; F(1, 601) = 11.212, p = .001, 𝜂
"
#
= .018, for the main effect contrast. These
observations are consistent with a fluency account and incompatible with an illusion-of-evidence
account.

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 17

Figure 4. Proportion of “true” responses to items that appeared with each photo type. Error bars
represent the 95% confidence intervals for the photo effects.

The fluency account further predicts that the impact of photos is moderated by
imageability. The results support this prediction. As Figure 4 shows, the overall truthiness effect
observed for claims presented with a photo of the subject of comparison held in the low
imageability condition, t(300) = 4.246, p < .001, mean difference = .089, 95% CI [.036, .142],
but not in the high imageability condition, t(301) = .142, p > .999, mean difference = .003, 95%
CI [-.05, .056]. This is reflected in a significant interaction of imageability and the main effect
contrast between subject- vs. no-photo conditions, F(1, 601) = 7.468, p = .006, 𝜂
"
#
= .012.
Conversely, the overall falsiness effect observed for claims presented with a photo of the
referent of comparison held in the high imageability condition, t(301) = 3.393, p = .002, mean
difference = .079, 95% CI = [.024, .133], but not in the low imageability condition, t(300) =
1.303, p = .606, mean difference = .029, 95% CI [-.026, .084]. However, the interaction
between imageability and the referent- vs. no-photo main effect contrast was not significant, F(1,
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 18
601) = 2.375, p = .124, 𝜂
"
#
= .004. We address this issue in a meta-analysis after reporting
experiment 3.
Discussion
Experiment 2 used comparative claims to examine divergent predictions of the illusion-
of-evidence and the fluency account. The obtained results are consistent with a fluency account
and difficult to derive from an illusion-of-evidence account. First, photos of the subject of
comparison resulted in a truthiness effect, whereas photos of the referent of comparison resulted
in a falsiness effect. Both presumably reflect the impact of the photos on the ease of processing a
comparative claim. Because comparisons begin with an assessment of attributes of the subject,
which are then compared with attributes of the referent (Tversky, 1977; Tversky & Gati, 1978),
presenting a photo of the subject facilitates the flow of comparison, resulting in more fluent
processing. In contrast, a photo of the referent disrupts the flow of comparison, resulting in less
fluent processing. As observed for many fluency manipulations--from repetition (Hasher et al.,
1977) and color contrast (Reber & Schwarz, 1999) to rhyme (McGlone & Tofighbakhsh, 2000) -
- fluent processing increases, and disfluent processing decreases, the acceptance of claims as true
(for reviews, see Dechêne, Stahl, Hansen, & Wänke, 2010; Schwarz, 2018).
Second, consistent with the relative nature of experienced fluency, truthiness and
falsiness effects depended on the change in perceivers’ processing experience. Seeing a photo of
the subject increases ease of processing more when the subject is difficult to imagine without a
photo than when it is easy to imagine to begin with. Hence, photos of the subject enhanced
acceptance of the claim more in the low than in the high imageability condition. Conversely,
seeing a photo of the referent while thinking about the subject impairs experienced fluency more
when the subject is expected to be easy to process than when the subject is expected to be
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 19
difficult to process to begin with. Again consistent with the relative nature of the fluency
experience, exposure to a referent photo impaired acceptance of the claim more when the subject
was otherwise easy to imagine than when it was difficult to imagine. We return to the broader
theoretical implications of these findings in the general discussion.
Third, neither the differential effects of subject and referent photos nor their moderation
by imageability is predicted by an illusion-of-evidence account. According to this account,
photos can appear as though they are supportive evidence and the sheer fact that the
communicator presents it can render normatively irrelevant information conversationally
relevant in context (Schwarz, 1994, 1996). From this perspective, recipients may assume that the
photo presents evidence relevant to their task, which may result in a higher acceptance of claims
that are accompanied by a photo, provided that the photos are not clearly unrelated to the claim.
The present results are incompatible with this main effect prediction, as were the results of
experiment 1.
It is worth returning to an ambiguity of experiment 1. There, subject photos did not
increase acceptance of noncomparative claims when the subject was easy to imagine without a
photo. Experiment 2 replicated this finding for comparative claims: subject photos did not
increase acceptance of comparative claims when the subject was easy to imagine without a
photo. While these observations are consistent with a fluency account, they are also compatible
with a simple alternative: Given the high correlation between imageability and familiarity, they
may reflect that knowledge about the subject limits the influence of nonprobative photos. The
referent photo results argue against this possibility. If knowledge about familiar commodities
limited the influence of photos under high imageability conditions, it should not only reduce the
emergence of truthiness effects in response to subject photos but also the emergence of falsiness
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 20
effects in response to referent photos. Instead, we observe a divergent influence of high
imageability/familiarity in these cases and the pattern of divergence followed the predictions of
the fluency account.
Finally, a caveat needs attention. In experiment 2, all photos had a label that identified the
commodity shown (as illustrated in Figure 3), repeating the name of one of the two commodities
mentioned in the claim. Such a label was missing in the no-photo condition, where the names of
both commodities were only presented as part of the claim itself. It is conceivable that the verbal
label of the photo shown increased processing fluency beyond the influence of the photo itself.
Experiment 3 addresses this concern by replicating experiment 2, using the same procedures and
materials, with the verbal labels removed from the photos. This also allows us to test the
robustness of the reported results in a direct replication that merely differs in the presence of
verbal labels on the nonprobative photos shown.
Experiment 3
Method
Participants. We aimed to recruit 600 Mturk workers to keep the number of participants
consistent with experiment 2. The eligibility criteria and compensation followed experiment 2
and a total of 597 participants completed the study. No participant was excluded from data
analysis.
Material, design, and procedure. Experiment 3 is a direct replication of experiment 2,
except that the labels that identified the commodity shown were removed from the photographs.
A separate follow-up study indicated that participants could identify which commodity was
being shown to them in the low imageability condition even without the labels, Maccuracy = 0.607,
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 21
SD = 0.154, t (80) = 6.261, p < 0.001. For a detailed discussion, see Appendix C in
supplementary materials.
Results and Discussion
Data analysis followed the same procedure as experiment 2. As shown in the right-hand
panel of Figure 4, experiment 3 replicated the results of experiment 2, although with slightly
smaller effect sizes.
Claims accompanied by a photo of the subject were again more likely to be judged true
than claims accompanied by a photo of the referent; F(1, 595) = 18.186, p < .001, 𝜂
"
#
= .030, for
the main effect contrast. Compared to the no-photo condition (M = .517, SD = .256), photos of
the subject produced truthiness effect (M = .558, SD = .258), F(1, 595) = 7.825, p = .005, 𝜂
"
#

= .013, whereas photos of the referent produced a falsiness effect (M = .495, SD = .256),
although the latter was not significant, F(1, 595) = 1.981, p = .16, 𝜂
"
#
= .003.
Further replicating experiment 2, simple contrasts between the subject- and no-photo
conditions within each level of imageability revealed that photos of the subject increased
acceptance of the claim in the low imageability condition, t(290) = 3.544, p = .004, mean
difference = .070, 95% CI [.018, .121], but not in the high imageability condition, t(305) = .648,
p > .999, mean difference = .015, 95% CI [-.036, .065]. This is reflected in an interaction
between imageability and the subject- vs. no-photo contrast, F(1,595) = 3.317, p = .069, 𝜂
"
#

= .006. Also replicating experiment 2, simple contrasts between the referent- vs. no-photo
conditions showed that photos of the referent decreased acceptance of the claim in the high
imageability condition, t(305) = 2.458, p = .038, mean difference = .053, 95% CI [.002, .104],
but not in the low imageability condition, t(290) = .483, p > .999, mean difference = .01, 95% CI
[-.042, .063]. This is reflected in an interaction between imageability and the referent- vs. no-
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 22
photo, F(1,595) = 4.350, p = .037, 𝜂
"
#
= .007. Detailed descriptive statistics are reported in Table
1 and a full omnibus ANOVA is reported in Appendix B.

Table 1
Mean Proportions of "True" Responses and Unstandardized Effect Sizes for Each Target Condition
Exp. Imageability Photo Type Mean (SD) Effect Size (ES) 95% CI for ES
Exp. 1 Low (N=199) Photo .527 (.188) .128 *** [ .097, .159]
No Photo .399 (.220)
High (N=197) Photo .518 (.189) .017 [-.015, .048]
No Photo .502 (.184)

Exp. 2 Low (N=301) Subject .615 (.269) .089 *** [ .036, .142]
Referent .497 (.274) .029 [-.026, .084]
No .526 (.268)
High (N=302) Subject .540 (.259) .003 [-.050, .056]
Referent .458 (.274) .079 ** [ .024, .133]
No .536 (.266)

Exp. 3 Low (N=291) Subject .580 (.239) .070 ** [ .018, .121]
Referent .521 (.269) .010 [-.042, .063]
No .510 (.257)
High (N=306) Subject .538 (.274) .015 [-.036, .065]
Referent .470 (.241) .053 * [ .002, .104]
No .523 (.256)
Note.
The unstandardized effect sizes are raw mean differences between the target and no-photo conditions.
*p<.05. **p<.01. ***p<.001.

In sum, experiment 3 replicated the results of experiment 2 using the same procedures
and photos, while removing the previously used verbal identifiers of the commodities shown.
However, not all the differences observed in experiments 2 and 3 were significant and we
address this using analyses suggested by Mcshane and Böckenholt (2017).
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 23
Meta-Analysis of Experiments 2 and 3
To calculate more precise effect size estimates for the influence of subject and referent
photos, we conducted a single-paper meta-analysis based on the comparative judgments
collected in experiments 2 and 3, following recommendations by Mcshane and Böckenholt
(2017). Appendix D presents the estimated unstandardized effect sizes (i.e., the raw mean
differences between the proportion of “true” responses to claims paired with each photo type).
As shown in Figure 5a, all three main effect contrasts were significant. 1) Photos of the subject
produced a higher proportion of “true” responses than photos of the referent. When compared to
the no-photo condition, 2) photos of the subject increased acceptance of the claim, whereas 3)
photos of the referent decreased acceptance of the claim. This confirms the emergence of a
truthiness effect for subject photos and a falsiness effect for referent photos.
Simple contrasts within each imageability condition further revealed 4) that the truthiness
effect produced by subject photos was reliable when the commodities were difficult to imagine,
but not 5) when they were easy to imagine (Figure 5b). In contrast, the falsiness effect produced
by referent photos was 6) reliable when the commodities were easy to imagine, but 7) not when
they were difficult to imagine (Figure 5c). These effects are also reflected in significant
interactions between imageability and the main effect contrasts (i.e., subject vs. no photo,
referent vs. no photo), as shown in the right panels of Figures 5b and 5c.
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 24

Figure 5. Estimates of unstandardized effect sizes for experiment 2 and 3 and the single-paper
meta-analysis (Mcshane & Böckenholt, 2017). The size of the squares represents the average
sample size per condition; the horizontal thick and thin lines around the squares represent 50%
and 95% CI intervals, respectively. The top panel (a) shows the three main effect contrasts. The
middle panel (b) and bottom panel (c) show the effects of subject and referent photos,
respectively, within each imageability condition, followed by their interaction with imageability.

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 25
General Discussion
In three experiments, we examined two potential mechanisms underlying the influence of
nonprobative photos on judgments of truth. Going beyond earlier research, we varied whether
the claim was non-comparative (experiment 1) or comparative (experiments 2 and 3), whether
the subject of the claim was easy or difficult to process (experiments 1-3), and whether a
nonprobative photo depicted the subject or the referent of a comparative claim (experiment 2-3).
These variations allowed us to further illuminate and disentangle the processes underlying the
truthiness effect (Newman et al., 2012), and to identify conditions under which reliable falsiness
effects emerge.
That nonprobative photos can influence the perceived veracity of a claim has received
broad attention because it is counterintuitive and violates normative standards of truth
assessment -- after all, nonprobative photos provide no evidence bearing on the claim. An
analysis in terms of Grice’s (1975) logic of conversation calls this assumption into question. The
mere fact that normatively irrelevant information is presented as part of a task renders it
conversationally relevant, which entices participants to draw on it in forming a judgment (for
reviews, see Schwarz, 1994, 1996). Presenting a photo as part of a claim may similarly convey
that it can be treated as evidence for the claim. Moreover, most photos that people encounter in
daily life capture real events and photos have long been offered as evidence that an event has
truly occurred (Mnookin, 1998). In addition, photos are rich in perceptual and semantic details
and people are more likely to assume that a mental event (e.g., an episodic memory) reflects
reality when it features such details (Johnson, Foley, Suengas, & Raye, 1988). In combination,
these factors may create an illusion of supportive evidence when a claim is presented with a
related photo, even when the photo does not have clear probative value. Plausible as this account
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 26
is, we have not obtained support for it. First, the illusion-of-evidence account predicts a main
effect of presenting a related photo that is independent of the imageability of the subject.
However, the imageability of the subject moderated the size of truthiness effects for comparative
as well as noncomparative claims in all three experiments. The illusion-of-evidence account also
fails to explain a falsiness effect when nonprobative photos of the referent are presented and its
moderation by imageability, in contrast to the results of experiments 2 and 3.
In contrast, the present results are fully consistent with a fluency account of truth
judgment. As observed in numerous studies, claims are more likely to be accepted as true when
incidental influences make them easier to process. Hence, repeating a claim (e.g., Hasher et al.,
1977), repeating its denial (e.g., Skurnik, Yoon, Park, & Schwarz, 2005), presenting it in an
easy-to-read color contrast (e.g., Reber & Schwarz, 1999), in high acoustic quality (e.g.,
Newman & Schwarz, 2018) or a rhyming form (McGlone & Tofighbakhsh, 2000) have all been
found to increase the claim’s acceptance (for a review, see Schwarz, 2018). These robust fluency
effects reflect that fluency can be brought to bear on key criteria that people use in judging truth
(Schwarz, 2015): Is it compatible with other things I know? Is it internally coherent and free of
contradictions? Does it come from a credible source? Do others believe it? Each of these criteria
can be evaluated analytically by drawing on relevant declarative inputs or intuitively by drawing
on the ease with which the claim can be processed. When statements are coherent (Johnson-
Laird, 2012) and compatible with the recipient’s knowledge (Winkielman, Huber, Kavanagh, &
Schwarz, 2012) they are more fluently processed than when they are incoherent or at odds with
other beliefs the recipient holds. Fluently processed names feel more familiar and endow their
bearers with higher trustworthiness (Silva, Chrobot, Newman, Schwarz, & Topolinski, 2017).
Widely shared beliefs have been encountered more frequently and are more easily processed due
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 27
to repetition, which increases estimates of social consensus even when all repetitions come from
a single source (Weaver, Garcia, Schwarz, & Miller, 2007). Finally, the amount of supporting
evidence is overestimated when some can easily be brought to mind (Schwarz, Sanna, Skurnik,
& Yoon, 2007), consistent with the availability heuristic (Kahneman & Tversky, 1973). In short,
the metacognitive experience of fluency provides an affirmative answer to each of the major
truth criteria people use in assessing the veracity of claims (Schwarz, 2015, 2018). From this
perspective, nonprobative photos should increase the acceptance of a claim to the extent that they
facilitate processing of the claim, but decrease acceptance to the extent that they impair
processing of the claim. Empirically, this was the case, resulting in a unique pattern of findings
across all studies.
First, relative to a no-photo condition, nonprobative photos of the subject of a claim
increased acceptance of non-comparative (experiment 1) and comparative claims (experiment 2
and 3), and more so when the subject of the claim was otherwise unfamiliar and difficult to
imagine. The moderation by familiarity and imageability parallels earlier observations using non-
comparative judgements that nonprobative photos exert less influence when the subject of the
claim is familiar (Newman et al., 2012; Abed et al., 2017).
Second, and more importantly, nonprobative photos of the referent decreased the
acceptance of comparative claims as true (experiments 2 and 3), and more so when the subject
was easy to imagine without a photo. Empirically, this provides first evidence that related photos
can produce falsiness effects. In research with non-comparative claims, falsiness effects were
only observed for unrelated photos (Newman et al., 2015, experiments 5 and 6), that is, photos
that had no substantive relationship to the claim (e.g., a photo of a pig shown with a claim about
shoe laces). Note that falsiness effects of unrelated photos are compatible with a fluency as well
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 28
as an illusion-of-evidence account. From an illusion-of-evidence perspective, a clearly unrelated
photo surely cannot be evidence for the claim and calls the communicator’s cooperativeness into
question, triggering more skeptical analysis of the claim. In contrast, photos of the referent of a
comparative claim are related to the substance of the claim and hence could serve as evidence.
From a fluency perspective, an unrelated photo impairs processing of the claim by introducing a
distractor. Similarly, a photo of the referent impairs processing of comparative claims because
comparative processing starts with the subject of the claim, not the referent. In both cases, the
emergence of falsiness effects can be plausibly traced to low processing fluency.
Third, truthiness as well as falsiness effects were moderated by imageability and
familiarity, consistent with the relative nature of the informational value of fluency experiences
(Newman et al., 2015; Wanke & Hansen, 2015). In the absence of any photos, non-comparative
claims with familiar and easy-to-imagine subjects were more likely to be judged true than claims
with unfamiliar and difficult-to-imagine subjects (Experiment 1). Presenting claims about
difficult-to-imagine subjects with a nonprobative photo brought acceptance of the claim to the
level observed for claims with easy-to-imagine subjects (Figure 2). However, nonprobative
photos of the subject did not affect truth ratings when the subject was easy to imagine even
without a photo, presumably because the photos did not enhance fluency relative to the already
high baseline. The findings for comparative claims parallel this pattern (Experiment 2).
Presenting a comparative claim with a nonprobative photo of the subject increased acceptance of
the claim relative to a no-photo condition when the subject was difficult to imagine; however, it
did not influence acceptance relative to a no-photo condition when the subject was already easy
to imagine (Figure 4). Conversely, presenting a comparative claim with a photo of the referent
decreased acceptance of the claim relative to a no-photo condition when the subject would
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 29
otherwise have been easy to imagine; however, it did not influence judgment relative to a no-
photo condition when the subject was expected to be difficult to imagine to begin with (Figure
4). Throughout, photos influenced judgments of truth when the resulting fluency experience
deviated from what it would have been without a photo but not otherwise.
As another alternative account one might also conjecture that participants thought that the
commodity that went up in price is the one shown in a photo; for example, for the claim “Cotton
will have increased more in price than salt” this would imply that cotton went up in price when a
photo of the subject is presented, whereas salt went up in price when a photo of the referent is
presented. Although this would produce a truthiness effect for subject photos and a falsiness
effect for referent photos, this possibility does not account for the moderation of photo effects
through the imageability of the subject in the absence of a photo, which was observed for
comparative as well as noncomparative claims.
Limitations and Future Directions
Some limitations are worth noting. First, we used statements about commodities because
they allowed us to have an objective standard to manipulate imageability/familiarity based on the
ratings in the MRC Psycholinguistic Database (Wilson, 1988). Future research may test the
robustness of the observed effects in other domains, including comparative claims about persons
and products. Second, imageability/familiarity was manipulated as a between-subject variable to
avoid large discrepancies in the imageability of the items that participants were asked to
compare. Theoretically, within-subject variations in imageability would further add to item-to-
item changes in fluency, which may attenuate or enhance the observed effects depending on the
resulting change in processing fluency. Finally, as is common in this type of research,
participants made numerous truth judgments within a short time, often bearing on unfamiliar
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 30
targets. Both of these aspects may have reduced task engagement compared to naturalistic
conditions under which a recipient may only be exposed to one or two claims at a time. Future
research may fruitfully address the role of task engagement by varying the number of claims and
their personal relevance.
Despite these caveats, our findings add to the converging evidence that processing
fluency is likely the main mechanism underlying the influence of nonprobative photos on the
acceptance and rejection of claims. They also suggest that much can be learned by moving from
omnibus manipulations of fluency (such as repetition, print font, color contrast, auditory quality)
to manipulations that differentially affect different components of a statement. Just as a photo of
the subject can facilitate, and a photo of the referent impair, processing of a comparative claim,
differential fluency of different components of complex claims may introduce complexities that
challenge the current assumption that easy processing of the parts always fosters easy acceptance
of the whole.

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 31
References
Abed, E., Fenn, E., & Pezdek, K. (2017). Photographs elevate truth judgments about less well-
known people (but not yourself). Journal of Applied Research in Memory and Cognition,
6, 203-209.
Berelson, B., & Steiner, G. A. (1964). Human behavior: An inventory of scientific findings.
Oxford, UK: Harcourt, Brace & World.
Davies, C., & Katsos, N. (2009, July). Are interlocutors as sensitive to over-informativeness as
they are to under-informativeness. In Proceedings of the CogSci Workshop on the
Production of Referring Expressions (PRE-CogSci 2009), Amsterdam, The Netherlands.
Dechêne, A., Stahl, C., Hansen, J., & Wänke, M. (2010). The truth about the truth: A meta-
analytic review of the truth effect. Personality and Social Psychology Review, 14, 238-
257.
Eskritt, M., Whalen, J., & Lee, K. (2008). Preschoolers can recognize violations of the Gricean
maxims. British Journal of Developmental Psychology, 26(3), 435-443.
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, 175-191.
Fenn, E., Newman, E. J., Pezdek, K., & Garry, M. (2013). The effect of nonprobative
photographs on truthiness persists over time. Acta Psychologica, 144, 207-211.
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and
semantics: Vol. 3: Speech acts (pp. 41-58). New York: Academic Press.
Hansen, J., Dechêne, A., & Wänke, M. (2008). Discrepant fluency increases subjective truth.
Journal of Experimental Social Psychology, 44(3), 687-691.
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 32
Hasher, L., Goldstein, D., & Toppino, T. (1977). Frequency and the conference of referential
validity. Journal of Verbal Learning & Verbal Behavior, 16, 107–112.
Johnson, M. K., Foley, M. A., Suengas, A. G., & Raye, C. L. (1988). Phenomenal characteristics
of memories for perceived and imagined autobiographical events. Journal of
Experimental Psychology: General, 117(4), 371.
Johnson-Laird, P. N. (2012). Mental models and consistency. In B. Gawronski & F.
Strack (Eds.), Cognitive consistency: A fundamental principle in social cognition (pp.
225-243). New York: Guilford Press.
Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological Review,
80(4), 237.
McGlone, M. S., & Tofighbakhsh, J. (2000). Birds of a feather flock conjointly (?): Rhyme as
reason in aphorisms. Psychological Science, 11, 424-428.
McShane, B. B., & Böckenholt, U. (2017). Single-paper meta-analysis: Benefits for study
summary, theory testing, and replicability. Journal of Consumer Research, 43, 1048-
1063.
Mnookin, J. (1998). The image of truth: Photographic evidence and the power of analogy. Yale
Journal of Law & the Humanities, 10, 1-74.
Newman, E. J., Azad, T., Lindsay, D. S., & Garry, M. (2018). Evidence that photos promote
rosiness for claims about the future. Memory & Cognition, 1-11.
Newman, E. J., Garry, M., Bernstein, D. M., Kantner, J., & Lindsay, D. S. (2012). Nonprobative
photographs (or words) inflate truthiness. Psychonomic Bulletin & Review, 19(5), 969-
974.
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 33
Newman, E. J., Garry, M., Unkelbach, C., Bernstein, D. M., Lindsay, D. S., & Nash, R. A.
(2015). Truthiness and falsiness of trivia claims depend on judgmental contexts. Journal
of Experimental Psychology: Learning, Memory, and Cognition, 41, 1337-1348.
Newman, E. J., & Schwarz, N. (2018). Good sound, good research: How audio quality influences
perceptions of the research and researcher. Science Communication, 40(2), 246-257.
Newman, E.J., & Zhang, L. (in press). How nonprobative photos influence judgments of truth. In
Greifeneder, R., Jaffé, M., Newman, E., & Schwarz, N. (Eds.), The psychology of fake
news: Accepting, sharing, and correcting misinformation. London, UK: Psychology
Press.
Reber, R., & Schwarz, N. (1999). Effects of perceptual fluency on judgments of truth.
Consciousness and Cognition, 8, 338-342.
Schwarz, N. (1994). Judgment in a social context: Biases, shortcomings, and the logic of
conversation. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 26,
pp. 123-162). San Diego, CA: Academic Press.
Schwarz, N. (1996). Cognition and communication: Judgmental biases, research methods, and
the logic of conversation. Hillsdale, NJ: Erlbaum.
Schwarz, N (2015). Metacognition. In M. Mikulincer, P. R. Shaver, E. Borgida, & J. A. Bargh
(Eds.), APA Handbook of Personality and Social Psychology: Attitudes and Social
Cognition (pp. 203-229). Washington, DC: APA.
Schwarz, N. (2018). Of fluency, beauty, and truth: Inferences from metacognitive experiences. In
J. Proust & M. Fortier (Eds.), Metacognitive diversity: an interdisciplinary approach (pp.
25-46). New York: Oxford University Press.
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 34
Schwarz, N., Sanna, L. J., Skurnik, I., & Yoon, C. (2007). Metacognitive experiences and the
intricacies of setting people straight: Implications for debiasing and public information
campaigns. Advances in Experimental Social Psychology, 39, 127-161.
Schwarz, N., Strack, F., Hilton, D., & Naderer, G. (1991). Base rates, representativeness, and the
logic of conversation: The contextual relevance of “irrelevant” information. Social
Cognition, 9, 67-84.
Silva, R. R., Chrobot, N., Newman, E. J., Schwarz, N., & Topolinski, S. (2017). Make it short
and easy: Username complexity determines trustworthiness above and beyond objective
reputation. Frontiers in Psychology, 8.
Skurnik, I., Yoon, C., Park, D. C., & Schwarz, N. (2005). How warnings about false claims
become recommendations. Journal of Consumer Research, 31, 713–724.
Tversky, A. (1977). Features of similarity. Psychological Review, 84, 327–352.
Tversky, A., & Gati, I. (1978). Studies of similarity. In E. Rosch & B. Lloyd (Eds.), Cognition
and categorization (pp. 79 –98). Hillsdale, NJ: Erlbaum.
Wänke, M., & Hansen, J. (2015). Relative processing fluency. Current Directions in
Psychological Science, 24(3), 195-199.
Wänke, M., Schwarz, N., & Noelle-Neumann, E. (1995). Asking comparative questions: the
impact of the direction of comparison. Public Opinion Quarterly, 59(3), 347-372.
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 35
Weaver, K., Garcia, S. M., Schwarz, N., & Miller, D. T. (2007). Inferring the popularity of an
opinion from its familiarity: A repetitive voice can sound like a chorus. Journal of
Personality and Social Psychology, 92, 821–833.
Westerman, D. L. (2008). Relative fluency and illusions of recognition memory. Psychonomic
Bulletin & Review, 15(6), 1196-1200.
Wilson, M. (1988). MRC psycholinguistic database: Machine-usable dictionary, version 2.00.
Behavior Research Methods, Instruments, & Computers, 20, 6-10.
http://websites.psychology.uwa.edu.au/school/MRCDatabase/uwa_mrc.htm
Winkielman, P., Huber, D. E., Kavanagh, L. & Schwarz, N. (2012). Fluency of consistency:
When thoughts fit nicely and flow smoothly. In B. Gawronski & F. Strack (Eds.),
Cognitive consistency: A fundamental principle in social cognition (pp. 89-111). New
York: Guilford Press.

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 36
Supplementary Materials
Appendix A: Table of Commodities with Imageability Ratings

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 37
Appendix B: Omnibus ANOVA from Experiment 2 & 3
Omnibus ANOVA of experiment 2. In addition to planned contrasts testing our
theoretical predictions, we also computed a 3 (photo condition) x 2 (imageability) mixed
factorial ANOVA. It revealed a significant main effect of photo type, F(2, 601) = 19.650, p
< .001, 𝜂
"
#
= .032, with subject photos resulting in a truthiness effect and referent photos resulting
in a falsiness effect, as well as a significant main effect of imageability, F(1, 601) = 8.618, p
= .003, 𝜂
"
#
= .014, with less imageable claims resulting in higher truth ratings overall. These main
effects are qualified by a significant interaction between photo type and imageability, F(2, 601) =
3.623, p = .027, 𝜂
"
#
= .006, consistent with the planned contrasts reported earlier.
Omnibus ANOVA of experiment 3. In addition to planned contrasts testing our
theoretical predictions, we also computed a 3 (photo condition) x 2 (imageability) mixed
factorial ANOVA. Replicating experiment 2, significant main effects of photo, F(2, 595) =
9.221, p < .001, 𝜂
"
#
= .015, and imageability, F(1, 595) = 5.246, p = .022; 𝜂
"
#
= .009, as well as a
marginally significant interaction emerged, F(2, 595) = 2.612, p = .074, 𝜂
"
#
= .004.

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 38
Appendix C: Accuracy of Identification of Commodities without Verbal Labels
We ran a separate study to determine whether participants can identify the commodities
shown in the photos when no labels are provided.
Method.
Participant. Eighty-one participants from Amazon Mechanical Turk were recruited, with
criteria and compensation identical to experiment 1, 2, and 3. No participant was excluded from
data analysis.
Material and procedure. The comparative claims used in the low imageability conditions
of experiments 2 and 3 were presented without labels. However, instead of judging whether the
statement was true, participants judged which of the two commodities mentioned in the
statement was shown in the photo. For example, the statement Marl will have increased more in
price than Tapis appeared with either a photo of Marl or Tapis and participants were asked “Do
you think this is Marl or Tapis?” After reporting their identifications for all claims, participants
were reminded of their answers and asked to rate their confidence (on a 0 to 100 scale) for each
identification made. The order of the comparison (e.g., whether Marl served as the subject or the
referent in the claim) and whether the photo depicted the subject or the referent were
counterbalanced, resulting in four versions of each statement. Each participant only saw one
version. The order of the questions was randomized.
Result. A one-sample t-test revealed that participants were able to identify the
commodity depicted by the photo significantly above chance, Maccuracy = 0.607, SD = 0.154, t
(80) = 6.261, p < 0.001, although their overall confidence was low, Mconfidence = 43.838, SD =
18.492. This provides evidence that participants were able to tell whether the subject or the
referent was being shown to them even without the label.
TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 39
Interestingly, a one-sample t-test also revealed that when in doubt, participants were more
likely to say that the photo depicted the subject, Msubject = 0.541, SD = 0.130, t (80) = 2.852, p =
0.006. This suggests that participants were more likely to expect to see a photo of the subject
than the referent when a photo is paired with a comparative claim, consistent with previous
research that comparative processes are dominated by the subject (1977; Tversky & Gati, 1978).

TRUTHINESS AND FALSINESS IN COMPARATIVE JUDGEMENTS 40
Appendix D: Estimated Effect Sizes from Single-Paper Meta-Analysis
Table C1. Estimates of Unstandardized Effect Sizes from Single-Paper Meta-Analysis
Contrast type Contrast Estimate of ES 95% CI
Main Effect Contrasts Subject vs. Referent .159* [ .120, .198]
Subject vs. No .087* [ .048, .126]
Referent vs. No -.072* [-.113, -.031]

Simple Contrasts Subject vs. No (low img.) .077* [ .050, .104]
Subject vs. No (high img.) .010 [-.017, .037]

Referent vs. No (low img.) -.008 [-.037, .021]
Referent vs. No (high img.) -.064* [-.091, -.037]
I = 0, 95% CI [0, 67.534].
Note. ES stands for effect size; "*" indicates estimates that have a 95% CI that does not include zero.

Abstract (if available)

Abstract Claims are more likely to be judged true when presented with a related nonprobative photo (Newman, Garry, Bernstein, Kantner, & Lindsay, 2012). According to a fluency account, related photos facilitate processing and easy processing fosters acceptance of the claim. According to an illusion-of-evidence account, related photos may increase acceptance of the claim because they are treated as tentative supportive evidence. To disentangle these potential mechanisms participants read statements about the future price of easy- or difficult-to-visualize commodities. Consistent with fluency predictions, photos only increased the acceptance of claims when the target was difficult to visualize without a photo (Experiment 1). In Experiments 2 and 3 we used comparative claims. In forming comparative judgments, people first assess attributes of the linguistic subject of comparison and subsequently compare them to attributes of the referent (Tversky, 1977). Photos of the linguistic subject in a sentence should facilitate, whereas photos of the linguistic referent impair, the fluency of this sequence. Consistent with fluency predictions, photos of the subject in a claim increased acceptance of comparative claims relative to a no-photo condition (a truthiness effect), and more so when the subject was otherwise difficult to visualize. Photos of the referent decreased acceptance of comparative claims relative to a no-photo condition (a falsiness effect), and more so when the subject of comparison was otherwise easy to visualize. All results are consistent with a context sensitive fluency account: increases in fluency foster, and decreases in fluency impair, acceptance of a claim as true.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Intuitions of beauty and truth: what is easy on the mind is beautiful and true

PDF

Only half of what I’ll tell you is true: how experimental procedures lead to an underestimation of the truth effect

PDF

Are jokes funnier when they’re easier to process?

PDF

“What difficulty means for me”: predictors and consequences of difficulty mindsets

PDF

Difficulty-as-sanctifying: when difficulties build character, purify the self, and elevate the soul

PDF

Metacognitive experiences in judgments of truth and risk

PDF

A roadmap for changing student roadmaps: designing interventions that use future “me” to change academic outcomes

PDF

Can I make the time or is time running out? That depends in part on how I think about difficulty

PDF

Cultural influences on mental health stigma in Asian and European American college students

PDF

Effects of racial exclusion from media

PDF

Brainstem structural integrity in the progression of Alzheimer's disease

PDF

The effects of bilingual acceptance and commitment training (ACT) on exercise in bilingual international university students

PDF

Learning repetition rules and non-adjacent dependencies from human actions in nine-month-old infants

PDF

Prenatal predictors of parental sensitivity in first-time mothers and fathers

PDF

Towards a correlational law of language: three factors constraining judgement variation

PDF

Examining the effect of fertility on lifespan using data from the Swedish Twin Registry

PDF

Altitude effect on tree wood carbon isotopic composition in humid tropical forests

PDF

The effects of acceptance and commitment therapy-based exercises on eating behaviors in a laboratory setting

PDF

Comparative study of the POG and FNCLCC grading systems in non-rhabdomyosarcoma soft tissue sarcomas

PDF

Increase in surface temperature and deep layer nitrate in the California Current: a spatiotemporal analysis of four-dimensional hydrographic data

Asset Metadata

Creator Zhang, Lin (author)

Core Title When photos backfire: truthiness and falsiness effects in comparative judgements

School College of Letters, Arts and Sciences

Degree Master of Arts

Degree Program Psychology

Publication Date 12/18/2019

Defense Date 12/17/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag comparative judgements,falsiness,fluency,nonprobative photos,OAI-PMH Harvest,truthiness

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Schwarz, Norbert (committee chair), Miller, Lynn Carol (committee member), Oyserman, Daphna (committee member)

Creator Email linzhan@indiana.edu,zhan306@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-255792

Unique identifier UC11675399

Identifier etd-ZhangLin-8098.pdf (filename),usctheses-c89-255792 (legacy record id)

Legacy Identifier etd-ZhangLin-8098.pdf

Dmrecord 255792

Document Type Thesis

Rights Zhang, Lin

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA