OUTCOME RELEVANT INVOLVEMENT AS A MOTIVATION FOR SPONTANEOUS
VERACITY JUDGMENT
By
David Clare

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Communication – Doctor of Philosophy
2014

ABSTRACT
OUTCOME RELEVANT INVOLVEMENT AS A MOTIVATION FOR SPONTANEOUS
VERACITY JUDGMENT
By
David Clare
Recently in deception detection research it has been shown that veracity judgments
infrequently occur without artificial prompting in paradigms representative of the area at large
(Clare, 2013). It is apparent through casual observation however that people do judge veracity in
natural contexts and the question is raised as to what does instigate those judgments. One
possible cause of veracity judgment is outcome relevant involvement. An experiment was
conducted testing this hypothesis, replicating Clare (2013), and testing the efficacy of explicit
instructions to consider veracity on the occurrence of veracity judgment. Veracity judgment was
measured in two ways. Results do not show support for the outcome-relevance hypothesis but
do replicate Clare (2013) and give a quantitative assessment of how strongly instructions to
consider veracity do in fact provoke veracity judgment.

ACKNOWLEDGEMENTS

If a scholar is lucky, completing their dissertation is bittersweet. It is obviously
wonderfully sweet to shrug the albatross project off your neck and enjoy sunlight and laughter
again. On the other hand, completing a dissertation is bitter because you are no longer formally
the student of wonderful mentors. I have been extraordinarily lucky and the bitter note is
palpable.
I have three professors from the department that I am grateful to know.
Dr. Timothy Levine has been my advisor for the last five years. Obviously, Tim taught
me a lot of stuff about science. More importantly he has modelled for me what it is to live the
life of the mind. Tim exemplifies what it means to be passionate about what you do. I hope
someday I can be as dedicated to something as Tim is to his work.
Dr. Franklin Boster has been my sensei for about five years. He too has taught me a lot
of neat things about science. If perhaps I did not always have the intellectual horsepower to get
everything from his technical lessons, I do know that I have learned a lot from him about being
an honorable man. I firmly believe that most heinous people are the result of not having
someone like Frank in their lives. For all we know the world is very lucky that he intervened
with me before I got too old.
Dr. Amanda Holmstrom has been my “boss lady” for about five years. She took me on to
her research team when I was in my first year and she deserves a lot of credit for housebreaking
me into the research team model. Frankly, Mandy taught me a lot about how to be an adult. I
will very much miss the Dream Team she founded.
The people to whom I owe a debt of gratitude extends.

iii

Dr. Joseph Cesario came on to my dissertation committee and officially I have had little
interaction with him. I have had the opportunity to learn from him informally however. I had
probably learned as much from him before the dissertation as I did during.
Dr. Timothy Pleskac sat on my preliminary paper committee and I had taken a course
from him in the past. He probably made the smartest “push” comments that any committee
member did across the four defenses of the preliminary paper and dissertation.
Dr. Hee Sun Park sat on my preliminary paper committee, been my professor, and
mentored me as an instructor. Hee Sun taught me a lot and always had patience with me.
Bri DeAngelis has been my fraternal twin for five years. I am so grateful that she has
been my friend.
Ashley Hanna was someone who I figured would be my friend and although my
hypotheses are usually wrong (see document below), I was right. I am very thankful that I was
right.
Rachel O’Connor can count putting up with me through this dissertation as her first
miracle for sainthood. At Thanksgiving, Rachel is at the top of my list when we go around the
table.
Julia Garvey, Alysha Roman, Chelsea Jolly, Domeda Duncan, Daniel Kelly, Wenjuan
Huo, Ellie Litteral, and Katelynn Dietzen all provided crucial help with this project’s data
collection. I thoroughly enjoyed getting to meet them. In all seriousness, getting to know
research assistants is one of my favorite parts of the research process.

iv

TABLE OF CONTENTS
LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
INTRODUCTION ...............................................................................................................1
METHODS ..........................................................................................................................7
Participants ...............................................................................................................7
Design ......................................................................................................................7
Outcome-Relevant Involvement Induction ..................................................8
Veracity Consideration Instruction Induction ..............................................9
Order of Veracity Consideration Measurement Induction...........................9
Constant Message ......................................................................................10
Instrumentation ......................................................................................................10
Thought-listing ...........................................................................................10
Consideration of veracity scale ..................................................................12
Message comprehension ............................................................................12
Perceived outcome relevance .....................................................................13
Procedure ...............................................................................................................14
RESULTS ..........................................................................................................................15
Measurement Model ..............................................................................................15
Replication of Clare (2013) Preliminary Paper .....................................................15
The Effects of the Outcome Relevance Induction .................................................16
Coded Thought-listing and Close-ended Measures of Veracity Consideration.....17
Message Processing and Comprehension ..............................................................20
DISCUSSION ....................................................................................................................21
Summary of Results ...............................................................................................21
Replicated Clare (2013) ............................................................................ 22
Two measures of veracity consideration....................................................22
The effect of instructions to consider veracity...........................................23
The effect of outcome relevant involvement .............................................23
Message processing and comprehension ...................................................23
Explanation of Results ...........................................................................................23
The hypothesis was wrong .........................................................................24
Lack of statistical power ............................................................................24
Measurement of veracity considerations ...................................................26
Perceived motive to deceive ......................................................................26
ORI, message processing, and message comprehension ...........................27
Implications of Findings ........................................................................................28
Measuring consideration of veracity ..........................................................28
Instructions work, but not amply ...............................................................29
Question of what instigates veracity judgment still not answered .............30
The salience of veracity judgment in interaction .......................................31
v

A Future Direction .................................................................................................33
Conclusion .............................................................................................................35
APPENDICES ...................................................................................................................37
APPENDIX A: OUTCOME RELEVANCE INDUCTIONS ................................38
APPENDIX B: TRANSCRIPT OF INTERVIEW CONSTANT MESSAGE ......39
APPENDIX C: CONSIDERATION OF VERACITY SCALE.............................43
APPENDIX D: OUTCOME-RELEVANT INVOLVEMENT SCALE ................44
REFERENCES ..................................................................................................................46

vi

LIST OF TABLES

Table 1: Means with standard deviations in parentheses for number of thoughts coded as
mentioning veracity in different instruction and measurement order conditions………………..16
Table 2: Cell sizes, means and standard deviations of closed-ended scale scores across
conditions………………………………………………………………………………………...18
Table 3: Cell sizes, means and standard deviations of thoughts coded as explicitly mentioning
veracity across all conditions…………………………………………………………………….19

vii

LIST OF FIGURES

Figure 1: Number of relevant thoughts per participant………………………………………..…11
Figure 2: Number of thoughts per participant coded as mentioning veracity…………………...12

viii

INTRODUCTION
Recently in deception detection research a new question has been highlighted. While
decades of studies have diligently sought to explain why people judge messages to be deceptive
or honest, no one had previously asked when people actually make those judgments of veracity.
The entire literature had explicitly asked receivers of messages to report veracity judgments,
presuming that people had made these judgments and that those judgments simply had to be
tapped by measurement instruments. Clare and Levine (2013) showed that there is indeed
variance in whether or not people do make any consideration of message veracity. This paper
will use the term “veracity judgment” to mean a conclusion that a receiver of a message comes to
about whether the sender was lying or telling the truth. A veracity judgment is not a perception
about a person’s honesty or trustworthiness, but specifically a judgment as to whether the
message they are providing is honest. A message is dishonest if the sender of a message is
intentionally providing a presentation that is meant to mislead a receiver. Receivers must
consider the veracity of a message before coming to a conclusive judgment.
Clare and Levine (2013) conducted an experiment wherein receivers of messages
completed thought-listing measures on paper while watching taped autobiographical interviews.
They completed these thought-listing measures either before or after completing traditional
prompting measures of veracity judgment. Those listed-thoughts were coded for evidence of
veracity judgments. Participants only mentioned veracity in their thoughts for 4.23% of
messages when they listed those thoughts before the prompting measure and for 23.93% of
messages when they completed the prompting measure first. Clearly there is variance in whether
people consider the veracity of a message. Furthermore, absent prompting, consideration of
veracity is relatively infrequent. A subsequent replication, utilizing an identical design but

1

different a procedure where participants interviewed a confederate face-to-face and completed
measures orally found similar results (Clare, 2013). In that study, participants’ thoughts were
coded as mentioning veracity for less than 1% of messages when not previously prompted by an
explicit measure of veracity judgment and 8.46% of the time when they completed the explicit
measure first. In the condition where the thought-list measure was completed before the
prompting measure, the confidence interval of the mean percentage of messages coded as
mentioning veracity included zero. Considering these results, there is substantial variance in
how often receivers of messages consider veracity and absent prompting they do so very
infrequently.
It is incumbent upon the deception detection literature to explain when the focal
phenomenon of the literature, veracity judgment, and prior veracity consideration does indeed
occur. It is possible that much of the research purporting to explain the process of veracity
judgment is based on observations of the phenomenon in situations that are importantly distinct
from the situations in which the phenomenon would actually occur, if for no other reason than
some situations are distinct from others because in those people do consider veracity and in other
situations they do not.
Veracity judgment may be instigated by a number of causes. As this is a first study in
this program of research it is important to be explicit about the expectation that veracity
judgment can be multiply caused. Plausibly, there are a number of conditions that may instigate
veracity judgment. Clare (2013) offered a number of possibilities, including, as a partial
recounting, the disliking of a source and aggression, ego defense, sociological role or occupation,
and simplification of cognitively complex message environments. Truth-default Theory (Levine,
2014) lists a projected motive for deception, behavioral displays associated is a dishonest

2

demeanor, lack of coherent message content, message inconsistency with factual knowledge, and
third parting warning of impending deceit as trigger events for veracity assessment. Some of
these information and motivating factors can co-occur and are not intended to be mutually
exclusive. Many of them may be sufficient but not necessary for veracity judgment to occur.
The purpose of this study is to make an initial attempt at identifying and testing one possible
cause of veracity judgments: outcome-relevant involvement.
While veracity judgments may be infrequent, there may however be conditions under
which veracity judgments are more frequent. One possible condition may be when a message
has important implications for a message receiver’s future outcomes. Receivers may be more
likely to make veracity judgments when they experience higher outcome-relevant involvement.
Involvement is largely defined as a motivational state induced by a stimulus’s activation of
attitudes related to one’s self concept.
Johnson and Eagly (1989) differentiate outcome relevance from other types of
involvement in regards to a message. Outcome relevance refers to a message’s implications for
the receiver’s ability to attain desired outcomes. This contrasts with value relevance which
refers to a message’s pertinence to attitudes that the receiver enduringly holds closely to in their
self-concept. Outcome relevance differs from impression relevance which refers to a message’s
implications for the receiver’s later self-presentation to others about the issue in question.
Persuasion researchers posit that outcome-relevant involvement can motivate receivers to
more carefully attend to the content of messages. Multiple explanations of persuasive processes
converge on the proposition that outcome-relevant involvement increases careful, deliberate,
attentive processing of messages. The dual process heuristic-systematic model says that
outcome relevance increases systematic processing of messages (Chaiken, 1980). The

3

elaboration likelihood model says that outcome-relevant involvement increases central
processing of messages (Petty & Cacioppo, 1979). Kruglanski and Thompson’s (1999)
unimodel eschews the distinction between processing modes but similarly argues that outcomerelevant involvement increases message processing.
When receivers are more careful in attending to the content of messages, they may be
more likely to consider the logical or empirical validity of claims made in the message. They
may pay more attention to possible contradictions. They may analyze more deeply the
plausibility of statements, either in terms of understatement or overstatement. Receivers who are
more carefully processing are more likely to pick up on aspects of a message that might cast
doubt on veracity, leading to judgments about veracity. Veracity judgment can be seen as a
subset of deliberate message evaluation processes. As receivers more carefully attend to a
message, part of what they will be attending to is the veracity of the content. All three of the
aforementioned models (ELM, HSM, and unimodel) assume that a basic motivation in attending
to messages is to hold correct beliefs. Veracity judgment is one of the ways to guard against
holding incorrect beliefs.
A second condition that may motivate receivers to judge veracity is when they have been
given a cue suggesting to them that they should be suspect of a speaker’s veracity. When cued to
be skeptical of a source’s veracity, receivers are more likely to be so. When such a cue is absent,
receivers should be substantially less skeptical. Receivers cued to be skeptical, will be,
regardless of their level of involvement with the message. In the deception detection literature,
explicit instruction to consider veracity is nearly universal. In the absence of a cue however,
receivers will be more likely to judge veracity as they are more involved and the message has
implications for future outcomes.

4

To test this idea an experiment was conducted in which receivers were exposed to a
video-taped job interview. In one condition, the job and who is chosen for it was intended to
have great implications for the receivers’ lives and should be highly involving. In the other
condition, the job and the choice of applicant had negligible, if any, impact on receivers’ lives.
This involvement factor was crossed with another factor regarding a cue to be skeptical of the
interviewee’s veracity. In one condition, receivers were told that the interviewee covets the
potential job and will try to present themselves in the best possible light. Making the most
favorable impression may be furthered by being deceptive and receivers should be vigilant in
trying to ascertain the interviewee’s honesty. In the other condition, receivers were not given
this instruction.
After these relevance and instructions inductions, the frequency of veracity judgment was
be measured. This was done in two ways. One of the ways, replicating Clare and Levine (2013),
was an open-ended thought-listing measure that will be coded afterwards. The other method was
a closed-ended self-report scale. This scale was newly developed for this study. It was hoped
that it would be more sensitive in detecting differences across conditions if they exist. On a
practical note, a closed-ended scale would be less labor intensive and may be useful for future
research. To assess the possible effects of these different measurement instruments, the order in
which receivers completed these measures was experimentally varied.
The inclusion of the measurement order factor also offered essentially a replication of
Clare’s (2013) preliminary paper. In the condition where receivers completed the closed-ended
measure first, the explicit nature of the questions could potentially prime receivers to list more
thoughts relevant to veracity in the thought-listing measure. In the condition where receivers
completed the thought-listing measure first, it was expected that they would produce fewer

5

thoughts that could be coded as indicating veracity judgment because they have not been primed.
This effect was expected to possibly not be as apparent however in the conditions where
receivers would list many veracity-relevant thoughts anyway, i.e., given instructions to consider
veracity and high outcome-relevant involvement, due to a ceiling effect and may only be
detectable in the conditions where participants were less likely overall to consider veracity.
It was expected that outcome-relevant involvement will have a minimal effect on the
frequency of veracity judgments when a cue suggesting skepticism was present; all participants
presented with that cue should more frequently make judgments than when the cue is absent.
When the cue is absent however, receivers should more frequently make veracity judgments in
the high outcome-relevance condition in comparison to the low outcome-relevance condition.

6

METHODS
Participants
One-hundred eighty-six undergraduates were recruited from the Communication
department participant pool. Recruitment was restricted to students who were not in their last
year of school so as to heighten their outcome-relevant involvement in the high outcomerelevance condition. Participants were predominantly female (58.60%) and were on average
19.52 (sd = 1.37) years old. Freshmen comprised 41.90% of the sample, sophomores 25.30%,
and juniors 32.80%. Whites comprised 72.00% of the sample, Blacks 16.70%, Hispanics 2.70%,
Asians and Pacific Islanders 6.50%, and those who indicated other races or ethnicities made up
2.20% of the sample.
Design
The study used a 2 (low vs. high outcome-relevant involvement) x 2 (no instruction vs.
instruction) x 2 order of veracity consideration measurement (open-ended, closed-ended vs.
closed-ended, open-ended) independent groups experiment. Participants were run in sessions of
between one and eight people. Each session was assigned to one of the eight experimental
conditions with the constraint that twice as many participants were assigned to the no instruction
conditions in comparison to the instructions present condition. It was anticipated that there
would be greater variance in responses in the no instruction conditions and more observations
were desired to obtain standard errors of means that would be comparable between conditions.
Participants reported veracity judgment on a closed-ended measure and thoughts coded as
pertaining to message veracity were the dependent measures. Message comprehension served as
an indicator of careful, attentive message processing. Perceived outcome relevance was
measured as an induction check and as a potential mediator.

7

Outcome-Relevant Involvement Induction. In the low outcome relevance condition
receivers were told that they were going to watch a video of an undergraduate who was
interviewing for a research assistant position in the department of communication. It is highly
unlikely that whatever tasks the person is assigned will affect other students in any way, shape,
or form. In the high outcome relevance condition receivers were told that they were going to
watch a video of an undergraduate interviewing for a position on an undergraduate advisory
board. The details of the information given to participants in each condition is available in
Appendix A.
In order to ensure a strong induction of ORI, pretests of the inductions were conducted in
the semester prior to data collection of the main study. The two a priori planned conditions were
compared to each other and two other job descriptions over the two studies.
A first pretest was conducted with the undergraduate research assistant (URA),
undergraduate advisory board member, and a fiscal responsibility advocate job description. The
fiscal responsibility advocate was described as a position where an undergraduate would have an
advisory role on a university level budgetary unit that had as part of its responsibilities managing
the university endowment and setting tuition rates. A repeated measures experiment was
conducted online where 212 undergraduates, recruited from the same participant pool to be used
in the main study, rated each job description on the ten-item ORI measure described in the
instrumentation section. The order in which job descriptions were presented and rated was
randomized. The advisory board position was rated highest on ORI (m = 4.57, sd = 1.07), the
fiscal responsibility advocate second highest (m = 4.48, sd = 1.16), and the URA position was
rated lowest (m = 2.70, sd = 1.14). The repeated measures ANOVA comparing the three groups
was significant, F (2, 211) = 210.29, p < .01, partial η² = .57.

8

A second pretest study compared four different job descriptions including the URA, the
fiscal responsibility advocate, and the advisory board conditions used in the first pretest. The
additional job description was for a Community High School Liaison who would represent the
university at various community outreach programs at high schools. A repeated measures
experiment was conducted online where 221 undergraduates completed the ten-item ORI
measure for each of the job descriptions. The order in which the descriptions were presented and
rated was randomized. The advisory board position was again the most highly rated on ORI (m
= 4.43, sd = 1.12) and the URA position was rated as the lowest on ORI (m = 2.83, sd = 1.07).
The fiscal advocate was the second highest rated (m = 4.37, sd = 1.09) and the community
liaison position was the third highest rated (m = 3.01, sd = 1.00). The repeated measures
ANOVA was obviously significant, F (3, 220) = 161.17, p < .01, partial η² = .42.
Veracity Consideration Instruction Induction. In the veracity consideration instruction
condition, participants were told: “As you can imagine, the applicant in this video will be doing
their best to present themselves as well as possible. They will want to make a good impression
and convince the interviewer that they should be selected. In trying to do their best to be
selected, the applicant may be somewhat deceptive. As you watch the video, do your best to
determine whether they are being honest or deceptive. Try to tell if they are lying.” In the no
instruction condition, participants were not given those instructions.
Order of Veracity Consideration Measurement Induction. In the closed-ended measure
of veracity consideration first condition, participants completed the closed-ended measure of
veracity consideration, described below, before all other measures after the video. In the openended measure condition first condition, participants completed the thought-listing measure and
then the closed-ended measure before all other measures after the video.

9

Constant Message. All receivers watched the same interview of a female undergraduate
of 9:53 in duration. The interviewee was instructed to try to as convincingly as possible present
herself as a good job candidate, a likely motivation in natural job interviews. The interview was
conducted by a graduate student other than the author. The interviewee was not given a chance
to rehearse and was not be made aware of the questions beforehand. The lack of rehearsal was
intended to produce sender demeanor that is not overly smooth or practiced. Overall, the
intention was to create a relatively representative sample of a natural job interview that might
include disfluencies, verbal contradictions, and some variance in the quality of answers both in
content and delivery. A transcript of the interview is provided in Appendix B.
Instrumentation
Thought-listing. After watching the video, participants were be presented with lined
sheets of paper on which to report any thoughts they had during the video. They were instructed
to bullet-point every separate thought they have. These thoughts were coded by three trained
undergraduate coders. They first eliminated any thoughts that were off topic or did not pertain to
the message. Thoughts were considered relevant if they talked about aspects of the interviewee’s
nonverbal presentation, the interviewee’s answers, the interview questions, the job description,
the interviewee’s qualifications or appropriateness for the job, or the participants’ attitudes
towards the interviewee. This was a relatively broad definition of relevance and the vast
majority of reported thoughts were considered relevant. The number of remaining thoughts
served as an indicator of depth of message processing, coders achieved reliability of α = .98 in
counting how many thoughts relevant thoughts were provided by participants. The average
participant was coded as providing 7.34 relevant thoughts (sd = 2.84). A histogram of the
frequencies with which participants offered relevant thoughts is provided in Figure 2.

10

Figure 1: Number of relevant thoughts per participant.

The relevant thoughts were then coded in the same fashion as in Clare and Levine (2013)
for explicit mention of consideration of veracity. If the participant directly talked in a thought
about the truthfulness or deceptiveness of the interviewee or the interviewee’s answers or
indicated being credulous or incredulous of the interviewee or her answers, explicit discussion of
veracity was considered to be present in that thought. The three coders achieved reliability of α
= .92 in counting how many thoughts explicitly mentioned veracity. The number of thoughts
coded as explicitly mentioning veracity served as an indicator of veracity consideration. The
average number of thoughts coded as mentioning veracity for each participant was .77 (sd = .87).
A histogram of the frequencies with which participants offered thoughts that could be coded as
explicitly mentioning veracity is provided in Figure 2.

11

Figure 2: Number of thoughts per participant coded as mentioning veracity.

Consideration of veracity scale. A nine-item self-report scale was used to measure
participants’ spontaneous consideration of veracity while watching the video. The nine items
were originally hypothesized to follow a Guttman simplex structure but due to a lack of
substantial differences between item means, the scale was assessed with the classical theory of
random measurement error. All of the items used seven-point Likert scales ranging from
“strongly disagree” to “strongly agree”. The items in the scale are available in Appendix C.
Three of the items in this scale were dropped following confirmatory factor analysis (CFA). The
remaining six items exhibited reliability of α = .87. The average score on this measure was 5.49
(sd = 1.01).
Message comprehension. Two measures were be used to indicate message
comprehension, a recall measure and a recognition measure.
12

The first message comprehension measure was a recall measure. Participants were given
lined paper and asked to write down as many of the questions that the interviewee was asked and
the answers they gave. Participants were instructed to write in complete sentences. The
interview transcript was unitized into sentences, including questions and answers. Recalled
sentences were coded for the presence of those sentences by two trained coders. The number of
recalled sentences was used as an indicator of message comprehension. Coded scores on this
measure were correlated at r = .97 between the two coders. The mean score on this measure was
11.34 (sd = 7.10).
The second measure was a recognition measure. This measure consisted of 15 true-false
questions. These questions asked about various details and aspects of the interview. This
measure was intended to be an indicator of message comprehension to be used in conjunction
with the recall measure. In assessing this measure after data collection, it appeared that the items
were too easy and the measure yielded essentially no meaningful variance. Scores on this
measured were highly skewed. The mean number of items correct was 13.72, the modal
summed scored was 15 (a perfect score), the median score was 14, the variance was only 1.81,
and the standard deviation was only 1.34. Because of the lack of information that this measure
could provide, it was dropped from further analysis.
Perceived outcome relevance. The seven items from Cho and Boster’s (2005) outcomerelevant involvement (ORI) measure were adapted to measure perceived outcome-relevant
involvement. Three new items were added to create a potentially more reliable ten item
measure. All of these items used seven-point Likert scales ranging from “strongly disagree” to
“strongly agree”. These items are included in Appendix D. Two of the items in this scale were

13

dropped after CFA. The remaining 8 items exhibited reliability of α = .94. The average score on
this measure was 3.27 (sd = 1.42).
Procedure
Experimental sessions were run in a classroom with a large screen video projector and
speakers. Participants were run in groups as large as eight. All measures were completed on
paper. After informed consent was obtained, participants were given one of the two descriptions
of the context of the video they would be watching, either the high outcome-relevance or low
outcome-relevance description. Participants in the veracity consideration instruction condition
then received those instructions to pay attention to the honesty of the interviewee. Participants
then watched the taped interview. After the video was complete, participants were given a
questionnaire packet with the thought-listing and veracity consideration scale. The order in
which the thought-listing measure and veracity consideration scale were included in the packet
instantiated the measurement order induction. After that packet was completed participants were
given a questionnaire with the message comprehension measures, perceived outcome-relevant
involvement measure, and finally demographic items. Participants were instructed to complete
all measures in the order in which they were presented in the packet and to not go back to parts
of the packet once they had turned them over.

14

RESULTS
Measurement Model
The nine item closed-ended consideration of veracity measure, the ten item ORI measure,
the number of relevant thoughts counted, the number of thoughts counted as explicitly
mentioning veracity, and the recall scores were analyzed using CFA using the LessR package
(Gerbing, 2014). Three of the items in the veracity consideration scale were dropped due to
internal consistency errors and two of the outcome relevance items were dropped due to internal
consistency errors. The recall, coded explicit mention of veracity, and relevant thoughts values
were used to check parallelism. No items were dropped from any scale due to problems with
parallelism. The average absolute value of residuals in the matrix without the diagonal was .03.
Replication of Clare (2013) Preliminary Paper
In Clare’s 2013 preliminary paper, two experiments were conducted wherein participants
completed an open-ended measure of veracity consideration after a message either before or after
a measure explicitly asking the participant to make a veracity judgment. Those open-ended
measures were coded, with an identical procedure to the current experiment, for the presence of
thoughts that explicitly mentioned veracity. In both of those experiments it was found that
participants were less likely to mention veracity if they completed the open-ended measure
before the explicit measure. In the first study of Clare (2013), with 72 participants, an effect size
of r = .58 (95% CI [.40, .72]) was observed. The second study in that paper, with 68
participants, observed an effect size of r = .40 (95% CI [.18, .58]). The current study offers an
opportunity to see if that finding replicates, albeit with a different type of explicit measure of
veracity consideration.

15

The most direct comparison to Clare (2013) would be in comparing the number of
thoughts coded as explicitly replicating veracity in only the conditions where no instructions
were given to judge veracity. Across outcome-relevance conditions in the no instruction
conditions, the effect of measurement order is statistically significant, F (1, 123) = 5.65, p = .02,
η² = .04, d = .44, r = .21 (95% CI [.03, .37]). This effect size is within the confidence interval of
the observed correlation observed in the second study of Clare (2013). Interestingly, this effect
replicates even in the conditions where participants were instructed to consider veracity, F (1,
61) = 9.26, p < .01, η² = .13, d = .79, r = .37 (95% CI [.13, .57]). This effect size is closer to the
larger effect size found in study 1 of Clare (2013). Across all outcome relevance and instruction
conditions, the effect replicates, F (1, 185) = 13.54, p < .01, η² = .07, d = .55, r = .26 (95% CI
[.12, .39]). The average r, weighted by sample size, for the two preliminary paper studies and
the current study is .36 (95% CI [.26, .45]). The means and standard deviations of number of
thoughts coded as mentioning veracity are included in Table 1 below.
Table 1: Means with standard deviations in parentheses for number of thoughts coded as
mentioning veracity in different instruction and measurement order conditions.
No Instruction Instruction Condition Across Conditions
Condition (n = (n = 62)
124)
Open-ended first (n = 91)
.42 (.60)
.79 (.70)
.54 (.65)
Explicit first (n = 95)
.72 (.76)
1.54 (1.21)
1.00 (1.01)
Across Conditions
.57 (.69)
1.17 (1.05)
.76 (.87)

The Effects of the Outcome Relevance Induction
To assess the efficacy of the outcome relevance induction, ANOVA was conducted with
ORI scores as the dependent measure and outcome relevance condition as an independent groups
factor. Instruction condition and measurement condition were also included to assess the
possibility of confounding interactions. The effect of the outcome relevance induction was

16

statistically significant, F (1, 185) = 64.09, p < .01, η² = .25, d = 1.30, r = .54 (95% CI [.43, .63]).
In the low outcome relevance condition the average outcome relevance score was 2.49 (sd =
1.08) on a seven point scale while in the high outcome relevance condition the average score was
4.03 (sd = 1.30). None other the other inductions had statistically significant main effects on
outcome relevant involvement scores nor did any interactions have statistically significant
effects.
Coded Thought-listing and Closed-ended Measures of Veracity Consideration
This study used two measures of veracity consideration, one closed-ended Likert-type
scale and another using coded thought listings. If they both are solely different methods of
measuring the same latent construct they should exhibit strong covariation and similarly relate to
other constructs. Alas, especially strong covariation between the two measures was not the case
as the two measures exhibited only r = .33 (95% CI [.20, .46]), p < .01.
An ANOVA was conducted with scores on the closed-ended measure of veracity
consideration as the dependent measure and outcome relevance condition, instruction condition,
and measurement order condition as independent groups factors. The primary hypothesis of this
study is that receivers in the high outcome relevance condition will have higher scores on the
veracity consideration measure. The null hypothesis could not be rejected for this relationship, F
(1, 184) = 3.66, p = .06, η² = .02, d = .29, r = .15 (95% CI [.01, .29]). Participants in the low
outcome relevance condition had average scores of 5.34 (sd = 1.05) while those in the high
outcome relevance condition scored on average 5.64 (sd = .96). Instructions to consider veracity
did increase scores on this measure, F (1, 184) = 21.11, p < .01, η² = .10, d = 1.30, r = .33 (95%
CI [.19, .45]). When instructed to consider veracity, participants’ average score was 5.96 (sd =
.65) and when they were not given instructions participants’ average score was 5.26 (sd = 1.08).

17

This result constitutes the first ever validity check applied to instructions to consider veracity
which are constant in the literature. The measurement order induction had only a trivial effect on
scores on the closed-ended measure of veracity consideration, F (1, 184) = 0.04, p = .83, η² = 0,
d = .08, r = .04 (95% CI [-.11, .18]). No statistically significant interactions qualified these main
effects. Descriptive statistics for closed-ended scale scores are provided in Table 2.
Table 2: Cell sizes, means and standard deviations of closed-ended scale scores across
conditions.
Low ORI Condition
Hi ORI Condition
CEF First TLF First Across
CEF First
TLF First
Across
Conditions
Conditions
Instruction n = 30
n = 31
n = 61
n = 30
n = 33
n = 63
Absent
m = 5.08
m = 5.15 m = 5.12
m = 5.59
m = 5.23
m = 5.40
sd = 1.05 sd = 1.22 sd = 1.13
sd = 1.06
sd = .96
sd = 1.02
Instruction n = 14
n = 15
n = 29
n = 16
n = 16
n = 32
Present
m = 5.72
m = 5.90 m = 5.82
m = 6.09
m = 6.09
m = 6.09
sd = .72
sd = .55
sd = .63
sd = .53
sd = .76
sd = .64
Across
n = 44
n = 46
n = 90
n = 46
n = 49
n = 95
Conditions m = 5.29
m = 5.40 m = 5.34
m = 5.77
m = 5.51
m = 5.64
sd = 1.00 sd = 1.10 sd = 1.05
sd = .94
sd = .98
sd = .96

An ANOVA was conducted with the number of thoughts coded as explicitly mentioning
veracity as the dependent measure and outcome relevance, instruction, and measurement order
conditions as independent groups factors. The outcome relevance induction had only a trivial
effect on the number of thoughts explicitly mentioning veracity, F (1, 185) = 0.28, p = .60, η² =
0, d = .11, r = .06 (95% CI [-.09, .20]). In the low outcome relevance condition, participants
produced on average .72 (sd = .91) thoughts that mentioned veracity and in the high condition
produced on average .81 (sd = .84) such thoughts. The instructions induction did have a
statistically significant effect on the number of thoughts coded as explicitly mentioning veracity,
F (1, 185) = 23.70, p < .01, η² = .10, d = .69, r = .32 (95% CI [.18, .44]). When instructed to
consider veracity participants produced 1.17 (sd = 1.05) thoughts that were coded as mentioning

18

veracity while when not instructed they only produced on average 0.57 (sd = .69) thoughts
mentioning veracity. This result is remarkably similar to the result for the closed-ended measure.
As described before, the measurement order induction did have an effect on the number of
thoughts coded as explicitly mentioning veracity, F (1, 185) = 19.09, p < .01, η² = .08, d = .55, r
= .26 (95% CI [.12, .39]).
The effects of measurement order, instruction, and outcome relevance conditions on
thoughts coded as explicitly mentioning veracity were qualified by interactions. The interaction
between outcome relevance condition and measurement order condition was significant, F (1,
185) = 4.47, p = .04, η² = .02, such that the effect of measurement order was greater in the low
outcome relevance condition due to a higher mean in the thought-listing first cell in the high
outcome relevance condition compared to the low outcome relevance condition. The interaction
between instruction and measurement order conditions was significant, F (1, 185) = 3.92, p <
.05, η² = .02, such that the effect of the measurement order induction was stronger when
participants were given instructions to consider veracity. Descriptive statistics for thoughts
coded as explicitly mentioning veracity are provided in Table 3.
Table 3: Cell sizes, means and standard deviations of thoughts coded as explicitly mentioning
veracity across all conditions.
Low ORI Condition
Hi ORI Condition
CEF First TLF First Across
CEF First
TLF First
Across
Conditions
Conditions
Instruction n = 30
n = 31
n = 61
n = 30
n = 33
n = 63
Absent
m = .67
m = .32
m = .49
m = .77
m = .53
m = .64
sd = .72
sd = .46
sd = .62
sd = .80
sd = .69
sd = .75
Instruction n = 15
n = 15
n = 30
n = 16
n = 16
n = 32
Present
m = 1.80
m = .56
m = 1.18
m = 1.31
m = 1.00
m = 1.16
sd = 1.31 sd = .68
sd = 1.20
sd = 1.09
sd = .68
sd = .91
Across
n = 45
n = 46
n = 91
n = 46
n = 49
n = 95
Conditions m = 1.04
m = .40
m = .72
m = .96
m = .68
m = .81
sd = 1.08 sd = .54
sd = .91
sd = .94
sd = .72
sd = .84

19

Both the closed-ended measure and the thought-listings coded for explicit mention of
veracity were not significantly correlated with ORI scores, r = .03 and .05, respectively.
Regressing either measure onto outcome relevant involvement scores and instruction and
measurement order conditions did not yield a statistically significant relationship between
outcome relevant involvement scores and the veracity consideration measures, nor substantively
change the magnitude of the relationship.
Message Processing and Comprehension
The number of relevant thoughts that participants provided, as an indicator of message
processing, was expected to be higher in conditions where outcome relevance was induced to be
high. Message processing was not higher in the high outcome relevance condition, F (1, 185) =
1.77, p = .19, η² = .01. Message processing was not correlated with ORI scores, r = 0, p = .95.
Interestingly, even while there would be dependence between the number of relevant thoughts
produced and the number of thoughts coded as explicitly mentioning veracity, the two scores
were not significantly correlated, r = .11, p = .12. The correlation between closed-ended veracity
consideration scores and the number of relevant thoughts was actually stronger while still not
significant, r = .12, p = .09.
Message comprehension, indicated by recall scores, was predicted to increase with
message processing. Message comprehension did not appear to be amply related to the extent to
which participants processed the message, r = .06, p = .39. Recall scores were not correlated
with either the thought-listing (r = .04) or closed-ended (r = .04) indicators of veracity
consideration.

20

DISCUSSION
This study sought to test whether veracity consideration is instigated by outcome relevant
involvement with a message. It was predicted that in the absence of an explicit cue to consider
veracity that outcome relevance could spontaneously motivate receivers to consider the honesty
of a message they are exposed to. To test this hypothesis an experiment was conducted where
receivers of a message were given information either heightening or lowering their perceptions
of how much their future outcomes and consequences were tied to the message. This induction
was crossed with another independent groups factor in which participants were either given a
cue, direct instructions to consider veracity, or they were not. The inclusion of this induction
constitutes the first validity check in the deception detection literature as to whether instructions
to detect deception in fact prompt participants to do just that. This is, to the author’s knowledge,
the first time in the literature that such an assessment has been conducted. Consideration of
veracity was measured by two separate methods, one thought-listing method replicating Clare
(2013) and a new closed-ended measure. In part to assess the methodological concern of order
effects, and more importantly to offer a replication of Clare (2013), the order in which those
measures were completed was crossed with the outcome relevance induction.
In line with past persuasion theory and research, it was expected that outcome relevant
involvement should increase message processing. Increased message processing should in turn
be related to how much receivers comprehend the message they were exposed to. Depth of
message processing was measured with a thought-listing task and message comprehension was
measured with recall and recognition scales. The recognition scale was not analyzed due to lack
of variance in the measure.
Summary of Results

21

Replicated Clare (2013). Clare (2013) reported two experiments wherein the
consideration of veracity, as indicated by coded thought-listings, was infrequent when
participants had not previously completed an explicit measure of veracity judgment. The
measurement order induction, varying whether participants completed a thought-listing or
closed-ended measure of veracity consideration first, offered a replication of those findings.
Across outcome relevance and instruction conditions, it was observed that thought-listings
contained fewer thoughts that could be coded as explicitly mentioning veracity when the
thought-listing was completed before the explicit closed-ended scale. This experiment replicated
Clare’s (2013) finding. This pattern of findings did not extend to the closed-ended measure of
veracity judgment, suggesting that the evidence of increased veracity consideration observed in
the thought-listing was not simply a result of sensitivity to any previous measure.
A word of caution is responsible to those wishing to directly compare mean differences
between conditions between the two preliminary studies and the current study. The two Clare
(2013) studies measured veracity judgments per message for 16 different messages in an
interview. Each of these messages consisted of one question by an interviewer and one answer
by an interviewee. The current study measured veracity judgments for an entire interview. Clare
(2013) showed that veracity judgment was infrequent at the message unit of analysis while the
current study shows that veracity judgment is infrequent at the interaction unit of analysis. On
the one hand this does show the robustness of the finding from the standpoint of inferential
statistics. Unfortunately direct comparison of descriptive statistics is not possible without rather
byzantine and overly complicated conversions of the metrics of the different studies.
Two measures of veracity consideration. It was observed that the two different measures
of veracity consideration exhibited only moderate covariation, suggesting substantive differences

22

in what latent construct they may be measuring. Further analyses also showed differences in the
patterns of results obtained with the two measures, although findings were never significantly in
opposite directions.
The effect of instructions to consider veracity. With both measures of veracity
consideration, it was observed that receivers do indeed consider veracity more when they are
instructed to do so in comparison to when they are not. These effects were remarkably similar
across the two measures. With neither measure however was this effect incredibly large (r’s =
.33 and .32).
The effect of outcome relevant involvement. As evidenced by ORI scores, outcome
relevant involvement was amply induced by the two different conditions. Neither ORI scores
nor outcome relevance conditions had statistically significant relationships with veracity
consideration scores, although with the closed-ended measure the relationship between outcome
relevance induction and veracity consideration teetered at the edge of statistical significance.
Outcome relevance, either in the form of measured ORI scores or manipulated conditions, did
not have any relationship with message processing as indicated by the number of relevant
thoughts produced in thought-listing.
Message processing and comprehension. Message comprehension did not appear to be
related to depth of message processing. Both message processing and comprehension had trivial
relationships with measures of veracity consideration.
Explanation of Results
This experiment did not find support for the hypothesis that veracity consideration can be
instigated by outcome relevant involvement with a message. There is one obvious explanation
for this finding: the hypothesis is wrong. There are also a few methodological considerations

23

that may explain why the hypothesis did not find support. Chiefly among explanations in that
vein is that there was a lack of statistical power, are that the measures of veracity consideration
were poor and that the experiment’s job interview paradigm evoked a perceived motive for the
interviewee to deceive, thus washing out the possible effects of outcome relevant involvement.
The hypothesis was wrong. Reassessing the hypothesis that outcome-relevant
involvement instigates veracity consideration, there may be weaknesses with the hypothesis and
the experiment offered to test it. Outcome-relevant involvement was proposed to be induced by
information before a message that would increase outcome-relevant involvement through the
course of message exposure that would in turn affect how receivers considered the message, as
evidenced by post-test measures. A more elegant, complicated experiment may have sought to
measure veracity consideration at multiple points across exposure to a message and
experimentally vary at different points in a message whether the specific topic at hand had
outcome relevance for the receiver. The blanket initial information paradigm may have
convinced participants that the interview mattered generally to them, but there may not have
been specific aspects of the message that held concrete consequential outcome to them. To an
extent, the original hypothesis was under-specified, treating outcome-relevance as a relatively
stable motivational state and not as a more specific cognitive consideration that would be taken
into account while engaging with various parts of a message.
Lack of statistical power. It is possible that the lack of significant findings for the
outcome relevance conditions on veracity consideration was a Type II error. For the closedended measure, the p-value of the effect of the relevance conditions was .06 but with rounding
the confidence interval of the effect size r did not include 0. The temptation with that finding
may be to strategically drop subjects that for some reason do not satisfy the researcher or fiddle

24

with the measurement model in order to p-hack to significance. A one-tailed significance test
could be used. A more ethical approach however would be to admit that the effect did not reach
statistical significance by conventional standards, argue that null hypothesis significance testing
is flawed, and then promise that future meta-analyses of similarly small findings would
eventually yield support for the hypothesis in the form of a non-trivial effect size estimate. A
Bayesian might argue that researchers have an updated view of the hypothesis that holds a zero
estimate of the effect size to be slightly less likely. At the very least, it is a counterfactual to
consider whether statistical significance would have been achieved with a few more subjects. In
order to reach statistical significance, future studies should use a stronger induction of outcome
relevance, more reliable measures, and larger sample sizes.
One argument against the findings being Type II errors is that only one measure flirted
with statistical significance. Results for the thought-listing measure did not offer the agony of
being near statistical significance. Also, this experiment was fortunate to be able to collect more
(n = 186) than was initially planned (n = 180). Given the expected effect size and the sample
required to have reasonable power, this study collected what was required. In the end,
considering both measures, it appears that if there is an effect for outcome-relevant involvement
on veracity consideration that it is an exceptionally weak effect and could plausibly be negative.
The desire for a stronger induction of ORI may be supported by the actual mean ORI
scores observed in the different conditions. In the high involvement condition, the mean ORI
score was barely above the mid-point of the scale. This difference between the conditions was
statistically significant, but it could be argued that the high involvement condition was truly not
that involving. While this first presents itself as a restriction in range, there may be greater

25

theoretical concerns such that the level of involvement did not cross some crucial threshold
required to instigate veracity judgment.
Measurement of veracity consideration. The two methods used to measure consideration
of veracity, a closed-ended scale and coded thought listings may do a poor job at measuring the
construct of veracity consideration with enough reliability or sensitivity to detect meaningful
variation in the construct with this sample size. The two measures were only correlated at r =
.33. This suggests that either one of the measures is substantially more valid than the other or
that both measures lack a good deal of validity. One pressing concern is whether perhaps the
measures are not measuring veracity consideration broadly, but are instead only measuring
judgment that the message is deceptive. The relatively high means across all conditions on the
closed-ended scale suggest that this might be the case, the interviewee may have seemed
relatively dishonest. This does not appear to be the situation with the thought-listing measure
however. The thought-listing measure on the other hand is difficult to assess for reliability in as
straightforward a fashion as the closed-ended measure. The thought listing measure also has less
opportunity for variance due to the substantial positive skew it produces. It may be that the
thought-listing measure is more validly measuring veracity consideration but is relatively crudely
imprecise. Unfortunately going forward in trying to create an alternative measure of veracity
consideration it may be wise to simply start from scratch yet again.
Perceived motive to deceive. The seemingly most plausible explanation for the lack of
support found for the ORI hypothesis is that another effect washed out whatever influence ORI
could have. It is possible that there was a lack of meaningful variation in consideration of
veracity scores because indeed participants were very actively considering veracity in all
conditions because they perceived the interviewee to have a motive to deceive. While this seems

26

more likely in the high outcome relevance condition because the contested position was more
prestigious, it may be that very little perceived motive is required to instigate consideration of
veracity. The thought-listing measure does seem to argue against this possibility as the majority
of participants had relatively low values on this measure. Then again, if that measure is not
valid, then the low scores do not argue against this possibility. Future studies may try to remedy
this problem by using a paradigm where the sender would have no incentive to lie but the
receiver may have more or less consequences contingent on the message. This may start to
stretch the ecological validity of the paradigm however as it seems difficult to find situations
where a sender would lie without incentive (Levine, Kim, & Hamel, 2010).
ORI, message processing, and message comprehension. One surprising finding in this
experiment was that while the outcome relevance induction did systematically vary with
measured ORI scores, neither the conditions nor the ORI scores systematically varied with
message processing as indicated by the number of relevant thoughts produced or message
comprehension indicated by scores on the recall measure. This finding seems to fly in the face
of well-tested hypotheses in the persuasion literature. Perhaps this finding is simply sampling
error. Another explanation for this may be that the thought-listing procedure measures
constructs in addition to message processing such as loquaciousness. It may also be the case that
while intellectually participants were abstractly aware that a message may have more or less
impact on their lives they were not very psychologically engaged with the message due to the
artificial laboratory setting.
Message processing and comprehension were also not related. One explanation for this is
that receivers who were diligently paying attention to the message may not have been trying so
much to memorize the message but instead engage with and evaluate the message’s content. As

27

explained, the recognition measure failed to produce useful variance and it may be that with the
recall instrument alone that message comprehension was inadequately measured.
Another possible explanation for the lack of findings here is related to the coding of
relevant thoughts. Two major possibilities present themselves. First, the definition of relevant
thoughts may have been too broad and more heuristic or peripheral cognitions may have also
been captured in measuring relevant thoughts. Second, if even only the most central, deliberate,
systematic thoughts are counted as relevant, not all relevant thoughts may be of similar quality.
Some thoughts may be more deliberate and focused on the arguments of the message than others.
This is basically a case for measurement error due to crudely imprecise instrumentation. This
problem may have attenuated any observed relationships regarding message processing as
indicated by relevant thoughts provided.
Implications of Findings
Measuring consideration of veracity. While a correlation of .33 between the two
measures used in this study to indicate veracity consideration may not be comfortably ample, it
may not seem as problematic when the unreliability, either estimable or not, or the measures is
accounted for and it is considered that the two measures are very different in the types of
responses required of participants. The two methods rely on very different cognitive response
models (Borsboom, 2006; Borsboom, Mellenbergh, & van Heerden, 2004), an aspect of
measurement that is often elided over in communication and social psychology. The thoughtlisting measure requires a very different behavior from the participant than the closed-ended
scale. Both methods do have some unique drawbacks. The closed-ended measure requires
reflection back to the cognitions that the participant had during the interview. The validity of
this method is contingent on participants being both able and willing to accurately report what

28

they were thinking in the past, two conditions that can be argued against in this case. The
thought-listing measure may fundamentally have the problem of not offering a representative
sampling of the participants conscious cognitions as they may be only reporting their most
salient and easily verbalized thoughts. Given some method variance with the two measures
however, the correlation between the two measures may not seem terrible. Compare the
correlation of .33 to agreement between different observers on otherwise trusted measures such
as with individual differences (e.g., Allik et al., 2010). The thought-listing measure is likely
confounded with other constructs such as loquaciousness or verbal facility but nonetheless may
share substantial variation with the construct of veracity consideration. The closed-ended scale
can be seen as being in an early stage of development and might be improved incrementally with
better or more items.
Instructions work, but not amply. This experiment showed that receivers of messages do
seem to attend more to veracity when they are instructed to do so. The effect size of this
instruction is not overly large however (that evaluation, however, should be tempered by the
qualification that the part of the design testing the effect of instructions was unbalanced in terms
of the number of participants in the no-instruction and instruction conditions and incongruity
between cell-sizes can attenuate relationships). For a literature that ubiquitously trusts their
studies to instructing participants to consider veracity, these findings should be alarming. If one
grants that the literature thus far has neglected to ask whether veracity judgments do indeed
occur in all of the situations that they are asked for in and accepts that the literature, with the
exception of this study and Clare (2013), assume that judgment occurs, it is not apparent that that
vein of research is often strongly inducing a situation wherein people are actively considering
veracity. The one half of the conceptual space that the field is looking at seems to be poorly

29

isolated in the studies that look at it. One is left to wonder how much variation there is across
studies in the extent to which participants are actually considering veracity, even when they are
asked to. Taking this study and Clare (2013) into account, it seems clear that the deception
detection literature has some empirical housekeeping to attend to in delineating what situations
exactly they are investigating.
Question of what instigates veracity judgment still not answered. The single largest
implication of these findings is that aside from ordering participants to consider veracity or
priming them to do so with explicit measures, the field still does not know what instigates
veracity consideration. There is still an open question. Clare (2013) offered a variety of
contextual, sociological, and personality variables that may instigate veracity judgment. Levine
(2014) offers a variety of more message-centric variables that may break receivers from passive
truth-default. Clare (2013) and this paper both replicated the finding that receivers are unlikely
to consider veracity if they are not directly told to do so. This is at the core of truth-default
theory (Levine, 2014): receivers are in a state of passive gullibility unless a trigger impels them
otherwise. This study posited that one blanket situational trigger could break receivers from
truth-default. This experiment investigated only one possible trigger however and future
research should look at the alternatives.
In assessing alternatives, it may also be fruitful to consider Clare’s (2013) initial assertion
that the veracity consideration could be multiply caused and that different causes may be either
necessary or sufficient. Future research may want to take this idea more seriously. From this
experiment’s findings, it does not appear that outcome-relevant involvement is a sufficient single
cause of veracity consideration. It may be that outcome-relevant involvement is a necessary but
not sufficient cause for veracity consideration and that a more immediate cue to deception is also

30

necessary. This cue may be one that Levine (2014) posits, such as demeanor. There is also the
possibility that no one factor is necessary to induce veracity consideration but that no one factor
is distinctly powerful. It is also possible that perhaps truth-default and veracity consideration are
distinctly discrete states and not simply ends of a continuum and that a number or intensity of
factors may be required to meet a latent threshold that tips a receiver from truth-default to
veracity consideration. More sophisticated theorizing that seriously considers multiple causes
may be necessary to actually explain variation in the frequency of veracity consideration.
The salience of veracity judgment in interaction. Any study seeking to explain deception
detection, either prompted or in the case of this study spontaneous, implicitly forwards veracity
judgment as an important and to some extent frequent phenomena in human processing of social
interaction. Seventy-six (40.90%) of the participants sampled in this study did not produce any
verbal behavior on the thought-listing measure that could be construed as considering veracity,
78.50% of the sample produced one or fewer thoughts addressing veracity. The maximum
number of thoughts that were coded as considering veracity for any participant was 4. As has
been explained, the variation in the extent to which participants considered veracity could be
partly attributed to artificial instructions and priming by explicit measures but not by a
theoretically interesting construct: outcome relevant involvement. Taking just the description of
the distribution of how much participants considered veracity, fairly little, and that the
distribution was not explained by what should be a potent construct, outcome relevant
involvement, deception detection does not appear to be central to human communicative
interaction.
In initially reviewing the relevant thoughts in the thought-listing data, it quickly became
apparent that many participants were making comments about the interviewee’s décolleté. The

31

three coders working on the thought-listing data also coded for thoughts about the interviewee’s
upper torso and choice of blouse and achieved reliability of α = .96. Compare the proportion of
people who said anything about honesty in their thought-listing, 59.10%, to the proportion of
participants that talked in their thought-listing about the interviewee’s cleavage, 59.70%. The
presence of thoughts about the interviewee’s appearance is partially explained by a seemingly
theoretically straightforward variable, participant sex. Women were substantially more likely
(68.80%) than men (46.80%) to mention something about the interviewee’s upper torso, χ² =
9.12, p < .01. Mentions of the interviewee’s décolleté were not systematically related to any of
the experimental inductions or whether participants mentioned considering veracity. While this
finding is not being exposited here as central to this study’s investigative endeavor, it may be
useful in illustrating the relative importance of veracity judgment in social interaction. As many
participants paid attention to the appropriateness of the interviewee’s blouse as paid attention to
the honesty of the interviewee. In the face of information explaining that the interviewee might
eventually be in a position to influence their lives, participants still paid as much attention to her
blouse as her veracity. After being instructed to consider the veracity of the interviewee,
participants still paid as much attention to her chest as they did the truthfulness of her verbal
presentation.
Detecting deception could be considered an exceptionally strategic phenomenon in
human interaction. There is a nontrivial prevalence of deception in the social world (Serota,
Levine, & Boster, 2010). Uncovering fraud, avoiding cons, and protecting one’s mental
environment from the pollution of falsity is useful to humans. Detecting deception during
message reception is one way to accomplish those things. However, if perhaps there is only one
take-away from this relatively unsuccessful attempt to explain variation in human interaction, it

32

may be that human experience in communicative situations involves a complexity of different
considerations and cognitions that might not all be obviously strategic.
The blouse choice of the interviewee may at first glance seem to be a rather flip, if not
tawdry, aspect of the interview to pay attention to. Comparing the interviewee’s blouse choice to
whether or not she is being honest in an interview, it would seem that the truth of her presented
qualifications and experiences is more important. On the other hand though, her wardrobe
choice may seem to be important to receivers because it signals her professionalism and
judgment. The presentation, or lack of presentation, of an interviewee’s décolleté may in fact be
quite relevant to a hiring decision. According to the participants, the signal was quite accessible
and easily perceived. One could debate if this is rationally a smart signal to be paying attention
to. Certainly it would be a cheap signal as better judgment could be easily feigned by the choice
of a higher necked shirt. This question aside, it is important to appreciate that the aspects of a
presentation that an audience attends to may not all be obviously strategic and that perhaps even
the non-strategic aspects that are paid attention to are considered to be quite strategic on the part
of the audience. Researchers should not underestimate the gulf between their normative
perceptions of what is important for an audience and what the audience actually believes is
important.
A Future Direction
Any study, especially one with mixed results, will have at many parts little quibbles that
the researcher hopes fellow researchers will heed and thus far a scattering of prescriptions and
exhortations have been made regarding future research. In the interest of clarity, and as a
comment on theory, one well-defined future study will be put forth herein. Parallel to this
program of research concerning what instigates veracity judgment, I have been endeavoring on a

33

few studies not yet to the point of even being citable as unpublished manuscripts that concern the
trajectory of veracity judgments over the course of a message. Basically, over time, how does a
receiver’s veracity judgment of a message change? Do receivers make definite judgments at a
single point in a message or do they gradually move closer to either disbelief or credulity?
Marrying this dynamic question with the current question a study may be sketched that would
show the importance in explaining what instigates veracity judgments and offer a better test of
the ORI hypothesis than this current study. In this study, a lengthy message would be created.
In one condition, throughout the entire message the sender would have exceptionally good
demeanor and seem very believable. In another condition the sender would have terrible
demeanor in the beginning half, in another they would have terrible demeanor in the later half,
and in a last condition they would have terrible demeanor the entire time. These four conditions
would then be crossed with another modification to the message wherein at the half-way point of
the message it would either be made clear to the participant that the message had great salience
to future outcomes of theirs or it would not be. An independent groups experiment would be run
with these eight conditions with the dependent variables at the end being measures of whether
veracity judgment occurred and then a forced-choice veracity judgment measure. The demeanor
inductions should only matter when outcome relevance has been highlighted and only those
demeanor changes to the second half of the message should matter. Demeanor will not be
picked up on until veracity judgment has been instigated via heightened ORI. It would be
expected that in the conditions where participants were not made aware that the message
mattered to them that they did not make judgments of veracity and their final forced-choice
judgments would exhibit high truth-bias. This would be regardless of the demeanor inductions at
any point in the message. In the conditions where the importance of the message was

34

highlighted, participants would report considering veracity. In the high ORI condition, forced –
choice judgments of veracity would be related only to demeanor in the second half of the
messages and not the first and forced-choice judgments should be equally credulous in the
conditions where demeanor was poor in the first half and in the condition where demeanor was
good throughout. Both of these conditions should be comparable to the conditions where
participants were not made aware of the message’s importance. Finally, in the conditions where
a salience cue was given and demeanor was terrible either throughout or at the end, forcedchoice judgments should be equal. This study would show that the instigation of veracity
judgment matters, at least so far as actual final judgments are concerned, and would cleverly
position this process in a more dynamic framework.
Conclusion
Past research (Clare, 2013) had pointed out that without explicit prompting to consider a
message’s veracity, message receivers are unlikely to do so. This paper sought to explain the
frequency with which receivers consider the veracity of messages that they are exposed to. It
was argued that receivers would be more likely to judge the veracity of a message if they have
high outcome-relevant involvement with the message. An experiment was conducted replicating
Clare (2013) with the inclusion of a outcome relevance induction. The inclusion of an
instruction to consider veracity induction was also included, constituting the first ever validity
check of the instructions constant in the deception literature that prompt receivers to judge
veracity. The experiment measured veracity consideration using one measure replicating Clare
(2013) and one new measure that was hoped to offer greater precision and practical benefit.
Results did not show support for the hypothesis that outcome-relevant involvement increases the
frequency with which receivers judge the veracity of the message. The experiment did replicate

35

Clare’s (2013) finding that absent prior prompting receivers are unlikely to consider the veracity
of a message and did show that instructions to consider veracity do increase consideration of
veracity but not entirely amply. This experiment leaves open the question of what factors, other
than artificial laboratory procedures, induce veracity consideration but offers some directions for
future research on that question.

36

APPENDICES

37

APPENDIX A: OUTCOME RELEVANCE INDUCTIONS
High outcome relevance induction
POSITION #ltn891125 - University Special Advisory Board
This position is for a seat on a university-level undergraduate advisory board. A roster of
applicants will be chosen by interview and the applicants will be voted on by the undergraduate
student body.
The advisory board is a new initiative, created to address and negotiate the many changes that
are occurring at MSU that directly affect undergraduates.
The board is going to be invested with direct influence in the form of regular meetings with the
university president and board of trustees.
The board member would have a real vote at the highest levels of the MSU organization.
It is expected that some of the issues that this board will have a part in will include things such
as:
• tuition levels
• curriculum and graduation requirements
• student housing
• Parking
• Welcome Week policies
• the role of the student government organization ASMSU
• student rights
• funding for student health facilities
• Scholarships
• campus rules.
It is important that the person elected to represent the students has the student body's interests at
heart.
It is also important that the student is competent because if they prove to be unreliable or unable
it will likely lead the university to discontinue the advisory board, cutting off undergraduates
from having such a large amount of influence on university policies.
It is similarly important that the representative be an upstanding student who puts forth a good
face for the student body.
The board will begin next semester.

Low Outcome Relevance Condition
POSITION #azd1196 - Comm. Dept. URA
This opening is for a research assistant position in the Department of Communication. Any
student who is enrolled in an independent study may apply. Applicants will be interviewed
about their academic and personal interests and career goals so that the specific duties of the
position may be tailored to them. Duties may include working in a lab, entering data, coding
data, or editing video or audio tapes.
This position will start in the next semester.

38

APPENDIX B: TRANSCRIPT OF INTERVIEW CONSTANT MESSAGE
Interviewer (1): So everybody has strengths and weaknesses. What are your strengths?
Interviewee (2): Well I think, um especially dealing with a job, this job, I’m a really good
communicator; I can talk to anybody and kind of persuade them to do whatever we want them to
do. Um I can, I am very personable so I can get them to be my friend and actually believe me
and think that what we believe in is what they should believe in. Um, yea
1: What about your weaknesses?
2: My weaknesses, um, (long pause) they definitely, sometimes I get really nervous and I can’t
talk to people, I um have trouble like in groups of people, I don’t really have a loud voice so I
can’t, sometimes they don’t hear me. Even if I try to talk loud I get scared and don’t wanna like
be loud enough, um I don’t know.
1: Okay, Well can you tell me about a time when you had to go above and beyond the call of
duty to get a job done?
2: Um, there was one time I was working with one of my friends on a project, and she wasn’t
really uh participating in the work and it was really hard for me because it was my friend so I
didn’t wanna be rude and I wanted to maintain a friendship with her, but I knew that the work
needed to be done so I kind of had to like talk to her and kind of not be mean but be a little
aggressive with her to make sure she got it done and I ended up having to do a lot of extra work
so we, but we got it done, but um just to make sure that it was done.
1: Okay, can you describe one or two of your most important accomplishments?
2: Um, one of my most important accomplishments well in college I guess is getting into college,
that was a uh a big thing, um staying in college, doing well that was something that was really
important to me, to be able to come to a good school and actually succeed which, I think I’m
doing, um to get summer internships and jobs and things like that so, that’s yea.
1: Can you give me an example of your ability to manage or supervise others?
2: um yea I mean like I said when working in groups, I’m I am able to identify what the issue is
with, whatever the issue is, and um give people responsibilities, tell people that, what needs to be
done. If they can’t do it take it upon myself to do it.
1: So tell me about a time when you demonstrated leadership? What were the results?
2: Um, one time we were working in a group and um, my friends, we were all in a group, and
they couldn’t preform the task that we were supposed to do um and I had to kind of step up and
uh, designate like who does what work. Try to get everyone to do the work and make sure that
they did it while like following them up to make sure they were actually doing what they were
told and um yea, and I’d uh make sure that everyone did it right.
1: and what was the outcome of that project?

39

2: we… ended up succeeding, we, everyone got what they needed to do done. Um, we did pretty
well on the project, the project looked good, so it was good.
1: Tell me about a difficult obstacle that you had to overcome in life? How did you handle it?
2: um, a diff, a difficult obstacle… ( long pause, laugh) um (pause) I have no idea( muttered
under breath) ( looks at camera and laughs) um…
1: we can come back to that question…
2: Yea let’s do that
1: if you like, yea, no problem… Uh, Do you consider yourself to be more of an analytical or
impulse driven decision maker?
2: I definitely think I’m a little bit of both, um I don’t do many things like impulsively, at least I
try not to. I usually think but… sometimes maybe I don’t like completely analyze which would
be like completely analytical, I think I’m kind of a mix. So think of things, and if I think right
away it’s a good idea, then I’ll just go for it and look back later.
1: What things would you say frustrate you the most? And how do you usually cope with those
frustrating things?
2: Um, I guess well dealing in a group setting when people are like very reliant on others, I try
really hard not to rely on anyone but myself in groups or anything, cause that’s uh I don’t know,
uh hard but um and it frustrates me when people think that they don’t need to do any work or
they don’t need to do anything and they can still kind of ride along and that frustrates me because
I just, you cant really do anything about those people, you cant tell them not to but there’s just
nothing you can do. But yea try to work with them.
1: So how would you cope with that in a working environment?
2: I guess in a working environment you would, well I would um if we were doing a project or
something and there was one person, or a few people trying to depend on everyone else, um I
guess try to maybe have a backup in case they don’t you know but try to believe that they will do
their best and they will get what they need to do done. Maybe remind them or help them and like
give them extra ways to actually be able to succeed.
1: How would you rate your oral presentation skills? And if you wouldn’t mind tell me about a
presentation where you’ve given uh the presentation successfully.
2: um, well for like formal speeches, that’s what you’re talking about?
1: Mhm
2: um, I get really nervous when I’m formally speaking, I, my freshmen year I had to take a class
that was, you do like a bunch of speeches in it, and wasn’t it, I thought it was going to be in front
of a ton of people, but it was in front of like ten people so it wasn’t awful but I, I mean I think I
did okay in the class like um I was able to get my speeches out with good time but I was just like

40

really really nervous, it was really, I don’t know why I get really nervous like in front of people
but I have huge stage fright, just for speaking though like.
1: So if this job requires you to give formal presentations, do you feel confident in your ability to
perform that duty?
2: I hope so, I think I would definitely try, I would definitely make an effort to be able to
succeed, and be able to get up there and, I know I can do it, it’s just I’d be really nervous doing
it.
1: Have you ever disappointed someone that you’ve worked for? And if so how did you deal
with it?
2: Um this summer in the internship working with a company basically trying to promote and do
things for like an event and there was one time he wanted, it was, I was working with another
kid, they wanted us to put out flyers like all over and we got the flyers and then like it was during
the summer and I was away for the weekend there was just a few things that just kind of
interrupted our process and we didn’t, they were kind of disappointed that we didn’t put them out
right away. We felt really bad and we ended up getting them out like after they spoke to us, we
got them like all out that day, so we really dealt with it but it was really hard because we felt, you
feel really bad when someone’s disappointed with you and all I tried to do was just make up for
it and I think we did. They were happy with us after.
1: Have you ever come upon an ethical quandary at school or work? And uh, tell me how you
approached that, and what you would do differently if you encountered a similar situation in the
future?
2: Oh my god, an ethical issue?
1: Mhm
2: Um, this one time I was, an ethical issue at school?
1: Or work.
2: Or work, um this one time I was dealing with a group and one of them members they wanted
to cheat with something they wanted to with like, I don’t remember what exactly it was but they
wanted to cheat to make it easier for the group and, it was kind of well you don’t really wanna
cheat with school because especially if they find out, even just you wanna do your own work and
just its better, also I think she was just trying to take the easy way out and cheat instead of just
actually doing the work which made us feel like she didn’t really care about the group, um but
we just kind of had to you know make sure she steered away from that, make sure that she didn’t
do that and kind of follow up on her and make sure she was doing her own work or ask her some
questions to see where her information was coming from and things like that but I think, it ended
up being okay. I don’t think she cheated so.
1: Would you do anything differently if you encountered a situation like that in the future?
2: Um, There’s definitely I mean depending on the situation. There’s always something different
you can do or there’s always something better you can do but with that situation I think we did a
41

pretty good job, I think we um, kind of just made sure she was doing the right thing and we let it
be up to her but we made sure so it was.

42

APPENDIX C: CONSIDERATION OF VERACITY SCALE
Asterisks indicate items that were dropped from scale following CFA.
1. It may have crossed my mind if the applicant was being honest or deceptive.
2. I considered at a couple points in the video if the applicant was being truthful.
3. I was on guard for deception, trying to determine whether the applicant was being honest or
deceptive.
4. I paid constant attention to whether the applicant was telling the truth or lying.*
5. Through much of the video I was attentive to whether the applicant was being honest or
deceptive.*
6. I barely paid any thought at all to whether the applicant was lying or telling the truth. (RC)
7. I thought a good bit about whether the applicant might be lying or telling the truth.
8. I was obsessed in trying to determine if the applicant was being honest or deceptive.*
9. Whether or not the applicant was being honest or deceptive was something I considered
during the video.

43

APPENDIX D: OUTCOME-RELEVANT INVOLVEMENT SCALE
Asterisks indicate items that were dropped from scale following CFA. Crosses indicate items
that were added to the original Cho and Boster (2005) scale.
1.
2.
3.
4.
5.
6.

Whatever person is selected for the job has little impact on my life. (RC)
All in all, the effect on my life of who gets selected for this job is small. (RC)
My life would change depending on who gets selected for this job.
The choice of what person gets selected for this job has little effect on me. (RC)
My life would not change much depending on who gets the job. (RC)
It is easy for me to think of ways in which the choice of who gets selected for the job affects
my life.*
7. It is difficult for me to think of ways that the choice of who gets selected for this job impacts
my life. (RC)
8. I would be personally affected by whether or not the right person is selected for this job. †
9. It might be harder or easier for me to get what I want out of life depending on the quality of
the person who gets the job.* †
10. I would not feel the effects if the wrong person was selected for the job. (RC) †

44

REFERENCES

45

REFERENCES

Allik, J., Realo, A., Mottus, R., Esko, T., Pullat, J., & Metspalu, A. (2010). Variance determines
self-observer agreement on the Big Five personality traits. Journal of Research in
Personality, 44, 421-426. doi:10.1016/j.jrp.2010.04.005
Borsboom, D. (2006). The attack of the psychometricians. Psychometrika, 71, 425-440.
doi:10.1007/s11336-006-1447-6
Borsboom, D., Mellenbergh, G. J., & van Heerden, J. (2004). The concept of validity.
Psychological Review, 111, 1061-1071. doi:10.1037/0033-295X.111.4.1061
Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source
versus message cues in persuasion. Journal of Personality and Social Psychology, 39,
752-766. doi:10.1037//0022-3514.39.5.752
Cho, H., & Boster, F. J. (2005). Development and validation of value-, outcome-, and
impression-relevant involvement scales. Communication Research, 32, 235-264.
doi:10.1177/0093650204273764
Clare, D. D. (2013). Spontaneous, unprompted deception detection judgments. Unpublished
preliminary paper.
Clare, D. D., & Levine, T. R. (2013). Spontaneous, unprompted deception detection judgments.
Presented at National Communication Association, Washington, D. C.
Gerbing, D. W. (2014). lessR: Less code, more results. R package v3.1, http://cran.rproject.org/web/packages/lessR/index.html
Johnson, B. T., & Eagly, A. H. (1989). Effects of involvement on persuasion: A meta-analysis.
Psychological Bulletin, 106, 290-314. doi:10.1037//0033-2909.106.2.290
Kruglanski, A. W., & Thompson, E. P. (1999). Persuasion by a single route: A view from the
unimodel. Psychological Inquiry, 10, 83-109. doi:10.1207/S15327965PL100201
Levine, T. R. (2014). A theory of human deception and deception detection. Journal of
Language and Social Psychology. Accepted for publication.
Levine, T. R., Kim, R. K., & Hamel, L. M. (2010). People lie for a reason: Three experiments
documenting the principle of veracity. Communication Research Reports, 27, 271-285.
doi:10.1080/08824096.2010.496334

46

Petty, R. E., & Cacioppo, J. T. (1979). Issue involvement can increase or decrease persuasion by
enhancing message-relevant cognitive responses. Journal of Personality and Social
Psychology, 37, 1915-1926. doi:10.1037/0022-3514.37.10.1915
Serota, K. B., Levine, T. R., & Boster, F. J. (2010). The prevalence of lying in America: Three
studies of self-reported lies. Human Communication Research, 36, 2-25.
doi:10.1111/j/1468-2958.2009.01366.x

47